Talk:Journal of Comparative Neurology 15 (1905)

From Embryology

The Journal of Comparative

Neurology and Psychology

Founded by C. L. Herrick


C. JUDSON HERRICK, Manager, Denison University


Harvard University


Columbia University



University of Pennsylvania


J. >rARK BALDWIN. Johns Hopkins University FRANK W. BAKCROFT, University of California LEWELLYS F, BARKER, University of Cliicago H. HEATH BAWDEN. Vas?ar College iLBRECHT BETHE, University of Strassburg G. E. COGHILI^. Pacific University FRANK J. COLE, University of Liverpool H. E. CRAMPTOX, Columbia University C. B. DAVENPORT, University of Chicago \VM. HARPER DAVIS, Lehigh University HENRY H. DONALDSON, University of Chicago LUDWIG EDINGER, Frankfurt a-M. S I. FRANZ, McLean Hospita', Waverley, Mass. THOMAS H. HAINES, Ohio State University A. VAN (JEHUCHTEN, University of Louva'in R. G. HARRISON, Johns Hopkins University C. F. HODGE, Clark University S. J. HOLMES, University of Michigan EDWIN B. HOLT, Harvard University G. CARL HUBER, University of Michigan JOSEPH JASTROW, University of Wisconsin J. B. JOHNSTON, West Virgi a University

B. F. KINGSBURY, Cornell University FREDERIC S. LEE, Columbia University JACQUES LOEB, University of California E. P. LYON, St. Ivouis University

ADOLF MEYER, N. Y. State Paibological Inst. THOS. H. MONTGO.MERV, Jr., Univ. of Texas WESLEY MILLS, McGill University

C. LLOYD MORGAN. University College, Bristol T. H. MORGAN, Columbia University

A. D. MORRILL, Hamilton College HUGO MUENSTERBERG, Harvard University W. A. NAGEL, University of Berlin G H PARKER. Harvard" University STEWART PATON, .Johns Hopkins University RAYMOND PEARL, University of Michigan C. W. PRENTISS, Western Reserve University C. S SHERRINGTON, University of Liverpool G. ELLIOT SMITH. Gov't. Medical School. Cairo EDWARD L. THORNDIKE, Columbia University JOHN B. W.ATSON, Universit\ of Chicago W. M. WHEELER, Am. Museum of Nat. History C. O. WHITMAN, University of Chicago



II o9

The Journal of Comparative Neurology and Psychology

Contents of Volume XV, 1905

Number 1, January 1905.

On the Areas of the Axis Cylinder and Medullary Sheath as seen in Cross Sections of the Spinal Nerves of Vertebrates. By Henry H. Donaldson and G. W. Hoke. (From the Neurological Laboratory of the University of Chicago.) With one figure. . . i

On the Number and Relations of the Ganglion Cells and Medullated Nerve Fibers in the Spinal Nerves of Frogs of Different Ages. By Irving Hardesty. (From the Hearst Anatomical Laboratory of the University of California.) . . . . . . . 17

Editorial '

Psychology and Neurology. ....... 57

The International Commission on Brain Research. ... 62

Literary Notices. ........... 65

Number 2, March, 1905.

Observations on the Spinal Copd of the Emu and its Segmentation. By Irving Hardesty. {From the Hearst Anatomical Laboratory of the University of California.) With four figures. .... 81

The Selection of Random Movements as a Factor in Phototaxis. By S. J. Holmes. (From the Zoological Laboratory of the University of Michigan.) 98

Notes on the Development of the Sympathetic Nervous System in the

Common Toad. By Walter C. Jones, M.D. With twelve figures. 113


Concerning the Genetic Relations of Types of Action. . . . 132

The Basis for Taxis and Certain Other Terms in the Behavior of

Infusoria. ........... 138

The Problem of Instinct. ........ 144

A Review of Some Recent Literature on the Chemistry of the Central Nervous System. By Isador H. Coriat. Worcester Insane Hospital. . . 148

Literary Notices 160

Number 3, May, 1905

The Morphology of the Vertebrate Head from the Viewpoint of the Functional Divisions of the Nervous System. By J. B. Johnston. (Frotn the Zoological Laboratory of IVest Virginia University.) With Plates I to IV 175

Literary Notices 276

Number 4, July, 1905.

The Sense of Hearing in Frogs. By Robert M. Yerkes. (From the

Harvard Psychological Laboratory.) With seven figures in the text. 279

The Reactions of Ranatra to Light. By. S. J. Holmes. (From the Zoological Laboratory 0/ the University of Michigan.) With six figures in the text. .......... 305

Literary Notices 35°

No. 5, September, 1905.

A Study of the Functions of Different Parts of the Frog's Brain. By WiLHELM LoESER, M.D. (From the Physiological Laboratory of the University of Kansas.) . . . . . . . -355

The Central Gustatory Paths in the Brains of Bony Fishes. By C. JUDSON Herrick. (Studies from the Neurological Laboratory of Denison University. No. XVLII.) 375

Literary Notices 457

Number 6, November, 1905.

Some Cellular Changes in the Primary Optic Vesicles of Necturus. By Clarence LoEB, A.M., M.D. St. Louis. (From the Anatomical Laboratory, St. Louis University.) With Plate V 459

Some Results of a Study of Variation and Correlation in Brain-Weight.

By Raymond Pearl. 467

The Relation Between the Occurrence of White Rami Fibers and the Spinal Accessory Nerve. By A. H. Roth, A.B., M.D., Lnstructor of Anatomy in the University of Michigan. (With an Addendum by JPlayfair McMurrich.) With one figure 482

Respiration and Emotion in Pigeons. By John E. Rouse. (From the

Harvard Psychological Laboratory.) . ..... 494

The Effect of the Bearing of Young upon the Body-weight and the Weight of the Central Nervous System of the Female White Rat. By John B. Watson. (From the Neurological Laboratory of tlue University of Chicago.) With Plate VL 514


The Work of Carl Wernicke. 525

Papers on Reactions to Electricity in Unicellular Organisms. By. H. S.

Jennings 528

Literary Notices. . .......... 535



Author's names are in small caps. References to subjects and authors of original articles are marked with an asterisk.

Acrididae, functions of nervous system, 164. Amoeba, reactions of, 170.

  • Action, instinctive, I44.
  • genetic relations of, 132.

Agababow, a. Nerves of the sclera, 70. Amphibia, eyes of, 163. Amphioxus, nervous system of, 350. Anatomy, of head and neck, 67. Anderson, H. K. (and Langley). Regeneration in nerves, 75. Union of nerve fibers, 74. Andrews, E. A. Breeding habits of

crayfish, 79.

  • Animal behavior, 98, 132, 174.

Anodonta, visceral ganglia of, 352. Audition in fishes, 166.

  • in frog, 279.
  • Axis cylinder, area of in spinal nerves, i.

'Dallowitz, E. Olfactory cells of

-»-' Petromyzon, 164.

Banchi, a. Abnormal brain and mind,

352. Baglioni, S. Physiology of reflex movement, 53 ). Barclay-Smith, E. (and Elliott).

Muscular activities of colon, 75.

  • Barker, L. F. Work of Carl Wernicke, 525.
  • Bawden, H. H. Psychology and neurology, 57- _ .

Behavior, determining factors in, 170.

  • forms of, 132.

of Collembola, 78. of crayfish, 79.

  • of leech, 102. .

of lower organisms, 166.

  • of Ranatra, 305.

of spider, 75. ^terminology, 138. BiGELOW, H. B. Hearing in goldfish, 78.

Birds, inheritance of song in, 174.

origin of migration of, 80. BiRUKOFF, B. Theory of galvanotaxis,

73 BoiNET, E. Evolution of medical doctrines, 458.

BoN, G. Le. Psychology of education,

458. Brain, anatomy of, 160. of Egyptians, 160.

  • morphology of, 175.
  • international commission, 62.
  • vveight, variation and correlation,

467. Bressler, J. Mental diseases, 541. BuRKHOLDER, J. F. Anatomy of brain, 160.

CAMUS, J. and Pagniez, Ph. Neurasthenia and hysteria, 76. Cardio-accelerator 'nerves, of molluscs, 72. Carlson, A. J. Heart-beat of Limulus, 72, 278. Heart-rhythm of molluscs, 72. Nerve cord of Myriapoda, 353. Cats, thinking, 539. Catfish, habits of, 278. Cerebellum, localization of functions of,

539tumors of, 3^2.

Cervical sympathetic, union with chorda tympani, 71 .

Child, C. M. Regeneration of nervous system, 354.

Chorda tympani, and cervical sympathetic, 71.

Ciliary movement, in metazoans, 540.

CoGGi, A. Ampullae of Lorenzini, 352.

COHN, P. Feeling and insanity, 173.

Cole, R. L. Thinking cats, 539.

Collembola, reactions of, 78.

•Color preference, in pigeons, 494. Comparative psychology, 164. Co-ordination, nervous nature of^in Lim ulus heart, 72.

  • CoRlAT, I. H. Chemistry of nervous

system, 14S. ♦Correlation, in brain-weight, 467. Cranial nerves, development of, ^^2 distribution of tifth, 161. Crayfish, breeding habits of, 79. CUSHING, H. Fifth cranial nerve, 161.

r^AVENPORT, C. B. Collembola, 7S.

■*-^ Statistical methods, 67.

Deaver, J. B. Surgical anatomy of

head and neck, 67. Deiter's nucleus, 70. Delphinin, action on nerve endings, 72. Disease, frontiers of, 45S. Dog, spinal reflex in, 75. ♦Donaldson, H. H. and Hoke, G. W.

Spinal nerves, i.

TIJ* ducation, psychology of, 458. ■'--'Egyptians, brain of, 160.

  • Electrotaxis, in unicellular organisms,

528. Elliott, T. R. and Barclay-Smith, E.

Muscular activities of colon, 75. •Emotion in pigeons, 494.

  • Emu, spinal cord of, 81.

Equilibrium, and forced movements,"

73EwiXG, H. Z. Functions of nervous

system in Acrididae, 164. Eye, in Amphibia, 163. physiology of, 73.

  • pishes, gustatory paths in brain of,

375- . hearing in, i66. sense of taste in 277. Fleury, M. De. Manual of diseases of

nervous system, 76. Freidenfelt, T. Visceral ganglia of

Anodonta, 352.

  • Frog, functions of brain, 355.

'hearing of, 279.

nervous system and musculature, 68. •number of ganglion cells etc., 17. FtJRBRlNGER, M. Fourth nerve, 69.

Goldfish, sense of hearing in, 78.

Goldstein, K.;^ Development of brain, 162. Influence of nervous system on development, 65.

Gross, H. Criminal psychology, 353.

T_J abit, migration' So. ' •■■ -^ Habits, of catfish, 278. of Collembola, 78. of crayfish, 79. •of Ranatra, 305

  • Haines, T. H. Instinct, 144.

Hair, sense, organs of, 16.). Halben, R. Pigment and light perception, 174. Hall, G. S. Adolescence, 77. Hardesty, I. Development and nature of neuroglia, 68. •Ganglion cells and nerves of frogs,


  • Spinal cord of Emu, 81.

Harris, W. Binocular and stereoscopic vision, 4;7.

Harrison, R. G. Nervous system and musculature of frog, 68.

•Head, morphology of vertebrate, 175.

Heath, H. Nervous system of Solenogastres, 163.

Hellpach, W. Psychology of hysteria,

541 Hericonot, J. Frontiers of disease, 458. •Herrick, C. J. Gustatory paths in fishes, 375. Sense of taste in fishes, 277. Hearing in fishes, 166.

  • in frogs, 279.

in goldfish, 78. HocHSTETTER. Fissures in brain, 351. •Holmes, S. J. Phototaxisand random movements, 98.

  • Reactions of Ranatra, 305.

HoRNADAY, W. T. Natural history, 173. Horse, Berlin "thinking," 539. Hyde, I. H. Respiratory centre in

skate, 70. Hypophysis, structure and function of,

i6^ 542. Hypochondria, 540. Hysteria, psychology of, 541,

Galvanotaxis, theory of, 73. * Infusoria, terminology of behavior of,

  • Ganglion cells, number of in spinal -^ 138.

nervesof frogs, 17. Inheritance, of song in birds, 174.

Gardella, E. Influence of phenic acid Instinct, and will, 276.

on taste, 71. *-Instinct, problem of, I44 Gehuchten, a. Van. Structure of Ioteyko, J. Reaction of degenerated

nerve cells and connections, 69. muscle, 72. Vestibulo-spinal tract, 70. Genital organs, nerve terminations in, 68

Irritability, and degeneration of muscle, 72.

JACOBSOHN, L. (and Mendel.) Neurology and Psychiatry, 67. Jennings, H. S. Behavior of lower organisms, 166. ^Reactions to Electricity in unicellular organisms, 528. •Taxis and other terms, 13S.

  • JOHNSTON, J. B. Morphology of vertebrate head, 175.
  • JONES, \V. C. Nervous system of toad,

"3 Joseph, H. Nervous system of Amphioxus, 350.

KAMON, K. "Geruchsknospen," 163. Kendall, W. C. Habits of catfish, 278. Kerr, J. B. Motor trunks in Lepidosir en, 352. Klein, F. Physiology of the eye, 73. KoELLiKER, A. Nerve fibers, 350.

T ANGLEY, J. N. Commissural fibers,

■*— ' between nerve cells, 74.

Union of cervical sympathetic with chorda tympani, 71.

Langley, J. N. and Anderson, H. K. Autogenetic regeneration in nerve fibers, 75. Union of nerve fibers, 74.

  • Learning in lower organisms, gS.

Leptoplana, regeneration of nervous system, 354.

Levi, G. Histogenesis of ammonshorn, 163.

Lewis, W. H. Development of amphibian eye, 163.

Limulus, heart-beat of, 278.

co-ordination of heart-beat, 72.

  • LoEB, C. Optic vesicles of Necturus,

359LoEB, J. General physiology, ^76.

  • LoEbER, W. Functions of frog's brain,

355LiBOsCH, W. Olfactory organs of

Petromyzon, 351. Lukas, F. Comparative psychology, 164.

TV/Tay, W. p. Innervation of muscu ■^'■*- lature of stomach, 75.

•Medullary sheath of spinal nerve, i.

•Medullated nerve fibres, in spinal nerves of frogs, 17.

Meehan, J. The "thinking" horse, 539.

Mendel, E. and Jacobsohn, L. Neurology and psychiatry, 67.

Mental traits, inheritance of, So.

Methods, statistical, 67.

Mettler, L. H. Diseases of nervous system, 354.

Migration of birds, origin of, 80.

Molluscs, heart-rhythm and nerves, 72.

  • Movement, random and phototaxis, 98.

MusKENS, L. J. J. Maintenance of equilibrium etc., 73.

Myriapoda, physiology of nerve cord, 353 ^Jagel, W. Hand-book of human

■^_^ physiology, 535. Natural history, 173.

of salmon, 351.

  • Necturus, optic vesicles of 359.

Nerve cells, structure and connections.

Nerve fibres, commissural, 74. development of, 350. union of 74. Nerves, of limbs, 75.

of sclera, 70.

  • Nerve, spinal accessory, 4S2.

trochlear, 69.

  • Nervous system, effects of bearing of

young on, 514. '^chemistry of, 148.

  • sympathetic in toad, II3.

diseases of, 76, 354.

  • functional divisions of, 175.

functions of in Acrididae, 164. influence of on development, 65. of Amphioxus, 350.

  • of fishes and gustatory paths, 375,
  • of frog, 355.

of Myriapoda, 353. peripheral, 350. regeneration of, 354. and musculature in frog, 68. Neurasthenia, and hysteria, 76. Neuroglia, 163.

development and nature of, 68. Neurology and psychiatry, 67.

  • Neurology and psychology, 57.

physiological principles in, 173. Neurones, connections of, 69.

Optic vesicles, of Necturus, 359. Ostrich, spinal cord of, 69.

PAGNiEZ, Ph. (and Camus). Neuresthenia and hysteria, 76. Pain, a pseudeffective reflex, 74. Parker, G. H. Ciliary movement in metazoans, ^40. Hearing in fishes, 166.

  • Pearl, R. Brain-weight, 467.

Petromyzon, chemical sense cells, 164.

olfactory organs of, 351. Phototaxis, 167.

  • and random movement, 98.

•of Ranatra, 305.

Physiology, hand-book of, 535.

of adolescence, 77.

of nervous system, 276. Pick. Zones of the head, 540.

  • Pigeons, respiration and emotion in,

494PiNKUS, F. Hair and sense organs,

162, 164. Prince, M. Sensory tibres in spinal

cord, 161. Protozoa, light perception of, 174. Psychiatry, and neurology, 67.

and zones of the head, 540. Psychology, abnormal brain, normal

mind, 352. Psychology, and neurology, 57. criminal, 353. of adolescence, 77. of hysteria, 541. of lowest animals, 164. PuTMAM, J. J. Principles in neurology, 173 T3 AMSTROM. Innervation of periton■»-^ eura, 351, Ranatra, reactions of, 305.

  • Rat, effects of bearing young on, 514.

Rawitz, B. Inheritance of mental characteristics, 80.

  • Reaction, method of in lower animals,


  • types of, 132.

of amoeba, 170.

  • of frog, 279,355.

of lower organisms, 166.

  • of pigeons to light, 494.
  • of Ranatra to light, 301;.
  • to electricity in unicellular organisms, 528.

Reflex, a pseudaffective and its spinal

path, 74. Reflex movement, physiology of, 539. the "sham death," in spiders, 75. Reflexes, nervous paths for, 73.

spinal in dog, 75. •Respiration, and emotion in pigeons,

494. Respiratory centre in skate, 70. Robertson, T. B. Sham death reflex

in spiders, 75.

  • RoTH, A. H. White rami and spinal

accessory nerves, 482. Rothmann, M. Nervous mechanism of

contact reflexes, 73. "*RousE, J. E. Respiration and emotion in pigeons, 494. Rubraschkin, W. Neuroglia, 163. Rutter, C. Natural History of Salmon,

35'Rynberk, G. van. Localization of

functions of cerebellum, 539.

Salmon, natural history of, 351. ScAFFiDi, V. Efferent fibres in the posterior root, 540. Hypophysis of man, 163. ScHAPER, A. Fissures in brain, 351. Schiller, V. Action of delphinin, 72. SCHULTZE, O. Peripheral nervous system, 350. Sclera, nerves of, 70. Scott, W. E. D. Inheritance of song

in birds, 174.

  • Segmentation, of spinal cord of Emu, 8 i .
  • Sense organs, morphology of, 175.

of hair, 164. physiology of, 173. Sensory fibers in spinal cord, 161. Sfameni, p. Nerve terminations of female genital organs, 68. Sheep, brain of, 160. Sherrington, C. S. Spinal reflexes in

dog, 75Skate, respiratory centre in, 70. Smell, in Petromyzon, 351. Smith, G. E. Brain of Egyptians, 160. ■■■■international brain research commission, 62. Solenogastres, nervous system of, 163 SouKHANOFF. Structure of spinal ganglion cells, 70. Spiders, sham death reflex of, 75.

  • Spinal accessory nerve, 482.

Spinal cord, blood supply of, 161. '•'■'Spinal cord, of Emu, 81. sensory fibers in, 161. structure of in ostrich, 69. Spinal ganglia, structure of cells, 70. Spinal nerves, development of, 353.

  • Spinal nerves of frogs, 17.
  • of vertebrates, i.

Spinal path, pseudaff"ective, 74. Statistical methods, 67. Sterzi, G. Blood supply of cord, 161. Structure of hypophysis, 542. Hypophysis of Petromyzon, 542. Stewart, T. P. A. A contrast experiment, 75. Stimuli and sense organs, 73. Streeter, G. L. Cranial and spinal nerves, 353. Spinal cord of ostrich, 69.

' I "aste, in fishes, 277. -•• influence of phenic acid on, 71. '*paths in brain of fish, 375. Tavernek, p. a. Origin of migration

of birds, 80.

  • Taxis, basis for term, 138.

Thought, in horse and cat, 539.

  • Toad, sympathetic nervous system, 113.
  • TriaI and error method of learning, 98.

method of reacting, 172.

  • Tropism, terminology, 138.
  • and random movement, 98.

theory of, 169.


nicellular organisms, reactions of to electricity, 528.

"\7'agus nerve, action of delphinin on, 72.

• Variation, in brain-weight, 467. ♦Vertebrate, morphology of head of, 175. binocular vision in, 457.

  • spinal nerves in, I.

Vision, and pigment, 174.

binocular and stereoscopic, 457.

  • V\/'atson, J. B. Effect of bearing

young on rat, 514. Werndly, L. Turning fork-sound etc.,

75Wernicke, C. Work of, 525. Weysse, a. W. Animal behavior, 174, WOLLENBERG, R. Hypochondria, 540. WooDWORTH, R. S. and Sherrington, C. S. A pseudaffective reflex, 74.

  • V7'erkes, R. M. Genetic relations of

■* types of action, 132.

  • Hearing in frog, 279.

The Journal of Comparative Neurology and Psychology

Volume XV 1905 Number 1


By Henry H. Donaldson and G. W. Hoke.

{From the .Veurologkal Laboratory of the University of Chicago.) With one figure.

Introduction. — The results presented in this paper are, in each case, based on averages of the measurements of twenty or more spinal nerve fibers. The nerves were taken from various animals representing the five great classes of vertebrates. The measurements show that the areas of the medullary sheath and enclosed axis are nearly equal, and by consequence that the volume of the substance forming the axis cylinders is equal to that forming the medullary sheaths. The relation constitutes a point of similarity remarkable for its wide extension through the vertebrate series.' It enables us, moreover, to estimate in any nerve the volume of the substance specialized for the conduction of the nerve impulse. Since this quantitative relation between axis cylinder and sheath is so close, it strongly suggests that in some way the axis controls the formation of the surrounding medullary substance.

In the spinal nerves of some animals this relation, as expressed by the equal areas of the axis and sheath when the fibers are seen in the cross section, was pointed out several years ago

^ The Acrania and Cyclostorai do not develop medullary sheaths on their nerve fibers, and are therefore not included.

2 Journal of Comparative Neurology and Psychology.

(Donaldson, 1895, p. 154). The fact that it was maintained in the growing fiber was also noted (Donaldson, 1901, p. 180; 1901, A, p. 326); and later Dunn ('go, '02) verified the relation by the study of the meduUated fibers in the sciatic nerve of the frog. It seemed desirable, however, to extend the observations on this point, and in 1901, Mr. Hoke took up the question and determined the relative area of the axis and sheath in cross sections of fibers from the nerves of 27 species of vertebrates representing the five great classes.^

Technique. — The animals were killed with chloroform. The nerve, usually from the brachial plexus, laid bare and partially fixed in sitii with osmic acid (i ^ sol.). After half an hour the nerve was removed on a piece of cardboard to prevent shrinking, and replaced in a i ^ solution of osmic acid for 24 hours. Then imbedded in paraffin by the usual method. The sections were cut 3.5 a thick and mounted in colophonium.

The measurements of the large fibers were made under the magnification of 340 diameters, or, in some cases, 265 diameters. The very small fibers were measured with the 1-12 oil immersion.

Tlic Effect of Osmic Acid Treatment upon the Size of MeduUated Peripheral Nerves. — This reaction was studied because of its obvious bearing upon the observations here presented.

Boll ('76) states that i % osmic acid causes a swelling of the sheath to almost double its normal size. This, however, was the result of putting a fiber which had been teased out ivJiile fresh, in a drop of the acid and examining it after a short time. Prolonged immersion in osmic, he says, in the same place, is followed by a shrinkage of the fiber. These statements lose value by reason of the fact that teasing out the fresh fibers stretches them and thus alters their reaction to the reagent.

BovEKi ('85) concludes that the normal form and structure of the medullary sheath is very little modified by the osmic acid

' Mr. Hoke's account of his work was accepted as a thesis for the degree of Master of Science in the department of Neurology of the University of Chicago in 1902. The data in that thesis form the basis of the present paper.

Donaldson and Hoke, Medullary Sheath. 3

treatment, and this view is shared by a number of other observers who have studied this question. The method employed in this investiijation was practically the same as that used by


Speaking generally, it may be said that mistreatment tends to cause a swelling of the medullated fiber (in i % osmic acid) in which the sheath becomes somewhat more swollen than the axis. The most reliable measurements, therefore, are those made on nerves which have suffered the least mechanical damage.

To determine whether our method of treatment produced permanent alteration in the size of the nerves used, a series of observations was made upon the eighth, ninth and tenth spinal nerves running free along the dorsal wall ot the body cavity of the frog, and also on the sciatic nerves of the white rat.

The nerves were laid bare and a bristle bent in the form of the letter U, exerting a tension of .25 to .40 grams, according to the size of the nerve, was rapidly tied to either end of the nerve. In this way about one centimeter of nerve was included between the ends of the bristle. The nerve thus prepared was then removed and before putting it in any fluid its diameter was carefully measured under the microscope. It was then placed for twenty-four hours in a shallow cell containing osmic acid. The subsequent treatment was exactly similar to that given under the paragraph on technique. Bristle and attached nerve were finally mounted in colophonium and the diameter again carefully measured.

The results of these observations are presented in Table I for the Frog and Table II for the Rat.

The column on the left gives the diameter in [x of the fresh nerves with the bri.sile attached. On the right is given the diameters in n while the nerves are in colophonium, after complete treatment by the osmic acid method as described above. In Table I, the middle column introduces the diameter of the nerves after having been in osmic acid twenty-four hours. The final number at the foot of each column gives the square of the average radius. This number, when multiplied by /r would

4 Journal of Comparative Neurology and Psychology.

give in sq. fi the total area of the nerves measured. As indicated, the nerves of the frog have increased in area 2.4%, those of the rat, .57%.

The measurements upon the rat were the last made, after some skill had been acquired in this manipulation, and are probably the more reliable.


To show the effect of the technique upon the eighth, ninth aud tenth spinal nerves of the Frog. Seven specimens.

Diameter in /.i. Fresh Fiber, Bristle Attached.

Diameter in u. Diameter in /U. After Being in In

Osmic Acid i off Colophonium. 24 hrs.

























3716.85 sq. H

65264.92 sq. //

Av. Diam.

Av. Radii Square

Percentage difference -f- 2./^cL.


To show the effect of the technique upon the sciatic nerve of the White Rat. Nine specimens.

Diameter in ix.

Diameter in //. Fresh Fiber,


Bristle Attached.












1 1 54.0








Av. Diam. 1039.5 1042.5

Av. Radii Squared 10S0560.2 sq. // 1086806.2 sq. //

Percentage difference -(- -57^

Donaldson and Hoke, Medullajy Sheath. 5

The results of these observations seem to justify the conclusion that the osmic acid treatment, followed by the preparation for examination, produces but little change from the normal diameter (or area) of peripheral nerve fibers.

Conditions Dcternwnng the Choice of the FibcTS to be Measured. — There were used for measurement only those fibers which had been cut at right angles to their long axis and which stood vertically. It is easy to see that any departure of a fiber from the vertical would make the measurements for the thickness of the sheath too large, and those for the axis correspondingly too small. From measurement M'ere excluded those fibers in which the medullary sheath appeared double, as occasionally occurs when the fiber has been cut through an enlarged cleft of Lantermann.

The rare instances in which the section passes through an internodal nucleus, or just above or below a node — where the relative area of the medullaiy sheath is very greatly increased — were easily avoided. A fiber in which the sheath was wrinkled, or which departed much from the circular form, was not measured. Where the section of the nerve fiber was suitable in other respects — and at the same time was slightly oval — two diameters were taken and the mean taken as the value to be used. Among the small fibers those that were stained gray and not black were classed as immature, and were not measured. It is among the small fibers that the greatest normal range in the relative development of the medullary sheath appears, and it is here too, that the greatest difficulties in making exact measurements are met, any departure from the vertical being especially disturbing.

The very small medullated fibers which appear in the rami communicantes were not studied in this investigation.

Method of Measurement. — On sections of fibers thus prepared and thus selected, the diameter of the entire fiber was first measured and then the diameter of the axis. In order to give an idea of this procedure we may take as an example the first group of ten measurements, the final results of which appear in Table VI after Specimen I. In the first column of Table

6 Journal of Co}}iparative Neurology and Psychology.

in (see below) is t,nven under A + S, (axis + sheath), the total diameter of each fiber, while under A (axis) is t^iven the diameter of the axis. The two columns to the ri_<j;ht i^ive the corre spondint^ radii. The measurements were first put down in terms of the divisions of the ocular micrometer — the subdivisions being estimated to tenths of these units.

'I'Aiu.i'; III.

Diameters ami llidii of entire nb;;r.s and axes, in ilivisioiis of the (xuhir micrometer.



A +



A + S









1 .41;











2. so

I .So












I. bo








1 .()0

To transform these measurements into // we multiply in each instance by the value of one division of the ocular micrometer which in this case was 3.06 /i. As the radii alone will be used, they only are reduced to //. (See Table IV.)



A + S




















7-2 +


It is on the basis of these radii that we calculate the areas of the fibers and of their axes, and for this we employ the formula tt?* giving tt the value of 3. 14.

As the square of the average radius for the grt)up is not equivalent to the average of the squares of the radii, and as the

Donaldson and Hoke, Medullary Sheath. 7

latter is the number desired, it is necessary first to find the square of the radius in the case of each of these ten fibers ; then to take the average of these squares and multiply by n in order to obtain the average area of the fibers.

TABLE V. The squares of the foregoing radii are as fo Hows : (Radius A + S)« (Radius A)^







40. (;6












SI. 84





493-3} ■

24^.29 sq. ti





24-53 sq- "

The average for the squares of the radii in the case of the axis and sheath and of tiie axis alone, must be multiplied by ^ (= 3. 14) to give in square // the areas of the entire fiber and of the axis respectively. In this instance the areas are as follows:

Areas. Entire Fiber Axis

154.9 sq. t.1 76.99 sq. u

The object of this investigation is to determine whether in the cross section of the fiber the area occupied by the ring-like sheath is equal to that of the enclosed axis.

By hypothesis they should be equal in area, hence in the case of the average entire fiber containing 154.9 ^q- /-< in its section we should expect to find one-half of this area 1^^= 77-5

sq. iL in the axis and the other half in the sheath.

With this ideal area, the area of the axis as observed is compared. Thus: Estimated area of sheath = 77.5 sq. /i. Observed " axis =76. 99 sq./i.

According to the hypothesis, the area of the axis should equal that of the sheath. The observed area of the axis is seen to be less by o. 5 i sq. //, or using the ideal area of the sheath

8 Journal of Comparative Neurology and Psychology.

as the standard, it is less by 0,6%. In this particular group, therefore, the actual area of the sheath is 0,6% greater than it would be if the assumed one to one relation of the axis and sheath were maintained.

In working out the results, it is this ideal of one-half the area of the total fiber which is always taken as the standard, and the observed area of the axis is compared with it. If the area of the axis is less than one-half the area of the total fiber, then it follows that the sheath must have been mor^ than onehalf and the percentage value of the difference is written in plus, to show that the sheath is too large by this amount, or, under the reverse conditions, as minus, to show that it is too small. This difference is designated the average percentage deviation.

In the case of each of the 1540 fibers here presented (Table VI) calculations similar to those just given, have been made. It is thus possible to say in each instance by what percentage the area of the sheath departs from that of the ideal, although in the table only the percentages for the extreme cases and for the average deviation are given.

Description of the Material Employed. — The following list of the specimens gives the common name ; scientific name (entered only once where several specimens of the same species were examined); length of body; weight; sex; age; season when killed; nerves taken; locality; and by whom prepared. Where no statement is made; the sex is male; the season winter; the nerves are taken from the brachial plexus, and the material killed and prepared by Mr. Hoke in this laboratory. The omission of any of the other data means that they are not available.

Donaldson and Hoke, Medullary Sheath.


s i

I- o



o -^ ■5 5j


< 5:


£ rt O -c

- or +












/ 22..




1 22.0



r] J



1 9.6




•« (








iS j

















16. 1
































2 \

15 1

^ /












1 1.6

1. 90

a \







2 < j

/ XV
















II. 2





Mud Puppy







- \








Horned Toad














« L







1. 00

-S 1












II. 6














10. 1
















White Rat









1 I.I















1. 10






24. s






























Wild Cat




. 10.3










Manila Monkey







Spider Mor














1. 00











I l.O












For the 1540 fibers the average percentage deviation := -j- 0.45%.

lo Journal of Comparative Neurology and Psychology.

Description of Specimens on which Measurements have been made.

Specimen i. Dog Shark {Miistelus cams), length 9^ cm., summer, Woods Hell.

Specimen 2. Dog Shark, length 9^ cm., summer, Woods Holl.

Specimen 3. Common Skate (Raja erinacea), length 49 cm., female, summer. Woods Holl.

Specimen 4. Common Skate, length 48 cm., female, summer. Woods Holl.

Specimen 5. Common Skate, length 45 cm., female, summer. Woods Holl.

Specimen 6. Summer Flounder [Paralicthys dentatus), length 40 cm., female, summer, W^oods Holl.

Specimen 7. Summer Flounder, length 49 cm., summer. Woods Holl.

Specimen 8. Summer Flounder, length 49 cm., summer. Woods Holl.

Specimen 9. Frog {Rana virescens), body-weight 22 grms., sciatic nerve, summer.

Specimen 10. Frog, body-weight 49. 8 grms., sciatic nerve.

Specimen 11. Frog, body-weight 31 grms., second spinal nerve.

Specimen 12. Frog, body-weight 31 grms., 8th spinal nerve.

Specimen 13. Frog, body-weight 74.5 grms., female, second spinal nerve.

Specimen 14. Frog, body-weight 74.5 grms., female, 8th spinal nerve.

Specimen 15. Frog, body-weight 74.5 grms., female, dorsal root of the 7th spinal nerve.

Specimen 16. Frog, body-weight 74.5 grms., female, ventral root of the 7th spinal nerve.

Specimen 17. Toad (Bufo lentiquiosus), sciatic nerve, summer.

Specimen 18, Mud-Puppy (Xecturus ??iacula(us) female, length 18 cm., summer.

Specimen 19. Common Lizard {Sceioporus tniiiulatiis), weight 8.2 grms.

Specimen 20. Horned Toad {Phrynosoma cormitum), hybernating.

Specimen 21. American Chameleon (Anolis principalis), weight 2.55 grms.

Specimen 22. Turtle [Chryseniys inarginata).

Specimen 23. Domestic Duck, female, 3 weeks old.

Specimen 24. Chick, female, young.

Specimen 25. Pigeon (Columba livia) .

Specimen 26. White Rat (Mus rattiis var. albiis), body-weight 104 grms.

Specimen 27. White Rat, body-weight 10 grms., 6th cervical nerve.

Specimen 28. White Rat, body-weight lO grms., 4th thoracic spinal nerve.

Specimen 29. White Rat, body-weight 10 grms., 2nd lumbar nerve.

Specimen 30. White Rat, body-weight 25 grms., 6th cervical nerve.

Specimen 31. White Rat, body-weight 65 grms., 6th cervical nerve.

Specimen 32. White Rat, body-weight 150 grms., 6th cervical nerve.

Specimen 33. Rabbit [Leptis cuniculus), adult, male.

Specimen 34. Agouti [Dascyprocta aguti), weight 1350 grms.. Dr. Wkv.DESTY, San Francisco.

Specimen 35. Black Bear (Ursics aimricanus), Dr. Hakdesty, San Francisco.

Donaldson and Hoke, Medullary Sheath. i i

Specimen 36. Gv&y Yoy. {Urocvon c/mro-nr<^entatus), body-weight 3940 grms.,

Dr. Hardesty, San Francisco. Specimen 37. Shepherd Dog, male.

Specimen 38. Wild Cat (Feiis rufus), body-weight 3050 grms., Dr. Hardesty, San Francisco. Specimen 39. Cat {Felis ,io))icsticus) , male. Specimen 40. Manila Monkey {.Macacus cynomolgtcs), body- weight 965

grms., Dr. Hardesty, San Francisco. Specimen 41. Spider Monkey (A(eles paniscus), body-weight 1650 grms.,

Dr. Hardesty, San Francisco. Specimen 42. Rhesus Monkey {Macacus rhesus), body-weight 2720 grms.

Dr. Hardesty, San Francisco. Specimen 43. Black Coated Ape ( Macacus inuus), body-weight 5216 grms.,

Dr. Hardesty, San Francisco. Specimen 44. Baboon {Cynofe/>halus bahuin), body-weight 6247 grms.. Dr.

Hardesty, San Francisco. Speciman 45. Man, adult, sciatic nerve.

Accuracy of Measurements and Sources of Error. — In estimating the value of the measurements given in Table VI, it is to be remembered that with the magnification most commonly used for the larger fibers, one-tenth of a division of the ocular micrometer or o. 306 ft was a shade over 3 5^ of the diameter of a fiber 10 [j. m thickness. In case, therefore, that a fiber was measured as 9.7 ijl or 10.3 // instead of 10 //, the deviation in the resulting aicas from that based on 10 a would be approximately, ± 6%. Individual cases might, therefore, readily vary by this amount as an error of observation.

With fibers of greater diameter the relative value of this error would decrease, while with those of less diameter it would increase. Similarly with the higher magnification used for the smaller fibers, the value of the one division of the ocular micrometer was 1.42 /i, and one-tenth, or 0.142 a would be 3 % of a fiber 4.7 ti in diameter. In this case also the error of onetenth of a division in the reading, either plus or minus, would give rise to corresponding deviations in the resultant areas, with a relative decrease as the fibers become larger and an increase as they become smaller.

In general, therefore, the errors of observation would increase as the fibers diminished in diameter. Such errors, however, would tend in all cases to balance out and so be reduced as the number of observations becomes large.

12 Journal of Comparative Neurology and Psychology.

Thus when we take the difference between the sum of all the minus percentai^es and the sum of all the plus percentages by which the sheath differs from the ideal half of the area of the entire fiber, we find that for the 1540 fibers, this amounts on the average to + -45 %• That is, on the average in this series, the area of the medullary sheath exceeds the ideal half by + .45%, or one-thirteenth of ± 6^, the deviai ion which might be expected as an error of observation in the case of a single fiber.

As tending to cause deviation from the ideal standard we have already noted mechanical injury of the nerve before hardening and to this must be added delay in dissection after death.

In this connection some of the material received through the kindness of Dr. Hardesty calls for special comment — comment which is made with the full recognition of the good fortune which brought the specimens into our hands and is merely intended to indicate how readily the nerves undergo slight alterations after death. Specimens 35-38 and 40-44 inclusive were given us by Dr. Hardesty, the animals in all cases coming from the Zoological Garden in San Francisco.

Dr. Hardesty writes as follows: "In every case the animals from which I sent you bits of nerve were dead when they came into my hands. Most of them came to me within a few hours after death, and none had been dead more than twentyfour hours. In case of the monkeys, most died of tuberculosis, some of the other animals died of pneumonia, one by dysentery^ one by accident and the bear from poison."

On looking- over our results on this material

Specimen 34,


" 36,


" 38,




show excessively large sheaths, giving 160 fibers with an average of + 4%, while among the remainder of the mammals from San Francisco, namely the bear, Manila monkey, spider monkey. Rhesus and baboon the average is only -(-1.14%. We conclude from this that the nerves of the first group had suffered a slight post-mortem alteration before they could be fixed.











Donaldson and Hoke, Medullary Sheath. 13

The deviations observed. — Table VI shows that the average deviation for all the 1540 fibers is +0.45 %.

On determining the deviations separately for the Fish, Amphibia, Reptiles, Birds and Mammals we obtain the following results :

, Number of Fibers Deviation

- 0.46% + i-70fo

- i.yofc

+ 1-40/0

This tabulation serves to show that in no class do the deviations rise very high, and that with the exception of the mammals, which have been commented on already, the average deviation tends to diminish as the number of cases increases.

On further looking over the individual deviations, as shown in the original records, it appears that in 580 cases (or 37% of 1540), they are more than 6% plus or minus — 266 cases being minus and 314 plus. Thus the excessive deviations are rather evenly divided on either side of the average.

It follows from this that 63 % of the cases deviate only 6% or less, or are within the anticipated error of observation.

From the foregoing it is concluded that despite the individual variations which may be often quite large, there is no tendency for the sheath to depart materially from the ideal standard, when this relation is determined by averaging the measurements obtained from groups of twenty or more fibers.

Figure illustrating variatiofis in the thickness of the medullary sheath. — In order to give a meaning to the numbers that appear in the foregoing Table VI, the measurements applying to Specimens I, XI, XXII, XXIV, XXVIII, and XLV have been selected for illustration (see Fig. i). This selection gives one specimen from each of the great classes below the mammals. In the case of each specimen, the diameter of the average fiber is taken as a standard, and this is shown, enlarged one thousand diameters, (i) with a sheath having just half the area of the entire fiber (Ideal); (2) with the sheath determined from the

14 Journal of Comparative Neurology and Psychology.

Showing Varutions in the NIedullaey Sheath.

Shark' Spec.




To-lle Spec.


Chick Spec.




Man Spec.

Averaoe percenla6e Ideal. deWalion.

Avera<)e Exlreme voriolious. observed. +

o - o o o o o o o

o • o o o

O ' o o o

+0.9 O O

O " o o o

Figure 1

Donaldson and Hoke, Medullary Sheath. 15

average of all the measurements (average observed) ; (3) with the smallest sheath observed in the series, and (4) with the largest sheath observed in the series (extreme variations).

B}' the study of Fig. i, some notion of the range in the size of the sheath is obtained, but at the same time it must be remembered that these extreme cases as seen in (3) and (4) are exceptional, and that most of the deviations in the sheath are of decidedly less extent.


The measurements in Table VI have been conipared with one another in so far as they permitted the comparison of (i) nerves having different diameters ; (2) those afferent or efferent in function ; (3) taken from various parts of the body ; (4) in males and females; (5) at different seasons; (6) at different ages ; and (7) in animals active and those slow in their movements.

From such comparisons there is no indication that any of these conditions modify the relative development of the medullary sheath.

To make these statements with regard to area the basis for a statement with regard to the relative volumes of axis and sheath, is justified for the reason that the medullary sheath in the peripheral system is very nearly coextensive with the enclosed axis cylinder, and the short stretches at either end of the fiber would serve to give only a slight increase in the volume of the axis. Since in cross sections the sheath is on the average a trifle in excess of the axis, this difference would tend to make the similarity in volume even closer.

If the foregoing statements are well founded it should follow that the figures which represent cross sections of nerve fibers treated by the osmic acid method and which have been published in special articles or in text-books should accord with the results here given.

We have examined such figures as they appear in Ranvier

('78), KUPFFER ('83), GasKELL ('85), CaJAL ('99), HUBER ('OO),

and find that when the conditions are observed of selecting only

1 6 Journal of Comparative Neurology and Psychology.

those fibers which are nearly circular, in which the sheath is single and where there has been no increase in sheath from careless printing, the measurements show the same relations which we obtained on the specimens themselv^es.

We feel justified, therefore, in making the broad statement that in the peripheral nerves of vertebrates above the Cyclostomes, we have in general a one to one relationship existing between the nerve sheath and the axis, whether we compare the volumes of the two or measure their areas in the cross sections of osmic acid specimens. It is to be anticipated, however, that somewhere among the fish we shall find examples of less complete development of this sheath, thus bridging the gap between the condition here found and that represented by the nerves of Cyclostomi and Acrania which are without the medullary covering.


1. Boll ('76). Atti della R Accademia del lincei. Vol. I, Series III, 1876.

2. Boveri {'85). Abhandl. d. K. bayer. Akad. d. Wiss., II CL. XV, Bd. II,

Miinchen, 1885.

3. Cajal ('99). Textura del Sistema Nervioso del Hombre y de los Vertebra dos, Nicholas Moya, Madrid, 1899.

4. Donaldson ('95). The Growth of the Brain. Contemporary Science Ser ies, Charles Scribner's Sons, Ne-w York, 1895, p. 154.

5. Donaldson ('00). hum. Nerv. and Ment. Dis., Vol. XXVII, Oct., 1900.

6. Donaldson ('01). The Central Nervous System, in an American Text book of Physiology edited by W. H. Howell, Vol. II, p. 180.

7. Donaldson ('01, A). An Article: The Growth of the Brain, in Reference

Handbook of the Medical Sciences, Vol. II, Wm. Wood & Co., New York, 1901, p. 319.

8. Dunn ('00). Journ. Comp. Neurol., Vol. X, No. 2, May, I900.

9. Dunn ('02). Journ. Comp. Neurol., Vol. XII, No. 4, 1902, p. 326.

10. Gaskell ('85). /oum. Physiol., 1885, Vol. VII, PI. II, Fig. 6.

11. Huber ('00). Bohm-Davidoff and Huber : Text-book of Histology, 1900,

p. 143, Fig. 114.

12. Kupffer ('93). Sitzungsberichte der Math.-phvsik. Classe der k. b. Akad emie der IVissenschaften zu Miinchen, 1883, H.III, PI. I, Fig. A.

13. Ranvier C'78). LeQons sur 1' Histologie du Systeme Nerveux, 1878, Vol.

I, PL 2, Fig. 7.


By Irving Hardesty.

[From the Hearst Anatomical Laboratory of the University of California).


I. Introduction.

II. Material and Technique.

III. General Results of the Enumerations.

IV. The Number of Spinal Ganglion Cells and their Relation to the Num ber of Fibers Connected with the Ganglion. V. Further Observations on the Conditions Determining the Excess of Fibers on the Distal Side of the Spinal Ganglion. VI. The Gain of Cells and Fibers with the Gain in Weight. VII. Summary. VIII. Bibliography.

/. Introductio7i.

That the number of medullated nerve fibers in the spinal nerves increases during the growth of the animal toward maturity is generally accepted. Aside from the evident requirements of growth, the fact has been established on several occasions, at least for the frog and mammal, by direct enumerations of the fibers.

In two papers of some years ago the author ('99 and '00) undertook to determine the rate at which the increase of fibers takes place in the frog. It was then found (i) that the increase occurs more rapidly in younger specimens ; (2) that in specimens of varying weights, counts of the fibers in transverse sections taken at intervals along the ventral root toward the spinal cord

1 8 Journal of Comparative Neurology and Psychology.

and along both the dorsal root and the nerve trunk towards the spinal ganglion, showed a gradual increase of fibers at the average rate of nearly 2 % per millimeter of length of the roots and of the trunk; (3) in considering the Vlth nerves alone of a series of 17 specimens ranging gradually in weight from 5 grams to 79 grams, it was found that in this one pair of spinal nerves there was for each gram of weight gained an average gain of 7.4 fibers in the two nerve roots and 10.4 in the trunk and dorsal branches. As to the numerical relations of the fibers present on the proximal and distal sides of the spinal ganglia, there was always found (4) an appreciable excess of fibers in the sum of the trunk and dorsal branches as compared with the sum of the dorsal and ventral roots. This "distal excess" was found to vary in the different spinal nerves, ranging from 7% to as much as 61 % of the sum of the two roots. It was found to be greater in those spinal nerves which have a relatively greater number of fibers in the dorsal branches, but showed neither a regular increase nor decrease with the increase in the size of the specimens.

BiRGE ('82) had previously counted the ventral root fibers of all the spinal nerves of one side of six frogs varying in weight from 1.5 to III grams and, assuming the fibers of the two sides approximately equal in number, his figures show an average total gain of 5 i ventral root fibers for each gram gained in body weight. Hatai ('02 and '03) has since determined the number of medullated fibers in the dorsal and ventral roots of three given spinal nerves of each of four white rats ranging in body weight from 10 to 264 grams. He finds that the number of fibers in the roots of these three nerves is 2.7 times greater at maturity than in the 10 gram rat, and that the increase is more rapid in the younger animals. Birge, though he did not take into consideration the dorsal branches, obtained noticeable distal excesses of fibers in the several nerves, and he was the first to attach significance to it. Gaule and Lewin ('96) investigating the sacral nerves of the rabbit, found an excess of fibers on the distal side of the spinal ganglion amounting to as much as 19%, and Buhler

Hardestv, spinal Ganglion Cells. 19

('98) mentions a distal excess of 25% in the IXth spinal nerve of the frog.

An investigation of the numerical relations of the axones on the two sides of the spinal ganglion naturally suggests an inquiry as to the number of ganglion cells in it and the relation they bear to the numerical arrangements of the fibers. This paper is offered with the hope of contributing something (i) to the knowledge of the relations between the number of the nerve fibers in the dorsal roots and the number of ganglion cells in the spinal ganglia whence the fibers are supposed to arise ; (2) toward throwing further light upon the conditions determining the excess of fibers on the distal side of the spinal ganglion ; and (3) concerning the influence of growth upon the relative number of ganglion cells and the numerical relations of the fibers in the region of the ganglion.

The investigation was begun some time ago when the author was connected with the neurological laboratory of the University of Chicago. It has been allowed to rest on account of the difficulty in obtaining the material required. Rana vircsccns having been employed in the beginning of this and and the previous papers with which the results of this must be compared, it was thought advisable to continue with this species though it is apparently very rare in California. It finally became necessary that the material from the specimens of the sizes required to complete the work be sent from the laboratory in Chicago, and I desire to thank Professor Donaldson and Dr. Dunn for their kindness and care in obtaining it for me.

Finally, let it be said that the author realizes that the results of an investigation of this kind are valueless unless obtained by most trustworthy procedure.

II. Material and technique.

Ouly three of the spinal nerves of each frog are dealt with. These are the Vth, Vlth, and IXth of each of seven frogs varying in weight from 10.4 to 63.4 grams. The Vth was chosen as a typical thoracic nerve and because, being near the lumbar region, its dorsal and ventral roots are long enough to render manipulation less difficult. The Vlth nerve in many respects is similar to the thoracic nerves anterior to it.

20 Journal of Comparative Neurology and Psychology:

but at the same time it is often larger than the Vth and in some respects resembles the Vllth which belongs almost entirely to the lumbar distribution. Morever, the Vlth was employed in a previous investigation ('oo) and is included in this in order to accumulate data concerning it. The IXth nerve was chosen because it is one of the largest spinal nerves in the frog, and differs from the Vth and Vlth in that all the fibers of its trunk are contributed to the lumbar plexus. These three nerves alone are considered, it being assumed that whatever conditions found true for these are approximately true for the remaining spinal nerves. To enumerate the ganghon cells and the nerve fibers in the dorsal and ventral roots and the trunks of the entire ten spinal nerves would have been so arduous a task that the additional information probably to be obtained by such was considered a reward scarcely sufficient for the ordeal.

To obtain a series ranging as gradually and as widely as possible, each specimen was chosen with reference to its body weight. The smallest specimen of Rana virescens obtained for this observation had a body weight of 10.4 grams. Smaller specimens may be obtained at certain seasons of the year just after the transition from the tadpole stage. The lai-gest specimen of the series weighed 63.4 grams and was, of course, female, since the males seldom exceed 50 grams whi^e the females have been known to attain a body weight of 80 grams \witji the ovaries removed.

In each case the anin^al was chloroformed in the usual way and, if female, the ovaries were removed. It was then weighed and its body weight and sex recorded. The procedure in dissecting, fixing, imbedding, sectioning, etc., was much the same as that followed in the previous investigation and it is there given in detail. The spinal cord was laid bare from the ventral side, the spinal ganglia and dorsal branches carefully exposed and dissected free with the aid of the dissecting microscope. Avoiding pulling and stretching, the dorsal and ventral roots were clipped with fine scissors close up to the spinal cord and the nerve trunk near its junction with the ramus communicans. Then the ganglion with the roots and portion of the nerve trunk attached was laid on a small piece of card-board. There the roots were arranged straight and parallel with each other, and the dorsal branches straightened out parallel with the nerve trunk in order to obtain transverse sections of the two with the same stroke of the knife. In doing this, careful use of the dissecting lens was always advisable lest some of the smaller of the dorsal branches should be overlooked and should remain tangled or doubled back upon themselves, and thus in the sec

Hardesty, spinal Ganglion Cells. 2 1

tions occasion a double counting of the fibers contained in them. In each case the strip of card-board with the specimen arranged upon it was placed into a vial containing i % osmic acid and bearing a label stating the date, the weight and sex of the animal, the number of the spinal nerve, and whether taken from the right or leftside. After being subjected to the action of the fixing agent for 15 or 20 minutes, the parts become stiffened sufficiently to maintain their arranged relations and may be gently lifted from the card-board that the fluid may penetrate more freely from all sides. The larger nerves were always removed into the fluid. After 12 to 24 hours in osmic acid, the specimen was washed about 12 hours in water frequently changed, and then transferred to 70% alcohol. While washing, an outline sketch of the specimen was made for the purpose of orientation.

Each specimen was embedded in paraffin separately in a small paper box on the side of which was copied the label on the vial in which it was fixed, washed, dehydrated, etc. Transverse sections of the two nerve roots of the trunk and dorsal branches were made of about 4 u in thickness, and those from the required localities were mounted serially. The sections of the nerve roots were taken from about midway their length, while those from the trunk and dorsal branches were taken close up to the peripheral side of the spinal ganglion, yet far enough away to avoid any portion of the ganglion itself being involved. Separate slides were devoted to each of the localities required, and the locality mentioned in their labelling. On approaching the ganglion, from whichever side the sectioning was begun, the thickness of the sections was changed. x\ll sections involving the ganglion were made 9 // thick and special attention was given to mounting them serially.

For counting the nerve fibers the photographic method, devised in the neurological laboratory of the University of Chicago ('99) was employed for every section except an occasional very small twig of the dorsal branches. The mounted sections from a given locality were first carefully looked over with the compound microscope, and one suitable for photographing was picked out, that is, a perfect section and one which high power showed to be flat and evenly adhering to the surface of the slide. Under low power this section was marked ■ by a ring of ink on the cover glass. The photographs were made with the aid of the Zeiss projection apparatus. The magnification employed was from 300 to 500 diameters, varying according to the size of the section None of the sections in this investigation when sufficiendy magnified were too large to get on a single plate of the

2 2 Journal of Comparative Neurology and Psychology.

largest size the apparatus accommodates. Since only medullated nerve fibers were to be counted and since, after osmie acid, these are indicated by the dense black deposit of the reduced oxide of osmium in their medullary sheaths, it was desirable to get photographs giving as much contrast as possible. For this reason, some form of the photographic plates used in "process work" was usually employed. These also allow considerable latitude in the exposure and always give strong negatives. With strong negatives, a good quality of blue print paper gives excellent results for the purpose required, and this was used throughout because of the simplicity of its manipulation as well as its inexpensiveness.

In making the counts of the fibers in a section, the photograph of the section was pinned out upon a soft ])ine board and the identical section itself placed under the compound microscope and subjected to higher magnification than that under which it had been photographed. After a few minutes study, any field in the photograph could be identified in the microscope. Then with a sharp pointed pencil, the photograph was divided into fields of about one inch square or less in area, the lines of division passing among and between the fibers instead of across them. The actual enumeration was done with an automatic tallying register, to the thumb-press of which had been fixed a needle holder. Each fiber in one of the outlined fields was punctured by the needle and all doubtful cases were settled by reference to the original under the microscope. The counting of one field was finished before beginning another. The counting apparatus, its use and the trustworthiness of the results obtained with it, are discussed in detail in the paper cited above. The method has since been used with little modification (Dunn 'go, Hatai 'c2, Inobert '03) and the results obtained have been uniformly considered satisfactory. The method has the advantage of being largely mechanical and eliminating as far as possible the psychological sources of error which are most to be feared in investigations of this kind.

It was not deemed neces.sary to photograph certain of the small divisions of the dorsal branches. Such as were small enough to bear a magnification of 700 diameters and still not occupy the field of the microscope, were counted by a modification of the more commonly used net-micrometer method. The procedure followed i.*^ fully described in the author's paper of 1900. By means of the projection of the squares of the micrometer and the completion of the count before the result is known, it is held that all errors of auto-suggestion are avoided.

Hardesty, spinal Ganglion Cells. 23

In counting the ganglion cells an application of the net method had to be used exclusively. Here all the sections involving the ganglion had to be gone over and photography was practically out of the question. Therefore, the problem had to be approached in another way. It was first determined that but few of even the very smallest of the spinal ganglion cells have a diameter of less than 9 /<, and further that 9 fi sections were thin enough to allow the light to pass through sufficiently to distinguish the intervening structures of the cell. For this reason 9 n was chosen as a convenient thickness. On the other hand, however, some of the largest cells of the ganglia may extend through as many as seven consecutive sections of this thickness. To obviate the danger of double counting, a procedure was adopted throughout, which I think was first described by Gaule and Lewin {'96) in their enumerations of the spinal ganglion cells of the rabbit. This consisted in counting only the nucleoli in the sections. As with them, the question of course arose as to the danger of double counting because 'of the existence of double nucleoli. In counting the cells in several of the ganglia here dealt with. I kept a record of the double nucleoli appearing in the sections and found that two nucleoli occurred in the same cell at an average of about 5 cells in 1000 or at the rate of about one-fifth of i % . Cells in which they were observed were usually among the smaller cells of the ganglia. They, of course, were not counted twice. Remembering that these double nucleoli were seen in sections of 9 yu, and that double nucleoli in a nucleus are never separated as much as 9 /< from each other, and further, that certainly not more than half of such cases existing have their nucleoli separated in the plane perpendicular to the plane of the knife in sectioning so that the knife may pass between them, it is therefore evident that the number of cells counted twice because of double nucleoli is probably less than one-fifth oi \%. The proabilities of double counting due to nuclei being cut in half by the knife and thus counted in the two sections, are also such as need only be considered in questions of the most absolute accuracy. In exceedingly few cases did appearances and comparison of the consecutive sections indicate this had happened. None of the conclusions drawn in this paper would be materially affected by an acknowledged error of one-fifth of i % .

The number of cells counted in each section was' recorded separately and not till all the sections of a ganglion had been counted were these numbers added to obtain the total number of cells in the ganglion. The larger ganglia often involved as many as 125 sectionR. In no case was the relation between the number of ganghon cells and the

24 Journal of Comparative Neurology and Psychology.

number of fibers in the dorsal root computed till all the nerves of a specimen had been completed.

The serial sections were always followed for some distance on either side of the spinal ganglion proper in search of out-lying ganglion cells. In one case these comprised the appreciable total of 31 cells. In most cases the number was very small. In some cases they were more abundant in the beginning of the dorsal root, in others there were more in the trunk. In the Vth and Vlth nerves, the mass of the ganglion shows a tendency to protrude distally along the dorsal branches instead of along the trunk. This is due to the probable fact that most of the fibers of the dorsal branches are cutaneous in their distribution.

///. General results of the enumeration.

Following the procedure indicated above, the Vth, Vlth and IXth spinal nerves were investigated from 7 frogs varying in body weight from 7 to 63.4 grams. The number of ganglion cells in the spinal ganglia of each nerve was determined and the number of nerve fibers present in the dorsal root, the ventral root and the number in the dorsal branches and in the trunk taken close up to the peripheral border of the spinal ganglion but beyond the region of outlying gangHon cells. The general results of these enumerations are given in the following tabulated form :

Table I. Showing in their respective columns the number of fibers in the dorsal roots, the number of cells in the spinal ganglia, the number of fibers in the ventral roots and in the trunks and dorsal branches of the Vth, Vlth, and IXth spinal nerves of 7 frogs varying in body weight as indicated in the first column. The ratios of the ganglion cells to the dorsal root fibers (col. C) are obtained by dividing the numbers in col. B by those in col. A. The figures in col. E result from subtracting those in col. F (fibers in the ventral roots) from those in col. G, and the ratios in col. D are obtained by dividing those in col. B by the respective numbers in col. E. The sums of fibers in the two nerve roots and the sums in the trunk and dorsal branches of each nerve are entered (col's. K andG) for purposes of comparison, but especially that the correctness of the amount of the excess of fibers on the distal side of the spinal ganglion (col. I) may be determined. The percentage of the "distal excess" (col. J) results from dividing the amount of the distal excess (col. I) by the sum of fibers in the dorsal and ventral roots (col. K). The ratios comparing the number of fibers in the ventral roots with the number of fibers in the dorsal roots (col. L) are obtained by dividing the numbers in col. A by those in col. F.

Hardest Y, Spinal Ganglion Cells. IXth NERVE Vlth NERVE Vth NERVE


O CT" 4^ Oj kj 11 >1

^ 00^^4OJ4k O 'f'-f^ bo-f>. VJ OJ 4^ O 5'-f^ 004^ ^4 UJ 4^

\0 -^ »-* O ^ C>J O 1-1 lyi Oj 00 ii i- « 

^Ji-iHiC^OUiOJ OJ OOvO u^ -^j ^j 4^

4k 4^ OJ 4^ U> 4^ C>J <^<0 i-n 00 " (yi 00 O VO ^J 00 O 004^ 00 ON — — 4^ — ►.!

lOOO'^ON^J'-" O^-'^J<-^C7^00•^

-J 4^ C>j OO^O -f' I— i-n 00~-J ^J ^J " ■ (^M4».4».WLn w^a Od4>> 4^ "-i 00

ivj u O) ro M N) u toc^ CO u ro M ro CO loo, co co to 10 ^j co co "o ikbb>'^ia)'vib)b co(;iui>4(ocoa)Nj bi^ioo'^i^^ui

hj M |o M ro to CO ro ^ jvj jo __i .* to to to ^ 10 ^, -» -» 10 10 to c>a)>icoai^^^s bui(;ib)N)'->ibs Kji^bo^OHi^ai

^a OO'-n \C Oj Oj G> - M M 00 "-■ On —

U) 00 00 (O ON "1 M

^^4^<-n4:.OjK3— 4^C^4^0jOjOjw ^Jt^Jt.n-^lO'OON Oni-iC^OjOjiviOO <JA On^J MOnOC/0 00"-4'Lri 00U\ O

vO ■" >-i •" -J ^J vC

OJ 00~^ l-n 4^ -• OJ On W On ON ^ NO « 

U i_r KJ — — — "-I ►1 nO ^J 0-) O *> — M On OOU) OOOj Oj

M Ul M U> K> kJ N2

On O ^4 — O O t^

^ 11 O 1>J <-" 004^


^J On 00 On+- 4^- OJ 00O> OJ O i-i OJ 00 ^J M (^ lyi ^ Oj »« 

On'^ On4^ 4i 4i Oj O 4^ W OnOj <~r\ <j\

Oj N ^4 M hJ 0^ >-<

'\>WOJOJOJ«--Hfj'\Jt^OjC^K)MM"-J)JOJOJOJ"W(0tNj On NO OnOJ nO Ln OJ ■-■ ^ ^I ^4 OO-M 4;. OnO>>. "- O Q4^0jOj i-i N00j4^ivt .-■4i4i. hjNO — OOO On C04^ ^4^0--4^ — MIJ-"

CnMOOi-^JOOjM^O'-'W— NO ONlyi ^J"*^^4^4l-n O tJ^^4 ~fti tOOJOi4i-iJiO ""On OOOi aNND~~4ivityiOj(.nOjU-i O -" OOOn

. CO to CO .^^ -» _» to , ^ tfl 0) to to lo to

On (0 p P f J" P 9"^ 9^^^^^^ ^^ P P !^ ^^ P" ."^ ."^ cnU'<^a)(D(ob)bNj co^jooocntocnoobo'rooobxOi^a)

M M W M « « vj 4* 004^ NO NO NO 4^.

to M 00 OnI-o \j\ kj ^•-JNO 4^ M O M

t-n Kj\ ON4k Oj OJ OJ On O •- Q 4- ^4 O NO ~-4 NO On-. 4^ -L^ C^ OJ Oj to U ON ^4 On "" <-n ^4 OOnO M (nJ " ^4 (VI


M W « "- bJ W tj

^ On-^ "" On 0\^4 On^ -^-I on M — M bN--4 0j

M Cn4^ OOOj OnvO

Body weight in grams.

No. of fibers in the >. dorsal root.

No. of cells in the spinal ganglion.

No. of cells per fiber — in dorsal root.

No. of cells per fiber in trunk and dorsal branches less the ^ ventral root.

No. fibers in trunk and dorsal branches M less the ventral root

No. of fibers in the ventral root.

No. of fibers in the dorsal branches.

'Distal excess.'

Percentage of distal excess.

Sum of fibers in the trunk and dorsal q branches.

Sum of fibers in dorsal and ventral roots

Ratio of fibers in

the ventral root to f fibers in dorsal root

26 Journal of Comparative Neurology and Psychology.

In this table the results for each of the three spinal nerves involved are grouped in the order of the body weights of the respective animals employed in order that the figures obtained for the different representatives of a given spinal nerve may be compared more readily with each other as well as with those of the other spinal nerves.

In the first place, it is apparent as far as the three nerves presented are concerned that neither the number of fibers nor the number of ganglion cells in a given nerve of a specimen is regularly dependent upon the body weight of the specimen. This is perhaps entirely due to a fact noted previously ('99) when was given a more detailed description of the macroscopic features of the different spinal nerves of the frog, viz. : that when the spinal neves of the different specimens are are compared, the relative proportion of fibers distributed to a given nerve is by no means fixed. The variations in proportional size are more frequent and marked in the Vllth, Vlllth, and IXth nerves. In the general arrangement, the Vlth is much smaller than the Vllth, and the Vlllth (the largest of the lumbar nerves) is considerably larger than the IXth; but occasionally the relative size of these nerves can be observed even with the unaided eye as decidedly different in different frogs and sometimes, indeed, on the two sides of the same frog. The variations are most frequent in the Vlllth and IXth nerves, the apparent condition being that a portion of the bulk usually possessed by the Vlllth may be contributed to the IXth instead, for when the proportional size of the IXth is greater than usual, that of the Vlllth is less. In the case of the frog weighing 7.0 grams, recorded in the table, it was noticed at the time of its dissection that the IXth nerves of both sides were appreciably larger than the Vlllth nerves. It was included, however, because at the time I was unable to obtain another specimen of Rana virescens as small as this. Some of the other discrepancies in the increase in the number of fibers accompanying the increase in body weight are without doubt due to similar variations. To determine the real rate of the increase, all the spinal nerves of the different specimens should of course be included in the

Hardesty, spinal Ganglion Cells. 27

counting. For the three nerves involved here, this table shows that, though the variation is not very regular, the larger specimens give the larger numbers. In some of the succeeding tables where the numbers are summed up and arranged according to the body weight, this fact is more evident. It is one of the features for which Table III is especially constructed.

In the last column of Table I the number of fibers in the ventral roots are compared with the number in the dorsal roots. The proportional relations are expressed in the form of ratios. It is seen (i) that in every case the dorsal root fibers exceed those in the ventral root ; (2) that in the Vth nerve this excess is greater than in the Vlth while in the Vlth it is greater than in the IXth, and (3) that the ratios do not vary with the variations in body weight. In five cases the dorsal root fibers of the Vth nerve are more than twice the number in the ventral root. In the Vlth there are but two such cases and in the IXth there are none, but in the latter the relation is much more uniform than in either of the other two. The average ratios, computed for each of the nerves, show more readily the differences between the excesses of dorsal root fibers in the Vth and Vlth nerves and the excesses in the IXth. Very similar ratios may be obtained from the numbers pertaining to these three nerves determined in the previous investigation ('99). Both the IXth nerves recorded there show the dorsal roots as having 1.5 times as many fibers as the ventral roots, thus agreeing with the average found here. There it was shown that of all the spinal nerves of the frog, only the 1st, Illth and Xth possess an excess of fibers in their ventral roots. That the excess of dorsal root fibers is greater in the Vlth and especially in the Vth than in the IXth nerve coincides with the fact that a greater proportion of their fibers are contributed to the skin than is the case in the IXth. They are much smaller than the IXth and in many cases the number of fibers in their dorsal branches exceeds the number in the nerve trunk (see Table III), and so great a proportion of fibers in the dorsal branches are sensory that the peripheral border of the spinal ganglion may nearly

28 Journal of Comparative Neurology and Psychology.

always be seen to protrude along the dorral branches instead of upon the trunk.

For the entire 21 nerves included here the general average ratio of meduUated fibers in the ventral roots to those in the dorsal roots is 1:1.8. Birge's work with Rana csculenta (^'62) gives the similar ratios of 1:1.7 and 1:1.8, and the author's previous work gives for the same three nerves an average ratio of 1:1.6, and for the Vlth nerve alone of 19 specimens, a ratio of 1:1.8. This similarity in the proportion of fibers in the motor and sensory roots indicates that whatever other changes may take place, a more or less fixed numerical relation is maintained becween the two.

For all of the spinal nerves of one side of the frog, Birge's figures give a ratio of ventral to dorsal root fibers of 1:1.2. Hatai ('02), investigating three of the spinal nerves of the white rat, finds in every case more fibers in the dorsal than in the ventral root. The excesses varied from 2 to 2.9 times the ventral root fibers and gave an average ratio of 1:2. 5. Ingbert ('04) has recently completed the enumeration of the fibers in the dorsal and ventral roots of all the spinal nerves of one side of an adult man and it is interesting to note that the ratio found by him is considerably higher than that found in either the frog or the rat. He finds in the sums of all the roots of one side that the ventral root fibers are exceeded by those in the dorsal roots in the ratio of 1:3.2. In comparing the different regions of the spinal cord, he further shows that the excess of dorsal root fibers runs appreciably higher in most of the cervical and lumbosacral nerves than in any of the thoracic region. Hatm's figures indicate this to be true for the white rat also, and to some extent the figures available show it true for the frog. It is explained as entailed in the requirements for the inervation of the relatively greater skin area of the upper and lower limbs.

Hatai, dealing with rats of different ages, found further that the excess of dorsal root fibers was greater in the young than in the adult, and thus with reason claims that the ventral root fibers increase in number more rapidly between the periods of medullation and maturity than do the dorsal root fibers. If

Hardesty, spinal Ganglion Cells. 29

the specimens of Table I be separated into two groups, letting the first three comprise a group of younger against the last four as a group of older specimens, and the sums of their root fibers compared, it will be found in each of the three nerves that the excess of dorsal root fibers over those of the ventral root is greater in the younger than in the older specimens. The differences in the ratios in favor of the younger, however, are not so marked as Hatai finds in the white rat. This is perhaps due to the fact that the frog has a relatively longer period of general growth than the white rat and, while its ventral root fibers do increase more rapidly than those in the dorsal root, the whole process of growth is slower. This point will be approached from another direction in the discussion of Table IV.

In every case Table I shows that there is a considerable excess of fibers in the sum of the trunk and dorsal branches above the sum of fibers in the ventral and dorsal roots (col. I). As compared with the sum of fibers on the central side of the spinal ganglion, this excess will be referred to as the "distal excess." It will be considered in more detail under a separate heading and with a differently arranged table, but since Table I better allows a comparison of the three nerves with each other it may be noted here (i) that in each of three nerves the distal excess shows a decided, though irregular, increase with the increase in body weight ; (2) that, though it is evidently dependent to a certain extent upon the amount of the dorsal branches, in neither of the nerves does it show marked coincident variation with the number of fibers in the dorsal branches (col. H); (3) that while the percentage of the distal excess (col. J) varies to a considerable exent with the proportional relation of the dorsal branches to the nerve trunk, the actual amount of the distal excess does not do so ; (4) the average amount of the distal excess is approximately similar in each of the three nerves and the same is true for the average amount of the dorsal branches in each of the nerves ; (5) the percentage of the distal excess ranges somewhat higher in the Vth than in the Vlth nerve and is much higher in both than it is in the IXth nerve.

30 Journal of Comparative Neurology and Psychology.

IV. The number of spinal ganglion cells and their relation to the number of fic>~i'e fibers connected zvith the ganglion.

It is seen in column A and B. Table I, that the numbers of ganglion cells vary much as do the numbers of fibers in the dorsal roots. This is to be expected. But it is also seen that there are many more cell-bodies in the ganglia than there are fibers in the corresponding dorsal roots. The counts show a general average of three times as many cells in the ganglia as fibers in the dorsal roots. The averages obtained for each of the three nerves are quite similar, being somewhat greater for the Vth and Vlth than for the IXth. The ratios (col. C) show no tendency in either of the nerves to vary with the weight of the animal.

This excess of cells in the spinal ganglion above the number of fibers in the dorsal root agrees fairly well with the results obtained by Hodge ('88) whose paper, as far as I am aware, is the only other dealing with an American species of frog. He counted the ganglion cells and dorsal root fibers in the VHth, Vnithand IXth nerves of a single specimen and obtained ratios of cells to fibers of 1:2.5, 1:2.9, ^-S-S respectively, giving an average of 2.9 cells per each dorsal root fiber. Buhlek ('98) investigating the spinal ganglia of Rana esculenta, counted the cells and dorsal root fibers of one of the IXth nerves" of one specimen and found 5 cells per fiber. This is considerably higher than any of my results. It may have been peculiar to the one nerve investigated or it may be peculiar to the species. For the mammals, Gaule and Lewin ('96) found in the XXXIIth nerve of the rabbit 6.4 spinal ganglion cells per fiber in the dorsal root of the nerve, and Hatai ('02) for three of the spinal nerves of each of four white rats of different ages, obtained excesses of cells ranging from 11.5 in the younger down to 2.7 per fiber in the adult, always finding twice as many cells as dorsal root fibers. On the other hand, Freud ('78) working with the less perfect technique of his time, concluded that in Petromyzon the number of spinal ganglion cells were equal tothe number of dorsal root fibers.

Hardesty, spinal Ganglion Cells. 3 1

The distal excess realized, it might be natural to suppose that many of these extra cells are such as send processes toward the periphery without sending corresponding processes into the dorsal root. After subtracting the motor or ventral root fibers (col. F) from the sum found on the peripheral side of the ganglion (col. G), it is seen that there is still an average of more than twice as many cells as there are fibers which can possibly be considered as arising in the spinal ganglia. These ratios (col. D) vary quite closely with the ratios having to do with dorsal root fibers (col. C). By comparing columns C and D it appears that there is in the Vth and Vlth nerves an average of about one cell less per distal fiber than there were cells per fiber in the dorsal roots, while in the IXth nerves this average is only 0.3 cells less. The Vth and Vlth are the nerves which have relatively the larger amount of distal excess and absolutely the larger percentage of distal excess.

Many of these extra cells are allotted to the category of the "DoGiEL spinal ganglion cell of Type II." However, none of the descriptions of this type indicate that they are sufficiently abundant in the ganglia to account for more than a very small proportion of the excesses of cells found there. A larger proportion of these extra cells belong probably to the anaxonic type of neurone, latent. cells which have not yet developed processes. BiJHLER and Hatai suggest that this type consists among the smallest cells of the ganglia. In the nerves of the frog here investigated, the proportion of small to large cells is greater in the ganglia of the Vth and Vlth nerve than it is in the IXth. This, however, is thought explainable on other grounds.

That the excess of cells above the number of dorsal root fibers is greater than the excess above the fibers on the distal side of the ganglion is no doubt due to some extent to certain of the so-called multipolar cells described in the spinal ganglia of various animals by Kolliker ('93), Lenhossek ('94), SpirLAS ('96), DoGiEL ('97), Cajal and Oloriz ('98) and others. DoGiEL thinks that some of these may be modifications of his cell of Type II. Others are thought to be of the character of

32 Journal of Comparative Neurology and Psychology.

sympathetic neurones and to send processes toward the periphery without sending a corresdonding process into the dorsal root. Another cause for the excess of cells above the fibers being less on the distal side of the spinal ganglion, is due to the bifurcation of fibers on the distal side. Both this and the multipolar cells of the character just mentioned contribute to the formation of the excess of fibers on the distal side of the ganglion and both will be considered in the discussion of the "distal excess."

On neither side of the spinal ganglion does the number of cells per each fiber supposedly connected with the ganglion show a regular decrease with the increase in the body weight of the specimens. Such a decrease might be expected with the usual conception that the number of ganglion cells becomes fixed at quite an early period in the growth of the animal, but my counts for neither of the three spinal nerves here represented show such with any regularity. Hatai ('02) found in the rat an appreciable decrease in the value of the ratios of cells to dorsal root fibers as the animal reached maturity. Either in this respect the processes of growth in the frog are again different, or the process is so slow in the frog as to require the investigation of a greater number of nerves of a greater number of individals. Could the cells be counted at different periods of the growth of the same frog, the result might be different, for it must be remembered that each set of figures in Table I but represents the conditions in 'the nervous system of a different individual at a certain time.

The growth relations are shown in another way in Table II which is constructed with reference to the body weights instead of grouping with reference to the number of the spinal nerve as in the table above. Such results as may be brought out by arranging the frogs into two groups, one containing the larger and the other the smaller frogs, and summating the numbers in each will be presented later.

Hardesty, spinal Ganglion Cells.













bjo 'S

pa u)

i3 >

C C/3

No. of cells in spinal ganglion of each nerve.

Sums of cells in spinal ganglia.

.S '^ t/i

S u

No. of cells per fiber in dorsal roots.

No. of cells per fibe in trunks and dorst branches less ventrt roots.

Sums of fibers in trunks and dorsal branches less ventral roots.

No. of fibers in trunk and dorsal branches less ventral root.

No. of fibers in dorsal root of each nerve.





7.0 j





9th =;th








149 1




10.4 ^





9th .Sth

458 1





























33-7 ]





9th ■;th












47-4 ]





9th Sth












61.8 j


















63-4 ]









3.2 - 3.6

2.5 2.4






Table II. Allow^ing a comparison of the number of cells found in the spinal ganglia (col. B), the number of fibers in the dorsal roots (col. I) and of the number of fibers found in the trunks and dorsal branches (col. H) of the Vth, Vlth and IXth spinal nerves of each specimen. The different specimens are entered in the order of their body weight. In columns C, D and G the sums of the ganglion cells, the sums of the dorsal root fibers and the sums of the fibers in the trunks and dorsal branches found in the three nerves of each frog may be compared with each other and, reading downward, the sums in each column may be compared with reference to the increase in body weight. Column E, obtained by dividing col. C by col. D, presents the ratios of ganglion cells to dorsal root fibers in the three nerves combined of each specimen and likewise col. F shows the number of cells in the combined ganglia per each fiber in the total of the trunks and dorsal branches less the ventral root fibers of each specimen.

34 Journal of Comparative Neurology and Psychology.

Table II shows (col. E) that when the numbers found in the three nerves of each specimen are combined, the ratios of the sum of the ganglion cells to the sum of the dorsal root fibers show even a less tendency to decrease with increase in the body weight than is manifested when each of the nerves is considered separately as in the previous table. Indeed, two of the three cases in which the sum ratio is highest are in the two largest specimens. • The sum ratios of cells to fibers on the distal side of the ganglion (col. F) follow so closely the variations of the ratios dealing with dorsal root fibers that whatever is true for the one is practically true for the other.

By comparing columns E and F it is seen that in the sums of the cells and fibers in the three nerves of the several frogs, the number of cells per fiber on the distal side of the ganglion is only from 0.2 to 0.7 of a cell less than on the central side. The cases in which the difference is greatest are in the largest specimens. The average difference is 0.6. In was seen in Table I that when the Vth nerves alone are considered this difference averages i.i cells, being less in the Vlth and much less in the IXth nerves.

In columns B and H, Table II, better than Table I, is also illustrated the fact that the relative number of cells and fibers apportioned to a given nerve varies greatly. To the unaided eye the Vth nerve usually appears smaller than the Vlth, but in three of the seven frogs here it contains appreciably more cells and fibers than the Vlth. This variation from the usual proportion is more frequent in the IXth nerve but its variation is less manifest here because of the absence of figures for the nerves adjacent to it. The relatively large sums for the 7 gram frog are due to its having comparatively a much larger IXth nerve than usual for its weight. Also the IXth m the 25 gram frog seems to have been small. In column I it is seen that the numbers of dorsal root fibers vary much as the ganglion cells and as the fibers on the distal side of the fjang-lion.

Hardesty, spinal Ganglion Cells. 35

V. Further observations on the conditions eietermining the excess of fibers on the distal side of the spinal ganglion.

The distal excess was described in the previous papers and some attention was given to probable arrangements of the fibers in producing it. It was shown in a series involving all the spinal nerves of one side of a frog, (i) that of all the nerves, those which in proportion to their trunks have the largest number of fibers in their dorsal branches have also the highest percentage of distal excess ; (2) that in some cas&s the amount of the distal excess may even exceed the number of fibers in the dorsal branches ; (3) that while in a given nerve (Vlth, considfrom a series of 19 frogs) the increase in weight is not accompanied by a regular increase in the value of the distal excess in this nerve, yet when the results are grouped for the small and large separately, the distal excess for this nerve shows a decided increase with the gain in weight. It was further suggested that the distal excess is due (i) to the medullated fibers from the sympathetic system which terminate in the spinal ganglion, (2) to cells in the ganglion which send processes toward the periphery but not into the dorsal root, (3) to the bifurcation of the peripheral prolongation of the spinal ganglion neurone of Type I and (4) to the bifurcation of ventral root fibers in the region of the spinal ganglion.

The observations as to the variations and causes of the distal excess made in these papers, one using the several nerves from one specimen and the other the one nerve from several specimens, should gain significance if further based upon the findings in several nerves from each of several specimens. Table III is constructed with special .reference to the distal excess and is given as indicating further some of the number relations involving it The table is compiled from the numbers in Tables I and II and represents some relations not brought out in the previous work.

Attention is called to the following relations as shown in Table III.

36 Journal of Comparative Neurology and Psychology.

1. When the different nerves of each specimen are considered indidually, the number of fibers in the nerve trunk (col. A) and those in the dorsal branches (col. B) do not show a regular increase with the increase in body weight. The same is true for the sums of these fibers in the three nerves of each specimen (col's. C and L). However, though the increase is not a regular one, the larger specimens have appreciably more fibers than the smaller ones. The same statements apply to the numbers of fibers in the dorsal and ventral roots (col's. D and E).

2. It is evident here (col's. C and D), and especially so when computed (see Table IV), that in the dorsal and ventral roots the increase of fibers with the increase in weight is not so great as in the trunks and dorsal branches. In other words, the fibers on the side of the distal excess increase more rapidly than on the central side of the spinal ganglion.

3. While the amount of the distal excess in each nerve (col. F) varies so greatly that its increase with the body weight is not very evident, yet when the sums of the distal excesses in each set of nerves are considered (col. G) there is manifest a decided and nmch more regular increase.

4. The most marked increase in the sums of the distal excess appears in this series between the weights of 25 and 61 grams. Between 25 and 33 grams the increase amounts to 27% and in passing from 33 to 47 grams the increase amounts to 23%, while the average increase from from one specimen to the

Table III. Giving number relations involving the distal excess. The ▼arious columns are either rearranged or derived from the columns of Tables I and II. The sums in columns C, D and G result from adding in adjacent columns the numbers found respectively in the three nerves of each specimen. The percentages in col. J result from dividing the sums in col. G by the respective sums in col. D of Table II ; the percentages in col. K from dividing col. G by the respective sums of ventral root fibers in col. C of Table IV. The sums in column P are obtained by substracting col. D of Table II from col. C of that table and the ratios in the last column result from dividing the numbers in col. P of this table by the sums in col. G. As indicated in the headings, all the other columns set in heavier type are derived from columns accompanying them in this table. The percentages are reduced to one decimal figure and the ratios are given in round numbers.

Hardesty, spinal Ganglion Cells.







^s-B- S-S-ff- 3-5-S' ?S-S-S'S'9' S-cr-pT- crrs-rj



o to -f^ -^ <.n OJ


t>J -t^ OJ 0-V-" M


^J Oj Ck) 4^ ^^ 10

\o to to O tJ to

O On O -(^ »- to OJ On O OnvO Oj


OJ - «  OJ 004^

- -P- o

u> - o

t>J Oj U) -f^ 00^

OJ Oj OJ (^ rr O

C>J M vO ^ -P " ON "

- - M l-n 4^ OJ -fi. 00 to

"- to 1- (0 - 'J

Oj O OJ -. vO — 4i. 4i. M to ^ « 

^ _(_ Ji ^ tj M C^J O UJ •- M vO O to

{^ oo <:> o o o^ ^I vo -F». o -^^ oj v; VI

U> OJ l-n OJ M 10 C*J

4- 00 Oj ^J ON O O 1^ O C« U> — 00 O ^ OJ O to o <-" ^

M -t>- tn-f ^I 'O 00


OOVi ^

to O ON

^4 ^-J VO


00 -- -J


^ 4^ OJ

On O O

4^ ON to

O Oj OJ ui 4^ — to ^J "

NO Oj Oj Ln --J Ui O 00^


4^ Oj to 10 O ^J

(0 ON-'.

Ml-" »wi 00U1

- M to

00 to -^

Oj t-n OJ

M to ►

4^ ^ O

O Cn M

Oj VI NO O VI 00

to ~a --4

" J, On

_ „ Ol ^ M N to 2 S 00 ■>! (D 00 -vi Oi 2 0) (d U b) to

Oj^ (0<-" OjOJ -f^lO •-itO -tj (OM \0 00 O OnJ- 00 OC-P>- W VivO^J VI vO OO OOj^J Oi^^^J

»«OjOO^i-^»- C3^0'30CvOOC^vOa^^O Cn|0 4^ 0<-nCN




















CO 00


CO 00


lO (0






  • >











  • >

  • t



















































Body weight in grams.

Spinal nerve.

No. of fibers in trunk.


No. of fibers in dorsal branches.


Sums of fibers in trunks and dorsal branches.

Sums of fibers in dorsal and ventral roots.

No. of fibers in dorsal and ventral roots.

The distal excess in each nerve.

Sums of distal excess. O

of distal excess in „_, I"^

iJ-l > W

each nerve.

% of sums of distal ex- Cij

cess based on sums of hh i— ( roots. .►-*


relations of sums of ciistal excess to sums of dorsal root fibers.

f/f relations of sums of distal excess to sums of f1 ventral root fibers.

Sums of fibers in dorsal branches.

(■^f, relations of sums of distal excess to sums of dorsal branches.

(fc relations of sums of distal excess to sums of I^ trunks & dor. branches.

(J^f relations of sums of dorsal branches to sums O of trunks & dor. branches

Sums of ganglion cells less sums of dorsal root h^ fibers.

Ratios of sums of distal excess to sums of gang- -^ lion cells less sums of dorsal root fibers.

38 Journal of Comparative Neurology and Psychology.

next in series is only about 13%. The first two and last two specimens differ less in weight than the others. The 47 gram frog was a male and the next in series, weighing 61 grams, was female. The distal excess in the two is about the same and the 6t, gram frog, also female, shows an increase above the 47 gram male of only about 4%. In the older specimens the males are always of less weight than the females and the frog of 47 grams was no doubt correspondingly as fully developed as the females weighing 13 and 16 grams more. This practical cessation of progressive changes in the relations of the distal excess at 47 grams is further indicated in some of the succeeding columns of the table.

5. The percentages of the distal excess in the individual nerves (col. H) are higher in the Vth and Vlth nerves than in the IXth, as was noted in Table I, and they also show a decided though irregular increase in value with the increase in body weight. When again the percentage of the sums of the distal excess based upon the sums of the fibers in the two roots of each set of nerves is considered (col. I) it is seen that the value increases gradually, and much more regularly, with the increase in weight. From this it is evident that not ojily does the general distal excess of fibers increase with the growth of the animal, but that it increases at a more rapid rate than the number of fibers in dorsal and ventral roots. This suggests the question as to which of the two roots undergoes growth changes more nearly corresponding to those manifested by the distal excess. In column J the relation of the sums of the distal excess to the sums of the dorsal root fibers in each set of nerves is expressed in the form of percentages and in column K the same is done tor the ventral root fibers. As shown in Table I, the number of dorsal root fibers is always greater than the ventral root fibers and so the sum of the distal excess in the smallest specimen is 14.4% of the sum of the dorsal root fibers and 24.4% of the ventral root fibers. The percentages of the ventral root vary in a more regular and constant progression than those of the dorsal root. From this it may be assumed that the changes in the number of the ventral

Hardesty, spinal Ganglion Cells. 39

root fibers more nearly keep pace with the changes in the distal excess. It is shown in Table IV that the ventral root gains fibers somewhat more rapidly than the dorsal root. The percentage variations in the ventral roots of the individual nerves cannot be shown so well in a single column as the sum relations, for, in 6 out of the 14 smaller nerves (the nerves which possess relatively the largest distal excess), the distal excess exceeds the number of fibers in the ventral root.

6. The variations in the distal excess do not closely coincide with the variations in the number of fibers in the trunks and dorsal branches (col's. F, A and B and col's. G and C). Yet when the percentage relations of the sums are considered (col. N) there is again evident a fairly regular and progressive increase. This indicates further that the fibers forming the distal excess increase at a proportionately more rapid rate than the sums of the trunks and dorsal branches in which the excess is contained. The excessive increase is similar to that relating to the two roots though less rapid and somewhat more irregular. In Table IV it is better shown that as the animal grows it gains fibers more rapidly in the trunks and dorsal branches than in the nerve roots. It is indicated here that some of this more rapid gain must take part in forming the distal excess.

7. As pointed out in the previous papers, the value of the distal excess is dependent upon or somehow correlated with the relative amount of the dorsal branches. Those nerves which have a proportionately large number of fibers in their dorsal branches as compared with the number in the nerve trunk always have a high percentage of distal excess. Most of the smaller nerves of the frog are of this type. Of the 14 smaller (Vth and Vlth) nerves employed here, 6 have dorsal branches containing even more fibers than the trunks themselves (compare col's. A, B and H). All of them have higher percentages of distal excess than the IXth nerve. While in the individual nerves (col. B) as well as in the sums (col. L) it is seen that dorsal branches increase with the body weight, the variations do not appear so regular or to closely coincide with the variations in the amount of the distal excess (col's. F and G). A

40 Journal of Comparative Neurology and Psychology.

comparison of their numbers alone shows that the dorsal branches contain many fibers which have nothing to do with forming the distal excess for in every case here the dorsal branches themselves contain at least a few more fibers than comprise the distal excess. However, in the nerves formerly dealt with it was found possible in some of the larger nerves, for the distal excess to exceed the amount of the dorsal branches thus making it necessary, in those cases at least, that some of the distal excess be contained in the trunk. The percentage relations of the amount of the distal excess to the amount of the dorsal branches are shown in column M. In this it is seen again that the distal excess increases at a more rapid rate than the number of fibers in the dorsal branches, though the progression is not so constant as in some of the other cases. If the fibers of the dorsal branches alone are concerned to any great extent in forming the distal excess, the relation of their number to the number composing both the trunk and dorsal branches should be somewhat similar to the same relation maintained by the amount of the distal excess. In column O such relations of the dorsal branches are expressed in percentages and when they are compared with the same values for the distal excess (col. N) the variations of the two are not very similar. Rather, a more fixed proportion is manifest, the dorsal branches comprising approximately 19% of the fibers on the distal side of the ganglion up to 47 grams at which specimen the proportion suddenly changes to 24%.

8. Certain of the excess of cells in the spinal ganglion above the number of fibers in the dorsal root may take part in contributing to the distal excess. In column P are given the numbers of cells in the spinal ganglia in excess of the fibers in the respective dorsal roots and in the last column of the table the relations of the fibers of the distal excess to these extra cells are expressed in the form of ratios. Though it appears that the cells themselves increase with the increase in weight, the increase of the fibers in the distal excess is such that while in the youngest specimens there are twelve times as many cells as fibers, in the 47 gram frog there are only six times as many.

Hardesty, spinal Ganglion Cells. 41

This shows it possible that some of these extra ganglion cells may send processes toward the periphery alone and thus contribute to the distal excess. It is further seen in the columns that if the cells of the ganglion increase in number, their increase does not keep pace with that of the distal excess, nor does the number of cells in excess of those giving origin to dorsal root fibers bear a constant relation to the fibers forming the distal excess.

As to tlie origin of the distal excess, it is indicated in the number relations noted above that it can hardly be due to any one cause. To explain its existence little can be added in this paper to that which was advanced in the previous papers. There are very {qw publications which even consider its presence, but certain of the findings of several investigators upon the structure of the spinal ganglia may be interpreted toward its explanation.

One of the first explanations which suggests itself is that the excess may be due to the splitting of fibers on the distal side of the spinal ganglion. This was one of the first evidences sought for by the author. Both methylen blue intra vitam, and gold chloride were employed in the preparations and serial sections as well as teasing were resorted to. The Vth and Vlth nerves were chiefly used for this because they are the smaller of the nerves used in this paper and because the smaller nerves have the higher percentage of distal excess.

A bifurcation of the fibers on the peripheral side of the ganglion was observed quite often but by no means was it observed in sufficient frequency to at all explain the entire origin of the distal excess. To give a distal excess of even 10%, every tenth root fiber must bifurcate at the peripheral border of the spinal ganglion and, while with the methods employed it was impossible to determine the exact proportion of dividing fibers, the author is convinced that it cannot occur even as frequently as this. It is seen in the tables that the smaller nerves always have a distal excess above \0% and that it may even exceed 50%. Most of the splitting observed occurred within the level of the peripheral border of the ganglion, and in order

42 Journal of Comparative Neurology and Psychology.

to correctly interpret each case, the ends of the fiber had to be followed sufficiently to preclude the possibility of the T-fibers of Ranvier or those of Dogiel cells being mistaken for the type of bifurcation sought. As the fibers from the two roots join to form the trunk and dorsal branches, they cross and intertwine to an extent which renders teasing or following a fiber in sections quite difficult. Owing to tearing and breaking and the difficulty in following a fiber in question through the maze, certain of the bifurcations observed had to be discarded as uncertain. But including the uncertain cases would not have made enough to account for the average distal excess. The great majority of the fibers on the distal side of the ganglion show no branching at all.

For the same reasons I was especially unable to determine the proportion of the dividing fibers belonging to each root separately. Bifurcations of ventral root fibers are somewhat more easily distinguished than those ot fibers arising in the spinal ganglion. Of them it can only be said that they are certainly not frequent enough to give percentage values similar to those in column K, Table III. In most of the cases a ventral root fiber was seen to bifurcate, sending one product of the division to the dorsal branches while the other remained in the nerve trunk. This is of interest physiologically for the muscle supplied by the dorsal branches and by the trunk are usually considered independent. It is not an unheard of complication however. Cajal ('99) pictures such an occurrence in the chick and discusses it physiologically, and Dunn ('02) finds frequent divisions of fibers in the sciatic of the frog, one product going to a branch supplying one muscle or set of muscles and the other going on in the trunk to be distributed to other muscles entirely distinct. Most of the fibers going from the ventral root to the dorsal branches do not bifurcate. Occasionally a fiber divides after entering the dorsal branches, its parts going to different divisions of the branches. If such a splitting occurs near enough to the spinal ganglion to be included in the section used in the counting, it of course would contribute to the distal excess.

Hard EST Y, Spinal Ganglion Cells. 43

Some years ago Stannius ('49) and Freud ('78) observed splitting of the peripheral process of the spinal ganglion cell, the former in fishes and the latter in Petromyzon. Dogiel ('96) describes the same for mammals and Buhler ('98) suggests it as an explanation of the distal excess found by him in the frog. It cannot be of very frequent occurrence in the trog. From my own observation I am led to believe that a splitting of the peripheral process of the dorsal root ^neurone is less frequent than in the ventral root fibers. Of the different frog's nerves examined here only three cases of the bifurcation of this process was observed. In these one product went to the dorsal branches and the other continued in the nerve trunk. Dogiel (p. 148) pictures this arrangement occurring in mammals. Dogiel ('97) also finds division of axones in the dorsal root or on the central side of the spinal ganglion.

It may be further added in explaining the distal excess that if it were mostly due to the splitting of fibers representing the two roots, then for it to obtain the rate of increase manifested, these fibers would have to either divide with increasing frequency as age advances or they would have to give off branches (divide) after they have grown into the nerve and become meduUated. The latter at least is hardly probable, though it might be urged in support of it that the bifurcation always occurs at a node in the sheath. I know of no observations directly supporting either idea.

The greater source of the distal excess lies perhaps in the presence of fibers connected with the spinal ganglion but not represented in the nerve roots at all. Such fibei-s are of sympathetic origin and have been repeatedly described (Cajal, '93. HuBER, '94, Dogiel, '96a, Cajal, '99 and others) as entering the spinal ganglion and breaking up into numerous twigs which terminate in telodendria about the cells there, mostly the Dogiel cells of Type II. Many of these sympathetic fibers are described as medullated. In my preparations of the frog, by comparing sections of the ramus communicans with sections of the trunk and dorsal branches, there may be seen in the latter numerous medullated fibers similar to those considered of

44 Journal of Comparative Neurology and Psychology.

sympathetic character in the ramus. They belong to the smaller type of fibers. The larger of them have a medullary sheath which is relatively thinner than that of fibers of undoubted spinal origin and which stains less black with osmic acid and shows a tendency to collapse in the sections. The smallest of the medullated fibers in the rami, however, cannot be distingushed from the smallest of the fibers known to arise in the spinal cord and spinal ganglia. Otherwise differential counts could be made to determine their exact proportion in the trunk. As is well known, the rami contain fibers from both the ventral roots and spinal ganglia. The larger of these may be distinguished by the character of their sheaths.

Fibers considered of sympathetic character were always observed in the dorsal branches. When the dorsal branches are much divided, often a small twig may be seen with the majority of the fibers in it of this type. This suggested that fibers from the sympathetic ganglia may enter the nerve trunk by way of the ramus, traverse it to the peripheral border of the spinal ganglion and there pass into the dorsal branches without connecting with the spinal ganglion. Such fibers would of course be counted twice, once in the trunk and once in the dorsal branches and thus contribute to the distal excess. A special search was made for such fibers and none were found which could be so construed with certainty. If any exist they must be very few and it was assumed that the distal excess cannot be very materially affected by them. I think it necessary to explain at least most of the sympathetic fibers in the dorsal branches in some other way, and suggest that certain of the so-called multipolar cells in the spinal ganglia have to do with them. There are numerous observations (cited above) noting the presence of these cells and many ascribed to them a sympathetic character — cells left over in the spinal ganglia during the period of the offshoot of the anlage of the sympathetic. Telodendria of the centripetal sympathetic fibers are described as terminating about these cells and the role presumed for them here is that they are merely interposed in a sympathetic chain of neurones and that the fibers given off by them pass by way of the dorsal

Hardesty, spinal Ganglion Cells. 45

branches to the blood vessels, etc., proximal to the vertebral column/ Such an arrangement would give two extra fibers on the peripheral side of the spinal ganglion. It should be mentioned, however, that Dogiel ('97) who has made a study of these cells, was unable to trace any of their processes beyond the confines of the spinal ganglion. He considers them either sympathetic or modifications of his spinal ganglion cell of Type II.

Finally let it be added that early in this investigation counts were begun of the medullated fibers in the rami communicantes with the hope of gaining some clue as to the proportional part they play in forming the distal excess. It was soon found that the number of fibers in the distal excess of a nerve is often greater than the number of medullated fibers in its ramus and since many of the fibers of the ramus are of undoubted spinal origin and many of the smallest of uncertain origin, the research was discarded as unprofitable in that the ramus cannot account for all of the excess and in that it was impossible to determine the exact proportion it does contribute. So far the conclusions must be general, namely, that the distal excess is due to several causes, though probably the greater amount of it is due to medullated sympathetic fibers connected with the spinal ganglion but which are not continued into the nerve roots.

VI. The gain of cells and fibers zvitJi the gain of zveight.

BiRGE ('82) determined that the frog while increasing from 1.5 grams to 1 1 1 grams gained in the ventral roots of its entire spinal nerve 5 i fibers for each gram of weight gained. This estimation was based upon counts of all the ventral roots of one side of 6 specimens. He also counted the dorsal root fibers of one side of two frogs, one weighing 23 and the other 63 grams. Computations based upon the numbers he obtained give for all the nerves of both sides a gain of j'] dorsal root fibers per gram of weight gained. His data are not sufficient upon which to base estimations of the gain of fibers on the distal side of the spinal ganglion. The author ('00) dealing with the Vlth spinal nerve alone of frogs varying from 5 to 79 grams, ob

46 Journal of Comparative Neurology and Psychology.

tained numbers indicating a rate of gain in this one pair of nerves of about 3 ventral root fibers, t^ dorsal root fibers, and

TABLE IV. (Frog)









0) OJ






■^ ^

S n-)

. . 1





rsin ches rves




of eel! anglia nerves

4J j£




^ 2


^ c

a; H i"

^ s

> ,„ m

y: c ^

of fi s a lies

of fi

1 bl


sof iber fiber


en t/) <D

■-n ?i

f. t


<Ji <A'-\


Sum trun bran


Sum dors the



Rati root root

33-7 47-4 61.8


6189 5201 7328 6771


1933 2294 2141

1416 1647

1509 1283

4205 4166 4384 4039

S08 1022 1049


^•7 2.7

3-2 3-2

1:1.6 1.2 1-5


Totals : Averages :

206.3 51.6

7.0 10.4 25-3

25489 6372 2

5105 6244 4462

8684 21710

1888 1723 1667

5855 1463. 7

1115 962


16794 4198. 5

3275 296S 2903

3853 963. 2

620 570 534

2.q 2.9


3-6 2.7


1.7 1.8 1.8

Totals : Averages :

42.7 14.2

15811 5270.3

s278 1759. 3

3020 1006. 6

9146 3048. 6

1724 5746

3-0 3.0


Differences of averages :






1149. 9

388. 6



Gained per gram of wt. gained :


20.9 fc



307 37.7C/C

% of gain :



Table IV. The base numbers are taken from Tables II and III. The specimens are entered in two groups, one of the four larger frogs, the other of the three smaller. Opposite the individual body weights are the sums of the ganglion cells, and of the nerve fibers in the localities indicated ot the three nerves. The table allows a comparison of the number of cells the average specimen of each group would possess in the spinal ganglia of the three nerves (col. A), and of the number of fibers in the dorsal roots (col. B), ventral roots (col. C), the trunks and dorsal branches combined (col. D), the number of fibers in the dorsal branches alone (col. E), and finallj- the ratios of the ganglion cells to the dorsal root fibers (col. F) and the ratios of the fibers in the ventral roots to the fibers in the dorsal root (col. G). In the last two lines across the table are given (i) the estimated number of cells and of fibers in the given parts of the nerve gained per gram of weight gained and (2) the gains in proportion to the numbers contained in each is expressed in percentages.

Hardesty, spinal Ganglion Cells.


lO fibers in the trunk and dorsal branches for each gram gained in weight. This estimation was based on upon counts from 12 specimens.

TABLE V. (White Rat)

.- o

1) ;z;

O rt C

S csi rj


a i2 Si

t« 1)

0) .4::


cv; .0 p

69.0 167.0

28897 29048

6420 7393

2567 3115

4-5 3-9



Totals : 236.0 Averages: 118.0



57945 28972. 5

26153 25061

'38 > 3 6906. 5

33^8 4343

S682 2841.0

"77 2139

4.2 4.2

7-9 5-8


2.8 2.0

Totals : 34.8 Averages: 17.4

5 1514 25757.0

7671 3835. 5


6.7 6.7


Difference of

averages: 100.6




Gamea per gram of wt. gained : 1




cji of gain


80. 1'^


Table V. Compiled from figures given by Hatai ('02) Table VII and ('03) Table III for the Vlth cervical, IVth thoracic and Ilnd lumbar nerves of four white rats of different ages. Constructed in the same way and showing the same relations as Table IV. (Note: In the two papers of Hatai there are some slight discrepancies in the body weights. For example, in one paper the results from a 68.5 gram rat are given in the other paper as obtained from one of 69 grams, and a specimen of 24.5 grams in one paper is given 25.4 in the other. These are considered oversights or misprints and of little importance anyway. The only specimen mentioned in the second paper as additional had a weight of 264.3 grams and it is not included here since its ganglion cells were not counted.)

These indications of the rate of gain of the nerve fibers with growth suggest similar estimations for both the nerve fibers and the ganglion cells of the nerves dealt with in this paper. The results of such estimations and the method by which they are made are shown in Table IV. And in order to compare the conditions

48 Journal of Comparative Neurology and Psychology.

in the nerves of the frog with those in a mammal, the figures obtained by Hatai ('02 and '03) for the white rat are compiled and arranged in the same way in the accompanying Table V. Hatai also used but three nerves. So far as I know his results are the only available for such a comparison. In both tables the numbers apply to the nerves of one side of the body only and therefore must be doubled for the approximate numbers for both sides.

In the first place, it is further shown in Table IV that the average younger frog possesses a good many more cells than fibers in proportion to its weight than the older. It was shown in the previous papers that the younger gain fibers at a more rapid rate than the older. This is not so evident in the three nervgs used here from the fact that the IXth nerves of the younger contain considerably above the usual relative proportion of fibers and thus produce larger sums for the three nerves. In the rat Hatai finds that the gain in the younger is considerably more rapid than in the older.

It is also shown in Table IV (col's. B, C, D and E) that as the frog increases in weight, for each gram of weight gained, there is an apparent gain of 1 1 dorsal root fibers, i 2. 2 ventral root fibers, 30.7 fibers in the trunks and dorsal branches and 10.4 fibers in the dorsal branches alone. Compared with results obtained in a similar way from Birge's figures, such gains in only three of the nerves of one side appear rather high. However, as was previously seen ('00), Rana virescens seems to possess an appreciably greater number of fibers in its spinal nerves in proportion to body weight than does the European Rana esailenta employed by Birge and therefore to maintain this, its absolute gain of fibers per gram of weight gained must be greater.

It is also evident that apparently the ventral root gains fibers at a somewhat more rapid rate than the dorsal root, the relations being 1 2. 2 to 11 per gram gained. In the previous paper, in the Vlth nerve alone, the advantage seemed to be with the dorsal root, the ventral root gaining but 1.4 fibers to 2.4 gained by the dorsal root. Computations from the figures available

Hardesty, spinal Ganglion Cells. 49

for the purpose in Birge's records give for all of the spinal nerves also a higher gain in the dorsal root. And Hatai's counts ('03) for the rat, when tabulated in a similar way (Table V), show a considerably greater absolute gain of fibers in the dorsal root per gram of weight gained. The dorsal roots of all the spinal nerves here used of the frog and all of those used by Hatai of the rat contain many more fibers than the ventral roots (col. L, Table I and col's. G and E, Tables IV and V). One would therefore expect the absolute gain to be greater in the dorsal root. But the gains in proportion to the number of fibers contained in each may show differently. It is seen in column C that the three ventral roots of the average frog of 14.2 grams gain about 45% in fibers with the increase to the average weight of 51.6 grams while the dorsal roots (col. B) gained but 23 ^. By comparing percentages of gain in this way it will be found that the figures for the Vlth nerve alone in the paper above referred to also give a somewhat greater gain in the ventral root. By grouping Hatai's figures in the same way (Table V) the dorsal roots appear to gain 30.5 fibers per gram of weight gained, while the ventral roots gain only 11.7 fibers and, in proportion to the fibers contained in them, the dorsal roots gain 80% and the ventral roots 71 % — a group result still in favor of the dorsal root. However, Hatai makes it one of his conclusions that the increase of meduUated fibers in the ventral root is more rapid than in the dorsal root. He reaches this not by grouping as above, but by simply comparing the relations of the nerve roots of the youngest rat with those of the adult individual. In column E of Table V it is seen that in the 10.3 gram rat, for each fiber in the ventral root there are 2. 8 fibers in the dorsal root, while in the 167 gram rat the ratio is 1:2.4, thus showing that, as far as these two individuals are concerned, the increase of ventral root fibers must gain on that of the dorsal root fibers during the growth of the animal. When, on the other hand, the ratios of the groups are considered it is evident that in the younger the average ratio of ventral root to dorsal root fibers is 1:2.3 while in the older it is 1:2.4, thus indicating a slightly more rapid increase in the dorsal roots. The same ratios

50 Journal of Comparative Neurology and Psychology.

in the frog (col. G, Table IV) do not vary as much as they do in the rat. In fact, the extremes, the specimen of 7 grams and that of 63 grams, show similar ratios between their ventral and dorsal root fibers. The group or average ratios, however, indicate the more rapid increase of ventral root fibers mentioned above.

Table IV further corroborates a relation indicated in the foregoing tables. Column D indicates that in the sums of the trunks and dorsal branches combined there is a gain of 30.7 fibers for each gram gained in weight. This is a greater gain than in either root or in both roots combined. In proportion to the number of fibers contained in them, the gain in the trunks and dorsal branches is only 37.7%, or less than the percentage gain in the ventral roots. But the percentage rate is made up in the dorsal branches considered separately (col. E). Here the gain is about 59%. This again suggests that some correlation exists between the dorsal branches and the distal excess, for as shown, the distal excess increases at a more rapid rate than the fibers in either the dorsal root, the ventral, or in the trunk and dorsal branches.

The question of the ganglion cells is rather puzzling. The tenet more usually held is that in the vertebrates the number of nerve cells is fixed at quite an early stage in the development. This belief applied to the frog precludes an increase in the number of spinal ganglion cells even at the youngest stages employed here. Yet, when the numbers of ganglion cells found in the different specimens are considered in the same way as the fibers in the nerves which admittedly increase with growth, the cells appear to undergo, in their sum relations at least, variations somewhat similar to those of the fibers. As arranged in Table IV (col. A) it is indicated that for each gram the frog increases in weight, there is a gain of 27.8 cells in the spinal ganglia, or, as the average specimen increases from 142 grams to 51.6 grams, the cells undergo an increase of 20.9%. This percentage increase is less than that of the fibers as must be the case, but it is surprising that it appears so nearly equal to it.

It is seen (col. F) that in the younger specimens the num

Hardestv, spinal Ganglion Cells. 5 i

ber of cells per fiber in the dorsal root is greater than in the older specimens, but the difference is not so great as one would be led to expect nor do the ratios of the individual sums show a variation decreasing with the increase in weight. Rather, the relations between the number of cells and the number of dorsal root fibers seem to be maintained approximately constant in the nerves here employed. All this can hardly be due to chance variations in the fixed number of cells in the individual ganglia of the different frogs.

With Hatai's enumerations for a mammal, the results are different. Using a similar tabulation of his numbers for the three nerves of the rat (Table V) it appears that, with the increase from the average weight of 17.4 grams to that of 118 grams, the ganglion cells undergo an increase of only \2% while the dorsal root fibers increase 80% and the ratio between the two is considerably higher in the younger than in the adult. Hatai ('02) notes that in some of the nerves of the rat of 10.3 grams, the ratio of fibers to cells is as much as 1:1 1, while in the older it may be as low as 1:3. Based on this great decrease of the ratio with the progress of growth and the relatively small actual increase in the cells of the individual nerves of the different specimens, he concludes that the number of ganglion cells in the white rat remains approximately constant between the weight of 10.3 grams and the adult. He interprets the differences as probabl)- due to individual variations in the ganglia of the different specimens employed and cites the fact that the 24. 5 gram rat gives a sum of cells in the three nerves chosen which is less than that of the rat of 10.3 grams. Donaldson ('02) supports the views of Hatai and in discussing the investigation, lends observations which strengthen it.

Hatai states that the only argument in favor of an increase in the number of cells with age is the fact that of the four specimens employed the two older gave combined a greater number of cells than the two younger. Table V here given groups the older against the younger in order to compare the results with those from the frog in Table IV. The difference between the averages of the two groups of rats amounts to about 3200 cells.

52 Journal of Comparative Neurology and Psychology.

In the frog at least, that in a Hmited number of nerves an older animal may show fewer cells than one somewhat younger, can be equally well explained as due to the fact that the proportional number of neurones apportioned to a given spinal nerve is by no means constant.

The progress of growth in the frog is evidently different from that in the mammal. The 7 gram frog is no doubt relatively more developed than the 10 gram rat and in the rat of 10 grams and younger the processes of growth must go on much more rapidly. This I think is indicated in the two tables.

While the number of ganglion cells may not increase with age in the rat, the evidence though slight is, I think, a little stronger for the frog. It is needless, perhaps, to go into the literature for the support of this view for so far as I am aware, there are no observations which directly maintain it. Buhler ('98) in discussing the cells of the spinal ganglia suggests that the large cells are continually degenerating while the small cells enlarge and replace them. If this were true, the cells would have to multiply or else the number would, on the contrary, •decrease with age. Of the 18 papers that I know of, which, dealing with tissues more or less mature, describe appearances in nerve cells (vertebrate and invertebrate) thought to be concerned in the processes of cell division, none of them describe cases of undoubted nerve cells which can be confidently considered the actual process of division. Of the list Lenhossek ('95), Dehler ('95) and Buhler ('98) describe such appearances in the spinal ganglion cells of the frog.


I. Due to variations in the relative number of neurones apportioned to a given spinal nerve, the number in a given nerve may not increase regularly with the increase in body weight but the sum obtained by adding the numbers in a given nerve of the larger specimens, as well as the sum of the several nerves of the larger specimens, is always greater than that of the smaller specimens.

Hardest Y, Spinal Ganglion Cells. 53

2. In the average there are nearly twice as many nerve fibers in the dorsal root as there are in the ventral root of the Vth, Vlth and IXth nerves. The average is the highest for the Vth and in both the Vth and Vlth is higher than in the IXth nerve. This agrees with the nature of the innervation supplied by the smaller nerves.

3. As compared with the corresponding trunk, the number of nerve fibers contained in the dorsal branches is reUtively much greater m the Vth and Vlth than in the IXth nerve, while in actual amount, the average number of fibers in the dorsal branches of each of the three nerves is similar.

4. The distal excess, or the excess in the sum of the fibers of the nerve trunk and dorsal brandies above the sum of the dorsal and ventral roots, occurs to an appreciable extent in every case. While it is less in the nerves of the smaller specimens, its average amount in each of the three nerves of the several specimens is similar.

5. The percentage of the distal excess ranges higher in the Vth nerve than in the Vlth and it is much higher in both than in the IXth nerve.

6. There is a general average of three times as many cells in the spinal ganglia as there are fibers in the dorsal roots. And there are more than twice as many ganglion cells as there are fibers in the trunk and dorsal branches less the number of fibers in the ventral root. The average ratios for each of the three nerves are similar, being somewhat higher in the Vth and Vlth than in the IXth nerves. On neither side of the spinal ganglion do the ratios of cells to fibers show a regular or marked decrease with the increase in body weight.

7. The distal excess in the three nerves of each specimen increases decidedly and with considerable regularity with the increase in body weight. 'The distal excess increases at a more rapid rate than the fibers in both or in either of the nerve roots, and also at a more rapid rate than the fibers in the trunks and dorsal branches combined or in either separately. With increasing weight, the variations in the ventral roots coin

54 Journal of Comparative Neurology and Psychology.

cide more nearly with the variations in the distal excess than do the variations in the dorsal roots.

8. With the increase in weight, the fibers on the distal side of the spinal ganglion increase in number more rapidly than do the sums of the dorsal and ventral roots. In proportion to the fibers contained in them, the dorsal branches gain fibers at a more rapid rate than any other part of the nerve. This more rapid increase of the distal fibers is in some measure an expression of the more rapid rate of increase of the distal excess.

9. Supported by direct observations, the distal excess is explained as due to three causes : (a) Centripetal medullated fibers from the sympathetic system which enter the spinal ganglion to branch and terminate about the cells there ; (b) The bifurcation of ventral root fibers on the distal side of the spinal ganglion ; (c) The bifurcation of the peripheral axone of the spinal ganglion neurone. In most of the bifurcations observed, one pioduct of the division joins the dorsal branches while the other remains in the nerve trunk. The observed divisions of the fibers of the dorsal branches themselves may take part in producing the distal excess. There are some indications supporting the assumption that certain of the excess of cells in the spinal ganglion may also contribute to the distal excess by sending processes into the periphery but not toward the central system.

10. With the increase in body weight the ventral roots of the Vth, Vlth and IXth nerves gain fibers at a more rapid rate than do the dorsal roots, and at a more rapid rate than the nerve trunks considered separately.

I I. With the increase in body weight there is an apparent gain in the number of spinal ganglion cells in the three nerves. Between the average weight of 14.2 grams and that of 5 1.6 grams the ganglion cells increase about 2 i ^r , or gain at the rate of 27.8 cells per gram of weight.

Hardesty, spinal Ganglic Cells. 55


Birge, '82. Die Zahl der Nervenfasern und dc^ Motorischen Ganglienzellen im Riickenmark des Frosches. Archiv filr Mat. und Physiol., Physiol. Abthl. H.VundVII.

BUhler, '98. Untersuchungen iiber den ^^.au der Nervenzellen. Verhandlungen der physik. vied. Gesellschaft zu V crzburg, N. F., 31.

Cajal, '99. Textura del Sisteirr ^rvioso del Hombre y de los Vertebrados. Tomo I, p. 364.

Cajal, '93. Neue Darstell' >om histologischen Bau des Centralnervensys tems. Archiv filr inat. und Physiol., Anat. Abthl.

Cajal and Olo'""" -^i. Los ganglios sensitives craneales de los Mamiferos. '\ijtsta triviestral micrographica, 2.

Dale, '00. On Some Numerical Comparisons of the Centripetal and Centrifugal Medullated Nerve-fibers in the Spinal Ganglia of the Mammal. Journal of Physiology (Foster), 25, No. 3.

Dehler, '95. Beitrage ziir Kentniss von feineren Bau der syrapathischen Ganglienzellen des Frosches. Archiv fi'tr Mikr. Anat., 16.

Disse, '93. Ueber die Spinalganglien der Amphibien. Verhandlungen d. Anat. Gesellschaft auf d. Versamml. zu Gottingen.

Dogiel, '95. Zur Frage iiber den feineren Bau des sympathischen Nervensystems bei den Saugethieren. Archiv fUr mikros. Anat., 46.

Dogiel, '96. Der Bau der Spinalganglien bei Saugethieren. Anat. Anz., 12, No. 6.

Dogiel, '96a. Zwei Arten Sympathischen Nervenzellen. Anat. Anz., \\,^o.Z2.

Dogiel, '97. Zur Frage iiber den feineren Bau der Spinalganglien und deren Zellen bei Saugethieren. Internat. Monatschr. fi'tr .Anat. und Physiol., 14, H. 4 und 5.

Donaldson, '02. On the Number and Size of the Spinal Ganglion Cells and the Dorsal Root Fibers in White Rats of Different Ages. Amer. Journ. Anat., 1, No. 4.

Dunn, '00. The Number and Size of the Nerve Fibers Innervating the Skin and Muscles of the Thigh in the Frog (Rana virescens brachycephala, Cope), four. Co/nf. Neurol., 10, No. 2.

Dunn, '02. On the Number and on the Relation Between Diameter and Distribution of the Nerve Fibers Innervating the Leg of the Frog, Rana virescens brachycephala, Cope. four. Comp. Neurol., 12, No. 4.

Freud, '78. Ueber Spinalganglien und Riickenmark der Petromyzon. Sitzungsb . d. K. Akad. d. Wien, 78, Abthl. 3, Juli Heft.

Gaule and Lewin, '96. Ueber die Zahlen der Nervenfaseren und Ganglienzellen des Kaninchens, Centralblatt filr Physiologie, Heft 15 und 16.

Hardesty, '99. The Number and Arrangement of the Fibers Forming the Spinal Nerves of the Frog (Rana virescens). Jour. Comp. Neurol., 9, No. 2.

Hardesty, '00. Further Observations on the Conditions Deterrnining the Number and Arrangement of the Fibers Forming the Spinal Nerves of the Frog (Rana virescens). Jour. Comp. Neurol., 10, No. 3.

56 Journal of Coi, arative Neurology and Psychology.

Hatai, '02. Number and S. e of Spinal Ganglion Cells and Dorsal Root Fibers

in White Rats of Dift'el "nt Ages. Jour. Comp. Neurol., 12, No. 2. Hatai, '03. On the Increase i" the Number of Medullated Nerve Fibers in the.

Ventral Roots of the Spin« ' Nerves of the Growing White Rat. Jour.

Comp. Neurol., 13, No. 3. Hodge, '88. Some Effects of Electi'cally Stimulating Ganglion Cells. Amer.

Jour, of Psychology, 2. Huber, '97. Four Lectures on the Symp -^ic Nervous System. Jour. Comp.

Neurol., 7, No. 2. Ingbert, '03. An Enumeration of the Medm . " Nerve Fibers in the Dorsal

Roots of the Spinal Nerves of Man. Jour. Co) p. Neurol., 13, No. 2. Ingbert, '04. An Enumeration of the Medullated NerVt "'bf^rs in the Ventral

Roots of the Spinal Nerves of Man. /our. Comp. Netirol. ar.<-. \y'h.,

14. No. 3. Kolliker, '93. Handbuch der Gwebelehre. Bd. 2. Kolliker, '94. Der feinere Bau und die Functionen des sympathischen Nerven systems. Sitziin^sbr. d. Witrzburi^er Physik-med. Gesellschaft. Lenhossek, '94. Zur Kenntniss der Spinalganglien. Beitrige zur Histol. der

Nervensystems und der Sinnesorgane, Wiesbaden, p. 129. Lenhossek, '95. Centrosome und Sphare in den Spinal Ganglienzellen des

Frosches. Archiv filr Mikr. Anat., 46. Spinas, '96. Zur Kenntniss der Spinalganglien der Siugethieren. .4nat. Am.,

11, p. 629. Stannlus, '49. Das periphere Nervensystem der Fische. Rostock. Stienon, '80. Recherches sur la Structure des Ganglions Spineaux chez les

vert^br^s Superieurs. Annales de /' Universiti libre de Bruxelles.



In an earlier number of this fouriial the need has been urged of some category common to the neurologist and psychologist in terms of which the problems of neural structure and mental function may be discussed without immediately arousing metaphysical prejudices. Such a category is action or behavior. The former is perhaps the more abstnict concept and hence will lend itself more readily to the discussion of the philosophical questions which sooner or later are bound to arise. The latter has the advantage of being a term of popular as well as of scientific usage and is more commonly employed to describe the action of organisms. Mr. Morgan's latest book is entitled "Animal Behavior" and under this term he is successful, for the most part, in discussing the actions of organisms without prejudging the nature of the question of the psychical and its relation to the material processes. Likewise Mr. Jennings in the valuable article which appeared in the last issue of the /<9«r««/ discusses the behavior of Paramecium in terms of "action-systems" in a way which does not preempt the field for either the mechanical or teleological interpretation of the phenomena. Researches carried on in this spirit are greatly needed at the present time. Only thus is it possible to construct a platform whose planks shall consist of facts interpreted in terms of a common technique.

It is too early in the history of the movement to predict in detail the lines along which the two sciences will get together, but it is safe to say that there will have to be considerable revision of working concepts on the part of both neurology and psychology. By this is meant that the newer insight into the energic nature of matter will in time inevitably

58 Journal of Comparative Neurology and Psychology.

affect the biologist's conception of the nature of what he calls an organism. Biology, in so far as it pretends to be an exact science, regards the organism as a complicated mechanism whose elements are to be understood in terms of the physical laws which hold for these elements outside the organism. Hence, if, for example, the study of the electrical properties of matter results in transforming our chemical and physical notions, and some form of an energic is substituted for the atomic theory, this dynamic view ultimately must reach into biology with transforming effect.

In a similar way, the conception of the nature of consciousness is undergoing reconstruction in psychological science, in part due to this same energistic theory which is transforming physical science. The traditional formula which is satisfied to postulate a soul back of consciousness, just as it postulates material atoms back of motion or force, appears likely to be eclipsed by the results of inquiries which seek to discover the nature of the intimate relation which certainly seems to exist between mind and matter. There is no blinking the facts of brain structure nor of mental functioning ; the problem is to understand what we mean by each in terms of the other.

This it has been almost impossible to do in the past because of the diverse historical conditions and techniques associated with the two sciences. Biology had its roots in the natural and positive sciences ; psychology arose as a branch of philosophy and was long known as "mental philosophy." But now that the basis has been laid for a scientific psychology, there is hope of its being possible for the psychologist and neurologist to get together in their work on this common problem.

As has been intimated, this will involve a revision of psychological conceptions on many fundamental points. That this is already taking place is evident from recent tendencies in psychological thought. Consciousness is coming to be stated more and more in terms of action, in terms of the motor aspect of the organic circuit, instead of being stated exclusively in terms of the sensory aspect, which was the tendency with the older intellectualistic psychology. Great emphasis is now being placed

Editorial. 59

on the motor character of attention, on the dynamogenic nature of ideas, on ideomotor impulses, on tactile-kinaesthetic imagery, on the emotions as vestiges of motor attitudes, on the growth of voluntary movement, on the constructive and reconstructive character of thought. Says Mr. Marshall: "We are compelled to assume a unity of process in conscious life. From this point of view, the distinctions between reflex and instinctive activities and between habit and instinct are not fundamental. The sharp distinction between instinct and intelligence implies denial of the unity of consciousness" {^Mind. Vol. XI, No. 44). Professor Lloyd says: "Nothing in philosophy is so much needed at the present time as the adjustment of the science of abstract thought to the science of organic action, and every little hint as to how this adjustment can be brought about cannot but be at least a little help. The evolution of consciousness must be almost meaningless until the simplest case of accommodation as seen by the biologist is identified with the most perfect case of abstract thought that the logician knows" i^Psy. Kev., Vol. Ill, p. 426). And a recent writer has gone so far as to define perception as "an attitude toward the object perceived." He says: "Perception is an attitude toward an object as well as a complex of sensations." "All that objects mean to us is largely due to the sensations that flow backward from the bodily reverberations they excite directly in us. Perception is an attitude toward the objects perceived" (Bolton, Biological View of Perception, Psy. Rev., Vol. IX, No. 6).

How far these particular suggestions may prove fruitful in bringing about the desired synthesis is a matter of relatively little moment here. The important consideration is to note the fact of this tendency in recent literature and to keep in touch with the almost kaleidoscopic changes which are marking the progress of the comparative method as employed in this field.

The further investigations of animal reactions are carried, the more difficult appears the problem of the distribution of consciousness. But, as if to counterbalance this, the further research in comparative psychology is carried, the more is the conviction forced upon the investigator that the reactions of

6o Journal of Comparative Neurology and Psychology.

human beings (including their psychical processes, their conscious acts) will never adequately be understood until we have formulated the laws of the behavior of these simpler types of organisms.

The value of a study of the animal mind for human psychology has been emphasized by various writers. But its full significance, methodologically, has not always been realized. This deeper significance lies in the dynamic conception of consciousness, as itself a phase of the ultimate energic system, a balance or tension of forces, admitting, like all other energic phenomena, of examination, description and explanation. The conditions of consciousness as represented in the complicated structures of the brain in the higher forms are too intricate to admit of exact statement as yet in scientific terms. Hence the promising ch.iracter of researches upon the lower forms where the conditions are simpler, and where, if anywhere, the precise function of the brain as an organ for the transformation of energy can be determined. Here first may we expect the laws of equilibration or tension of energies which we call conscious to be elucidated. The solution of the deepest problems of psychology, there is good reason to believe, lies in the hands of the comparative psychologist.

ScHULTZ, in a recent article entitled "Gehirn und Seele" (^Zeitschr. f. Psy. u. Physiol, d. Sinncsorg., XXXH, Heft 3 u. 4, pp. 246-7) calls attention to the apparent dilemma in which the comparative psychologist finds himself. It is certainly a safe assumption that the higher, more complicated mental life of man and the higher animals can best be explained by a knowledge of the simpler conditions of mental life in the lower forms. On the other hand, it is a general principle that in explanation we should proceed from the known to the unknown. Now my own human individual consciousness is best known to me and most immediately given. We here seem to be under the compulsion equally of following what Professor Baldwin has called the "leveling up" and the "leveling down" methods, the mechanical and the teleological (or what some would call the anthropomorphizing) tendencies.

Editorial. 6 1

The limitations of the one method he in the incredible chasm in degree (if not in kind) which must lie between my complicated conscious life and that of the simplest organisms (if they have any at all). This would seem to check any anthropomorphizing tendency at the start. Binet's mistake, for example, lies not chiefly in his assumption that the lower organisms have consciousness (this may or may not be true), but in his uncritical use of the categories of adult human psychology in describing the reactions of micro-organisms. It is, of course, an inference that any organism besides my own has consciousness, but it is an inference, in certain cases, of extreme probability. But that perception, association, preference, choice, mean the same in these lower forms is a point to be demonstrated, not to be assumed. The great need of comparative psychology at the present time is the reduction of human conscious reactions to the lowest terms, especially as they are represented in the human infant, in the savage, and in primitive man, in order to make the comparison between human and animal behavior more direct.

On the other side, the difficulty lies in the fact that the terminology of tropisms and animal reactions has grown up almost exclusively under the domination non-psychological science, with the result that the answer to the question as to the presence of mental life in these lower forms is prejudged from the outstart. Evidently there is need of some common basis of method in biology and psychology. This is supplied, in a general way, in the conception of conscious states as themselves acts, as truly as the more obvious activities of the motor organs, but more subtle because remotely conditioned in the brain processes. One of the common problems thus, of comparative neurology and comparative psychology becomes, as has been said before, the problem of the evolution of action, and particularly the problem of the determination of the conditions of conscious action. h. heath bawden.

62 Journal of Comparative Neurology and Psychology.


The idea of appointing a special commission to advise the International Association of Academies as to the means best calculated to advance and coordinate research work on the brain originated, so far as I am aware, with the late Professor His. He formulated a somewhat ambitious scheme, the main idea of which was the foundation in each country of a central institute to, in a sense, control the research work being done in that particular country and to serve as a means of communication with similar institutes in other countries. The function of these institutes was to receive material for research, sent by people who did not particularly want it, and to distribute it to workers to whom it would prove of special value ; to receive and store specimens, photographs and other records of research, so that any worker might have the opportunity of examining the actual material upon which published memoirs were based. The author of the scheme hoped that by means of such institutes more uniformity might be introduced in the methods of research, and in the presentment of results; that the data upon which investigations were founded might be rendered more accessible than heretofore and so a common source of disagreement among workers might be removed ; and especially that valuable material might be directed into those channels where the best use might be made of it.

To discuss this proposed scheme a special commission of thirty-five members representing fourteen nationalities was appointed. It was subdivided into seven sub-sections of five members each to consider the scheme from the standpoints of (i) Human Anatomy and Anthropology, (2) Comparative Anatomy, (3) Histology, (4) Embryology, (5) Physiology, (6) Pathology and (7) Clinical Medicine.

The Commission met in London in the last week of May without its leader and prime mover, without the one man whose quiet persistency could have brought any measure of success in the realization of his scheme: Professor His died in Leipzig three weeks before the meeting.

Editorial. 63

Under these depressing circumstances only about twenty of the remaining thirty-four members of the Commission were able to attend the meeting, over which Professor Waldeyer presided. The subdivision into special subsections was abandoned and a general discussion took plaee as to the feasibility of establishing such institutes as the late Professor His had suggested. In the public discussion the chief difficulty brought forward against the realization of the scheme was financial — the need for funds to establish and maintain the institutes ; but in private conversation with the members there seemed to be a general concensus of opinion that the scheme was too Utopian ; that it was hardly likely that any considerable body of men would be so self-denying as to present their material to an institute for distribution and that the possibility of accomplishing the other objects aimed at in the general scheme seemed to be very slight. However, the members present agreed to strive to make the existing institutions in which each of them was working serve as far as possible the function of such a central institute as had been outlined in the general scheme. This platonic resolution was the only result of the general meeting of the Commission.

At the general meeting Professor Edinger remarked that this exceptional meeting of neurologists afforded an excellent opportunity to discuss certain problems of general interest, and he proposed that a special meeting be held to discuss the primary subdivision of the vertebrate cerebral hemisphere. At the special meeting, which was presided over by Professor J. N. Langlev, Professor Edinger explained that his chief reason for calling the meeting was to discuss the possibility of devising some primary subdivision of the lowlier vertebrate types of cerebral hemisphere such as I had proposed for the Mammalia. I was requested to explain to the meeting the nature of my subdivision of the mammalian hemisphere and especially the significance of the neopallium. In the discussion, which was carried on chiefly by the chairman. Professors Retzius, Edinger and the writer, it was agreed that it was not possible at present to suggest any satisfactory mode of subdivision which

64 Journal of Comparative Neurology and Psychology.

could be applied to all vertebrates, because the differentiation of structure in the higher groups rendered useless the subdivision which would apply to the lowlier groups. It was therefore decided to submit the question to further investigation and Professor Edinger invited me to prepare a report proposing a subdivision, which might be submitted to all those interested in the problem, whether members of the Commission or not, for criticism and suggestions.

The other question brought forward for discussion was the possibility of describing cerebral sulci from their relationship to areas of known physiological significance. I explained the definite relationship which the calcarine sulcus, the sulcus lunatus ("Affenspalte") and the superior and inferior occipital sulci present to the visual cortex. The essential part of the suprasylvian sulcus is a superior limiting furrow of the auditory cortex. I'he central sulcus in the Primates is a posterior limiting sulcus of the excitable or motor area, whereas the crucial sulcus of the Carnivora is an anterior sulcus of the motor cortex. In time it will probably be possible to describe all the important furrows of the hemisphere in terms of their relationship to certain definite cortical areas and so to correlate the data of morphology and physiology. The excellent researches of Dr. A. W. Campbell of Liverpool and the well-known work of Professor Flechsig are rapidly preparing the way for such an advance.

In the discussion of this matter, in which Professors HenSCHEN, Retzius, von Monakow, Edinger and Langley took part, it was agreed that it was too early to adopt the proposed method of describing sulci.

At other informal meetings various members of the Commission gave demonstrations. Professor Ramon y Cajal showed extraordinary specimens of neurofibrillae in ganglion cells stained by his new method and Professor Henschen showed many sections of the calcarine region exhibiting various forms of degeneration in the visual area.

The Commission is to meet again three years hence.



Goldstein, Kurt. Kritische und experimentelle Beitrage zur Frage nach dem Einfluss des Nervensystems auf die embryonale Entwicklung und die Regeneration. Three plates and two text-figures. Arch. f. Entwkmech.,

1904, 18, sy-i'o.

Schaper's experiment which showed motihty in a frog larva in which he had destroyed the brain and found the spinal cord in a state of disorganization, left a certain desire for more evidence. Wolff failed to obtain the same decisive result, and Moskowski actually considered the claim refuted. Goldstein, a pupil of Schaper, now supplements the first description by a drawing which is more convincing than Schaper's original one, and he adds new experimental material, which shows Wolff's error and firmly establishes very important data in harmony with Schaper's observations.

Wolff divided frog larvae of less than 5 mm. so that the dorsal part contained the entire neural tube, and failed to corroborate Schaper. Goldstein succeeded in keeping both parts alive for five days, through the use of Locke's solution, and he showed that they recover motility in two days. Hence, spontaneous and reflex motility in an early embryonic period does not depend on the existence of nerve conduction of a central organ. Moreover, the ventral piece showed further development ; notwithstanding the elimination of the neural tube it reached the size corresponding to a larva of about 6.5 to 7 .0 mm. Against these facts any arguments based on laws of regenerative processes have absolutely no weight, since we deal here merely with a primary condition of development.

Goldstein next turns against certain views of Neumann. The latter had concluded that at least for a start in the development of muscles, nervous centers were necessary ; that, once started, they would develop independently from the central organ, and, in post-embryonic life the trophic center of cord and brain would again put them into a dependent position. The first point is contradicted by various facts. Bardeen found that muscle differentiation began before the nerves grew forth from the tube. Also Harrison demonstrated an inde

66 Journal of Comparative Neurology and Psychology.

pendent development of muscle with fibrils, striation and sarcolemma after excision of the spinal cord and ganglia in larvae of 2.9 and 3.7 mm. NussBAUM, too, admits an independent development of embryonic muscles up to a certain degree — all agree with Goldstein's result that Neumann's first claim is incorrect. His second claim, made in order to explain the persistence of muscles in the amyelic monsters, would not be conclusive on his own assumptions. The muscles need not have degenerated within the short time between the lesion (3d or 4th month) and birth (usually in the 6th or 7th month). In WeberAlessandrini monsters the lesion must have occurred at 2-3 months and the animal reached full term ; the muscles were, therefore, degenerated and mere fat layers and tendons. Leonowa also found the muscles of her case of amyelia extraordinarily fatty. (It seems, moreover, that the condition of the motor nerves in these monsters is not satisfactorially ascertained, but should be of great importance in view of Bethe's claims). Herbst's attempt to attribute a trophic control over muscles to the spinal ganglia is refuted. Taking all the facts together, Goldstein comes to the conclusion that the central nervous system during a certain early period of development has no demonstrable morphogenetic influence on the developing organism.

The second part of Goldstein's article furnishes evidence showing that this same rule holds for regeneration. Regeneration need not follow the rules of development. It also depends largely on the age of the animal or embryo. The results in invertebrates are contradictory ; those on vertebrates (Barfurth), probably favorable to the theory of independence from the central nervous system. In adult Tritons Wolff thought he had jiroved the necessity of a nervous influence. He obtained regeneration of a leg after removal of the cord, but with intactness of the ganglia ; when the operation was done while regeneration had begun it was arrested in all but six cases. In an experiment of Schaper on a Triton larva of 30 mm., an extremity was regenerated after destruction of the cord, and although there was complete absence of sensibility and motility. What nerves there were, came "largely" from the spinal ganglia ; the muscles were normal ; even a piece of i mm. of spinal cord had reformed at the posterior end of the cut of the cord.

The conclusion is : In the stage of organ formation (Rorx) the normal development and regeneration take place quite independent of the nervous central system. In the stage of functional development there is, however, a decided influence from the central organ.

a. m.

Literary Notices. 6y

Davenport, C. B. Statistical Methods with special reference to biological variation. New York. John Wiley and Sons, viii -\- 223, second, revised edition, 1904. In this edition Davenport has revised and enlarged his handbook of statistical methods in a manner which greatly increases its value to the student of biological variation. An important new chapter deals with the results of statistical work, several new methods are described, and the bibliography is much enlarged.

The present scope and nature of the book is well indicated by the titles of the several chapters : I. On methods of measuring organisms, II. On the seriation and plotting of data and the frequency polygon, III. The .'classes of frequency polygons, IV. Correlated variability, V. Some results of statistical biological study.

The work of the publisher is no less admirable than that of the author of this guide to statistical methods. Every student of exact science will find the book serviceable, and no student or investigator of biological variation can afford to be without it. Dr. Davenport deserves much credit for the impetus which his energy and enthusiasm have imparted to biometric research in America. r. m. y.

Deaver, John B. Surgical Anatomy of the Head and Neck. Philadelphia, P. Blakistoti' s Son ^ Co., 1904, pp. 770.

This book is printed from the same plates as the author's three volume work on Surgical Anatomy, those sections being assembled which will be of greatest interest to specialists in diseases of the eye, ear, nose, mouth, throat and nervous system. The book will prove useful to these and also to physiologists, psychologists and general readers who require a manual for ready and rapid consultation. The text is brief and clear and the illustrations are admirable. There are 177 full page plates drawn from original dissections, which by themselves constitute a useful atlas of topographical anatomy. All parts are designated on the plates in full, thus permitting the reader to glean much of his information by simple inspection without consultation of the text. The external and gross features of the brain are fully and clearly figured, without, however, any attention to histological detail. The printing and binding are exceptionally good.

c. J. H.

Mendel, E. and Jacobsohn, L. Jahresbericht iiber die Leistungen und Fortschritte auf dem Gebiete der Neurologic und Psychiatric. VII JahrgangBericht iiber das Jaht 1903. Berlin, S. Karger. 1904. Price, M. 35.

The Jahresbericht is issued this year in two volumes and, like its predecessors, is indispensable to all who wish to keep abreast of the vol. uminous literature of neurology and psychiatry. The bibliographies are

68 Journal of Comparative Neurology and Psychology.

grouped by topics and accompanied by critical annotations on the contents of nearly all of the papers cited. Compendious indexes make the whole mass of material instantly available. c. j. h.

Sfameni, Pasquale. Sulle terminazioni nervose nei genitali femminili esterni e sul loro significato morfologico e funzionale. Archivio dt Ftsiologia, 1904, 1, 345-3'^4In this careful and thorough investigation the author has confined his attention to the nerve terminations iu the clitoris and the vulva. Topographically these are of three classes : (a) Intrapapillar nerve terminations ; (b) nerve terminations iu the reticular layer of the derma ; (c) terminations in the subdermal connective tissue. The first and second of these are by far the most important. All of these terminations reduce to the single type of '"a nervous organ, with or without an envelop of connective tissue, composed of one or more nerve fibers which, after divesting themselves of their myelin sheaths, if they have any, ramify in and around a granular, nucleated substance."

The nerve corpuscles are not the terminations of the sensory ■neryes but are peripheral ganglia, corresponding to spinal ganglia, and their function is to bring about a more subtle division and modification of the stimuli. The real terminations are always differentiated ectoderm cells scattered through the epithelium and the superficial layers of the derma, and connected with sensory nerve fibers. There is a bibliography of 46 citations. j. c. bell.

Harrison, Ross Granville. An Experimental Study of the Relation of the Xervous System to the Developing Musculature in the Embryo of the Frog. The American Journal of Anatomy, 1904, 3, 197-220.

Observations upon embryos the spinal cord of which had been completely isolated before the appearance of either nerve fibers or contractile substance in the musculature, and upon larvae which were reared in a state of constant narcosis by means of a dilute solution of acetone chloroform. Conclusion: "all of the constructiye processes inyolved in the production of the specific structure and arrangement of the muscle fibers take place independently of stimuli from the nervous system and of the functional activity of the muscles themselves. G- E. c. Hardesty, Irving. On The Development and Nature of the Neuroglia. The American /ournal of Anatomy, 1904, 3, 229-268.

A study of pig embryos to demonstrate the syncytial nature of the neuroglia as proposed in the author's earlier paper on the spinal cord of the elephant. .Attention is called to cells resembling the nervecorpuscles of the peripheral nerve which encircle the medullating axones of the cord. g. e. c.

Literary Notices. 6g

Streeter, George L. The Structure of the Spinal Cord of the Ostrich. T/ie Aniencan Journal of Anatomy, 1 904, 3, I -27.

A description of the meninges and the macroscopic and microscopic features of the cord ; inchiding noteworthy contributions on thearachnoidea, the relation of the peripheral gHa sheath to the sinus rhomboideus, Reissner's fiber, and the nuclei marginales. Tabulated measurements and a diagram of the cross-section area, in each segment of the cord, of the funiculi ventro-laterales, substantia grisea, and funiculi dorsales. G. e. c.

Fiirbringer, Max. Morphologische Streilfragen. i. Nervus trochlearis. 2. Rabl's Methode und Behandlung der Extremitatenfrage. Morph. Jahrb., 1902, 30, 85-274. Part I (pp. 86-143) is an important contribution to the morphology of the fourth nerve. It is, in the main, a reply to the criticisms made by Rabl and Dohrn upon the author's theory to account for the dorsal origin and the crossing of the nerve in question : viz. that the superior oblique muscles were originally muscles of the parietal eye, and that with the disappearance of that organ the originally right oblique muscle became associated with tlie left eye, and vice versa. A bibliography of 296 titles. G. e. c.

Van Gehuchten, A. Considerations sur la structure interne des cellules nerveuses et sur les connexions anatomiques des neurones. Le Nevruxe, 1904, 6, S3-116. The author places great emphasis upon the fact that the anatomical independence of neurones, as they are demonstrated by the methods of Golgi and Ehrlich, is the substance of the neurone theory. He considers that, so interpreted, the neurone theory is not contradicted by a single anatomical fact. Intracellular continuity by means of nets and "nervose Grau" as proj^osed by Bethe, Nissl others is purely hypothetical. Even the auto-regeneration of the axone as demonstrated by Bethe, and va.v Gehuchten repeats this experiment successfully, affects only our idea of the origin of the neurone and does not bear upon the neurone theory proper. While in some nerve cells the fibrillae seem to be independent, in many they clearly anastomose in the dendrite and especially in the perikaryon. This condition refutes Bethe's and Nissl's opposition to the idea of polarity of the nerve cell, and supports the neurone theory.

g. e. c.

Van Gehuchten, A. Connexions centrales du noyau de Deiters et des masses grises voisines (Faisceau vestibulo-spinal, Faisceau longitudinal posterieur, Stries meduUaires) . Le Nevraxe, 1904, 5, 19-74A critical review of recent literature on the subject, and a report

70 Journal of Comparative Neurology and Psychology.

on several of the author's degenerative experiments on rabbits. The Marchi method was supplemented with the methods of indirect Wallerian degeneration and of Nissl. The vestibulo-spiual tract arises exclusively after Deiter's nucleus and descends in the anterior column as far as the lumbo-sacral region. The spinal portion of the posterior longitudinal fasciculus is exclusively descending. Ascending fibers are found in this tract only within the bulb and mesencephalon. Both ascending and descending heterolateral fibers probably come as inferior arcuate fibers either from the terminal vestibular nucleus or from the tuberculum acusticum. The ascending homolateral fibers arise higher up, probably from the nucleus of Bechterew. The striae medullares arise exclusively in the tuberculum acusticum. G. E. c.

Soukhanoff. Contribution a I'etude du reseau endocellulaire dans les elements nerveux des ganglions spinaux. Le A'ez'rctxe, 1904, 6, 75"^° The endocellular net as observed by the Kopsch osmic acid method is identical with the Golgi endocellular net, and is not the same structure as the canaliculi of Holmgren and others. G. E. c.

Agababow, A. Ueber die Nerven der Sclera. Archiv f. viik. Anat., Bd. 63, H. 4, pp. 701-709, 1904.

Hyde, Ida H. Localization of the Respiratory Center in the Skate. Amer. four. F/iysiol., 1904, 10, 236-258.

By the employment of careful and precise methods of experimentation on living skates, Miss Hvde has demonstrated the segmental arrangement of the respiratory center. The animal under observation was placed on a board, and sea-water was passed in a continuous stream through a tube into the mouth. Artificial respiration could in this way be maintained for days. In a skate thus situated the medulla may be separated from the spinal cord and from those portions of the brain lying anterior to it without destroying its function as a respiratory center.

Medisection of the medulla is followed, after the inhibitory effects of shock have passed off, by a resumption of coordinated respiratory movements on both sides of the body. The gill arches of one side may move in a rhythm quite different from that of the opposite side. The spiracles may keep time with the gill arches of their respective sides, or both spiracles may be in rhythm with the arches of one side. From the results of median section of the medulla it be. came evi<lent that "the centers for the nervous respiratory mechanism in the skate were bilateral, each half controlling the movements of its respective side."

Evidence of the segmental character (jf these bilateral respiratory

Literary Notices. yi

centers was obtained by following medisection with liemisection. One lateral half of the medulla was separated into anterior and posterior divisions by a transverse cut. The arches and spiracle of the uninjured side continued their normal movements. The spiracle and first gill arch, controlled by the anterior division of the opposite side, sometimes exhibited a rhythm which differed both from the rhythm of the uninjured side and also from that of the remaining gill arches of the same side, which were under the control of the posterior division of the lateral half of the medulla. At times the respiratory mechanisms connected with all three divisions of the medulla moved in unison.

Lesions of the different lobes of the medulla indicated that the ganglion cells of sensory respiratory neurones, those of the seventh, ninth and tenth cranial nerves, are situated in the lobus vagi. Motor ganglion cells and neuraxones were found ventrad of the lobus vagi and in the fasciculus longitudinalis posterior. F. w. c.

Gardella, Eloisa. Azione dell' acido fenico sulla sensibilita gustativa. Archrvio di Fisiologia, 1904, 1, 398-402.

From its general anaesthetic properties we would expect phenic acid to cause a rise in the threshold of sensibility to taste stimuli. On the contrary, increasing strengths of phenic acid solution up to 3 % , applied for 10 seconds, cause a lowering of the threshold for all four of the tastes, bitter, sweet, salt and acid, most of all for bitter. With a constant 2.5% solution and with varying time the threshold falls rapidly during the first 10 seconds, then rises to the normal at 30 seconds, and from there continues to ascend more slowly to 45 seconds.


Langley, J. N. On the Effects of Union of the Central Part of the Cervical Sympathetic with the Peripheral Part of the Chorda tympani. Archivio di Fisiologia, 1934, 1, 505-511.

Three cats were operated upon and in 89, 96 and 176 days respectively they were killed and the nerves examined. Experiments upon the operated cats and examination of the nerves led to the conclusion that "the sympathetic fibers had united with the nerve cells on the course of the chorda tympani, and in consequence that sympathetic preganglionic fibers are capable of uniting with other similar cells . . . which occur on the course of the cranial nerves." The observations also showed that "the formation of meduUated fibers in a peripheral nerve may be caused by fibers reaching it which are themselves non-meduUated." j. c. bell.

72 Journal of Comparative Neurology and Psychology.

Carlson, A. J. The Rhythm Produced in the Resting Heart of Molluscs by the Stimulation of the Cardio-Accelerator Nerves. Amer. /our. of Physiol., 1904, 12, 55-66.

The author has proved that the cardio-accelerator nerves of molluscs when stimulated electrically produce rhythmical contractions of the resting heart. He is inclined to believe, also, that under favorable conditions the resting mammalian heart may likewise be caused to contract by stimulation of the accelerator nerves. The physiology of the cardiac nerves of invertebrates is of special interest because in some forms accelerator nerves alone are present, in others inhibitory nerves alone, and in others nerves which are both inhibitory and acceleratory.

R. M. V.

Carlson, A. J. The Nervous Origin of the Heart-beat in Limulus, and the Nervous Nature of Co-ordination or Conduction in the Heart. Amer. Jour, of Physiol., 1 904, 12, 67-74.

"It can now be stated as a fact that in Limulus the origin of the heart-beat is nervous, not muscular, and that conduction of the impulse or the co-ordination of the different parts of the heart takes place through the nerves, not through the muscular tissue." This important conclusion Dr. Carlson draws from the results of a series of ideally definite and well chosen experiments. He has shown (i) that lesion of the median and the two lateral nerves which supply the heart destroys coordination, (2) that cross-section of the heart does not interfere with co-ordinated rhythm so long as the nerves are intact, (3) that heartbeat ceases when the median nerve is severed. The author has found material which is particularly favorable for the study of heart action in its relation to the nervous system, and the importance of his work will be recognized at once by all who have interest in physiology.

R. M. Y.

Schiller, V. Ueber die physiologischen Wirkungen des Delphinins (Hegl). Arch. f. {Anat. u.) Physiol., 190^, 248-255, Taf. VI.

Delphinin paralyzes the vagus endings iu the heart to a marked degree. . r. p.

loteyko, J. Mecanisme physiologique de la reaction de deg^n^rescence des muscles. Travaux du Laboratoire de Physiologie, Instituts Solvay, 1904, 6, 59-78. The reaction of the muscle which has undergone degeneration because of separation from its nerve argues for contractility of the sarcoplasm, and a qualitative difference in irritability of the sarcoplasm as compared with the fibrillae. The author questions whether there may not be a similar qualitative differentiation of irritability in the substance of the nerve. <>• e. c.

Literary Notices. 73

Klein, Fr. Das Wesen des Reizes. Ein Beitrag zur Physiologic der Sinnesorgane, inbesondere des Auges. Arch. f. [Anat. u.) Physiol., 1904, 305-342. A fundamental attribute of protoplasm is that it is stimulated only by changes in the normal or usual external condition. The same thing applies to the nerve endings of the retina, hence certain conditions are possible under which we are blind ("rest blindness"). Two such conditions inducing "rest blindness" have been investigated. These are (a) hindering as much as possible every eye movement, (b) illuminating the eye from an equally lighted ("reizlose") surface. The author thinks it probable that only an increase in the intensity of light can act as a stimulus, never decrease. R. p.

Rothmann, M. Ueber die Leitungsbahnen des Beriihungsreflexes unter Berilcksichtigung der Hautreflexe des Menschen. Arch. f. {Anat. u.) Physiol., 1904, 256-270.

In dogs and cats the conduction in either ascending or descending direction of the impulses in Munk's "contact" reflexes of the extremities, is not confined to a single sensory or motor fiber tract. The ascending impulses of the reflex are conducted in two fiber tracts : one in the anterior and one in the posterior column. There are also two paths for the descending impulses : the pyramidal tracts and MonaKow's bundle. The "contact" reflex of lower animals agrees in many points with the skin reflex in man. R. p.

Birukoff, B. Zur Theorie der Galvanotaxis. Arch. f. {Anat. u.) Physiol., 1904, 271-296.

Galvanotaxis a combination of two factors : a physiological (general irritability) on the one hand, and a purely physical (kataphoric action on the other. r. p.

Muskens, L. J. J. Studies on the ^Maintenance of the Equilibrium of Motion and its Disturbances, so-called "Forced Movements." Jour, of Physiol., 1904, 31, 204-221.

Experiments producing forced movements after lesions of the central nervous system were performed on octopods, selachians (Acanthias, Scyllium), frogs, cats and rabbits. The conclusions with reference to lower vertebrates are : i. That rolling after unilateral lesion of the hind-brain, and circus movement after unilateral lesion of the base of the mid-brain appear to be, in lower vertebrates, fundamental phenomena, esjjecially localized in these parts; 2. That rolling to the impaired side and circus movement to the non-impaired side are phenomena narrowly related; 3. There is no reason to believe in any diametrical opposition of forced movements of an irritating and of a paralyzing nature. In cats and rabbits unilateral lesions of the middle and upper cerebellar peduncle and unilateral (partial more than com

74 Journal of Comparative Neurology and Psychology.

plete) removal of a lateral lobe of the cerebellum cause rolling movements to the side of the lesion, or a tendency to it. The most marked rolling was observed after lesion or section of the middle peduncle. Studies ou the changes in the mobility of the eyeballs were also conducted. R. p.

Woodworth, R. S. and Sherrington, C- S. A Pseudaffective Reflex and its Spinal Patli. Jour, of Physiol., 1904, 31, 234-243.

The lateral column furnishes the headward path in the spinal cord for nociceptive (algesic) arcs ; each lateral column conveys such impulses from both lateral halves of the body, and somewhat preponderantly those from the crossed half; these relations are true whether the arcs be traced from skin, muscle or viscus. r. p.

Langley, J. N. On the Question of Commissural Fibers between Nerve C'ells having the Same Function and situated in the same Sympathetic Ganglion, and on the Function of Post-Ganglionic Nerve Plexuses. Jour, oj Physiol., 1904, 31, 244-259. Concludes that since pilo-motor nerve-cells, in a given ganglion of the sympathetic system can be shown not to be connected with each other by commissural fibers, and that the same is true of pupillo-dilator and vaso-motor nerve-cells, it is fair to assume "that none of the nerve-cells of the sympathetic system, and none of the similar nervecells on the course of the cranial and sacral nerves are so connected."

R. p.

Langley, J, N. and Anderson, H. K. The Union of Different Kinds of Nerve Fibers. Joitr. of Physiol., 1904, 31, 365-391.

The nerve fibers dealt with are divisible into four classes, (a) the efferent fibers which run from the central nervous system and end in multi-nuclear striated muscle cells (efferent somatic fibers), (b) the efferent fibers which run from the central nervous system and end in ganglion cells (pre-ganglionic fibers), (c) the efferent fibers given off by the peripheral ganglia (post-ganglionic fibers) and (d) afferent fibers connected with nerve cells in the posterior root ganglia. The general conclusions drawn from the experiments are, (i) The central end of an efferent fiber can make functional connection with the peripheral end of any other efferent fiber of the same class whatever be the normal action produced by the two fibers. (2) The central end of any fiber of class (a) can make functional union with the peripheral end of any fiber of class (b), and the central end of any fiber of class (b) can make functional connection with the peripheral end of any fiber in class (a). (3) The peripheral ends of cut nerve fibers appear to have a chemotactic influence on the central ends. The chemotactic action is strong

Literary Notices. 75

€st among fibers of the same class. (4) It is provisionally concluded that (A) efferent fibers proceeding from the spinal cord, (B) postganglionic nerve fibers, and (C) afferent nerve fibers are incapable of coalescing with one another. r. p.

Robertson, T. B. On the "Sham-Death" Reflex in Spiders. Jour, of Physiol., 1904, 31, 410-417. Three species of spiders were used, two classed as "active" {Epeira producta (?) and Aniaurobhis sp.) and one as "sluggish" (Celacnia excavata). To produce the "sham-death" reflex in the active species the stimulus must be sharp and sudden. In the sluggish species the "sham-death" posture is continuous and therefore practically independent of the stimulus. The reflex is carried out without the aid of the abdominal nerves, and in the active species can be performed in an altered manner by the thoracic ganglia alone, or even by the ganglia of the two posterior or two anterior segments of the thorax alone. In the sluggish species the reflex cannot be induced without the head ganglion. The "sham-death" reflex is a complete tetanus when the nervous system is intact. In conclusion the author says : "In the active species of spiders, in which the "sham-death" posture is probably a means of escape in emergency even if they be regarded as possessing "conscious volition" at all, unless we suppose that this is an attribute of each of the thoracic ganglia, this reaction cannot be due to their conscious volition. That is to say, purposeful as it may seem as a means of protection from their enemies, this reaction is a necessary consequence of the structure of their nervous system and the functional activities of the nervous elements, independently of any mental processes connected with the cerebral nervous masses."

R. P.

Langley, J. N. and Anderson, H. K. dn Autogenetic Regeneration in the Nerves of the Limbs, four, of Physiol., 1904, 31, 418-428.

May, W. P. The Innervation of the Sphincters and Musculature of the Stomach. Joitr. of Physiol., 1904, 31, 260-271.

Elliott, T. R. and Barclay-Smith, E. Antiperistalis and other musular Activities of the Colon. Jour, of Physiol., I904, 31, 272-304.

Sherrington, C. S. On Certain Spinal Reflexes in the Dog. Jour, of Physiol., 1904, 31, (pp. 17-19 of Proc. Physiol. Soc.)

Stuart, T. P. A. A Contrast (?) Experiment. Jour. Physiol., 1904, 31, (pp. 6 and 7 of Proc. Physiol. Soc.)

Werndly, L. U. H. C. Aequisonore FlSchen rings um eine ertonende Stimmgabel. Arch. f. (Anat. u.) Physiol., 1904, 297-304.

']6 Journal of Comparative Neurology and Psychology.

De Fleury, Maurice. Manuel Pour L'Etude des Maladies Du Syst^me Nerveux. Pans, Felix Alcan, 1904.

This work is essentially the digest of De Fleury's effort to review the status of neurology after several years of special work on the neuroses. With his clearness of distinction and presentation he hoped to do a valuable service to the student by putting together the notes from his reading. The result is a volume of 1000 pages, beginning with very full directions as to how to examine a case (pp. 1-105). This is followed by a brief medical anatomy of the nervous system (pp, 109-216), very elementary and schematic, but filling its purpose for beginners. The rest of the book is devoted to the individual diseases, very concise and lucidly written. There is, of course, a great latitude of personal predilections ; the chapter on neuroses is especially well treated.

As a student's manual and for rapid orientation the book deserves full praise. a. m.

Camus, Jean and Pagniez, Ph. Isolement et psychotherapie. Traitement de I'hysterie et de la neurasthenic. Pratique de la reeducation morale et physique. Preface de M. le professeur Dejerine, medecin de la Salpetridre. Paris, F. Alcan, 1 904.

For a number of years Professor Dejerine has paid attention to the method of isolation and to psycho-therapeutics in hospital and private practice.

His pupils here present a work which gives on ground of 60 well chosen observations the statement of therapeutics which they developed and extended on the ground of Dejerine's principles. After four historical chapters, they discuss the isolation and the various forms of its application, the rest and over-feeding, and other physical helps, such as electrization and hydrotherapeutics. And after a discussion of the reciprocity of physical and mental life, they enter upon hypnotism ; the comparative value of suggestion and persuasion ; the importance of confidence and attention of the patient, and the means of obtaining them ; and finally the rules for psycho-therapeutic conversation and the principles of reeducation.

A final chapter speaks of prophylaxis, heredity, education, and the question of overwork.

Considering the material presented, the book is an extremely valuable help in many practical and general questions concerning the neuroses. A. M.

Literary Notices. yy

Hall, G. Stanley. Adolescence. Its Psychology and Relation to Physiology, Anthropology, Sociology, Sex, Crime, Religion and Education. New York, D. Appleton and Company, 1904. Vol. I. xx -\- t;88, Vol. II, 784 pp.

This work, by intent at least, is a biological psychology. The author has devoted many years to the study of various aspects of adolescence. During this time he has influenced and encouraged a great number of students in the investigation of problems of genetic psychology ; unceasingly he has urged that "the studies of the mind need new contact with life at as many points as possible." Instead of speculating concerning the future of the soul we should study its past, if we are desirous of understanding its present condition. "We must collect states of mind, sentiments, phenomena long since lapsed, psychic facts that appear faintly and perhaps but once in a lifetime, and that in only few and rare individuals, impulses that, it may be, never anywhere arise above the threshold, but manifest themselves only in automatisms, acts, behavior, things neglected, trivial and incidental, such as Darwin says are often most vital. We must go to school to the folk-soul, learn of criminals and defectives, animals, and in some sense go back to Aristotle in rebasing psychology on biology, and realize that we know the soul best when we can best write its history in the world, and that there are no finalities save formulae of development."

The work is much more than an assemblage of facts, for the author has convictions as well as ideas. His concern, as early appears, is with the bearing of his researches upon education. A few words from the introductory remarks will serve to indicate the practical trend and tone of the book. "Never has youth been exposed to such dangers of both perversion and arrest as in our own land and day. Increasing urban life with its temptations, prematurities, sedentary occupations, and passive stimuli just when an active, objective life is most needed, early emancipation and a lessening sense for both duty and

discipline " And again, "In education our very kindergartens

tend to exterminate the naive which is the glory of childhood.

Everywhere the mechanical and formal triumph over content and substance, the letter over the spirit, .... information over education, marks over edification, and method over matter."

The Psychology of Adolescence, as the book might be called, is full of the earnestness of purpose and enthusiasm of the author. Its facts are vitalized by the sense of contact with life which each page gives. President Hall has handled with admirable skill a subject which was at many points extremely difficult of treatment.

R. M. Y.

78 Journal of Comparative Neurology and Psychology.

Davenport, C. B. The CoUembola of Cold Spring Beach, with Special Reference to the Movements of the Poduridae. Co/J Spring Harbor Monographs, II. 32 pp., I pi., 1903.

This is a description of the form, systematic relationships, habitat and modes of behavior of the simple insect CoUembola. Apart from the new results of observation, which are interesting even to non-technical readers, the paper is valuable for its suggestions of methods and interests in the study of animals.

The spirit and purpose of the writer in his research can best be understood with the aid of the following paragraph from the monograph : "By this analysis I wish to enforce the idea that the movements of the podurids are not to be referred to so many 'vohtional' acts ; nor that, as instincts, are we to think of the tendency to particular kinds of behavior as having been inherited. But rather, the animal being provided with a sensitiveness of a certain sort to oxygen, to water, to currents of air or other movements, to contact, to gravity, and to light, it will behave on the beach as we see it behave there. Under another set of environmental conditions it will behave very differently. The movents of the podurids ott the beach are the result equally of the specific, inherited capacity of response on the one hand and the particular stimuli afforded by the conditions of the beach."

And again, in a brief but effective summary the author thus states the chief matters of observation : "The podurids of the beach live between tide-marks, go into the sand at high tide and rise to the surface when the tide is out to take air. They run up surfaces in the face of the wind and leap when they reach the top, being blown back to the starting point. They are exceedingly sensitive to gravity, to contact, to moisture, to currents of air and to light, and these elementary reactions are so combined as to bring about their normal movements. They are provided with these instincts before they reach the beach. Had they not had such instincts they could never have lived on the beach. The instincts have selected the habitat."

R. M. Y.

Bigelow, Henry B. The Sense of Hearing in the Goldfish, Carassius auratus L. American Naturalist, 1904, 37, 275-284.

The experiments which form the basis of this paper were performed to test the accuracy of Kreidl's conclusions, "that goldfishes do not hear" — a conclusion reached after studying the sound reactions in fishes from which the semicircular canals with the attached parts had been removed.

The author first studied the reactions of normal fishes to sound vibrations and found that at least 78% of them gave definite well

Literary Notices. yg

marked reactions to such vibrations. He then studied the sound reactions of fishes on which various operations had been performed.

In one lot of fishes the skin was rendered insensitive by cutting the spinal cord and the lateral branches of the loth, yth and 5th nerves. The fishes with few exceptions recovered from the operations and many of them lived for weeks afterward. 80% of the animals with the nerves of the skin destroyed still gave normal reactions to sound vibrations.

In a second lot the ears were made insensitive by cutting the 8th nerves. This resulted in complete loss of response to sound vibrations.

In a third lot the 8th nerve on one side was first cut in each fish. After this operation the fishes gave practically normal sound reactions, but after the nerve of the opposite side had been cut the fishes no longer responded to sound vibrations.

In a fourth lot the semicircular canals and attached portions of the ears were pulled out. After this operation, the fishes still reacted to sound, but the reactions were less marked in normal fishes.

The author thus repeated Kreidl's experiments and confirmed his results, but on dissecting these fishes he found that a portion of the ear largely imbedded in bone had not been removed. This portion he thinks probably represents the combinrd sacculus and lagena of higher vertebrates. It contains two otoliths and is well supplied with nerves from a branch of the 8th.

The experiments seem to show beyond a doubt, that the ear in goldfishes functions as an organ of hearing. A somewhat similar set of experiments, performed a year earlier by Dr. G. H. Parkee on Fundulus heterocUtiis, led to the same conclusion with reference to that species. Dr. Parker, however, was unable to note any reaction to sound vibrations in the smooth dog fish. s. o. mast.

Andrews, E. A. Breeding Habits of Crayfish. American Naturalist, 1904, 37, 165-206.

In this paper the author describes the habits of Camhariis affinis taken from the Potomac river. The habits of C. affinis are found to vary considerably from those of Astacus, the European crayfish. This fact lends additional interest to the observations of Andrews.

The most important points brought out in the paper may be briefly stated as follows :

Mating takes place during February, March and April, and sometimes in October and November. Each male may mate with more than one female. Sex union continues from 2 to 10 hours. The sperm is transferred to external seminal receptacles. The eggs are

8o Journal of Comparative Neurology and Psychology.

laid at night, often not until "some weeks" after sexual union. After the eggs are laid the mortality is great among the females. It is, however, still greater among the males after sexual union, many dying immediately thereafter. Before the eggs are laid, the female thoroughly cleans her ventral surface. The eggs, from 400 to 600 in number, are fastened to the pleopods in clusters, much like bunches of grapes. Immediately after laying the eggs, the female seeks a dark place and there rolls from side to side, resting from i to 6 minutes on either side and on the ventral surface. This reaction continues for several hours and probably serves to fasten the eggs to the pleopods. Eggs removed from the female die unless the embryos in them are well developed. It requires from 6 to 8 weeks for the embryos to hatch, after which they remain attached to the mother continuously during the first week and part of the time during the second. After the second week they no longer return to the mother. If several mothers with embryos are in the same dish, the embryos frequently crawl onto the wrong mother. There is no evidence that the mothers recognize their own young. The embryos moult 2 days after hatching. The intervals between the following 5 moults vary from 5 to 18 days.

Practically all the observations were made in the laboratory. It is to be regretted that more work was not done in the field. Most of the descriptions are worked out in detail. These together with 10 very good figures convey a very clear conception of the habits studied and seem to show that the author made keen, thorough, careful observations. The literary merits of the paper, however, might have been improved by careful revision. s. o. j^iast.

Taverner, P- A. A Discussion of the Origin of Migration, Auk, 1904, 21, 322-333 Return of birds to the winter home is easy, that to the summer home is hard"; to explain. The author discusses previous theories, and concludes that sufficient emphasis has not been placed on the fact of the great increase in the bird population at the beginning of the breeding season. In southern latitudes some birds begin to breed early ; when later breeders are ready to commence they find that food has become scarce. The scarcity of food thus brought about was, Mr. Taverner thinks, the chief cause of birds' having commenced the habit of seeking a new home at the beginning of the breeding season.

WALLACE CRAIG. Rawitz, B. Die Unmoglichkeit der Vererbung geistiger Eigenschaften beim Menschen. BioL Centralbl., 1904, 24, No. 12, 396-408.

Rawitz argues vigorously against the inheritability of mental traits in man. His standpoint is that of extreme materialism, g. w.

The Journal of Comparative Neurology and Psychology

Volume XV 1905 Number 2


By Irving Hardesty.

[From the Hearst Anatomical Ladoratoiy of the University of California.) With four figures in the text.

Some months ago a zoo in San Francisco lost by sudden death two specimens of Emu [Dromaeus novae-hollandiae) and the management very kindly delivered the carcasses to me within a few hours after death. The birds were immediately dissected and, with certain of the other organs, the central nervous system of each was removed and preserved in io% formalin.

The specimens were full grown and considered of large size, being nearly as large as the adult ostrich. With the exception of the head and upper part of the neck, the bird resembles the Cassowaries ; but it is larger, its neck relatively longer and it belongs to a different genus. In general size of the body it is intermediate between the Rhea, or South American ostrich, and the ordinary or African ostrich [Strut/no camelus). The head, though less bare, is very similar to that of the ostrich, while most of the neck, unlike the ostrich, is covered with the long streaming plumage common to the rest of the body. The wings are even more rudimentary than those of the ostrich, being nothing more than slender stubs entirely hidden in the plumage and even void of spiny rudiments of wing feathers. Similar to the ostrich, the legs are bare but somewhat shorter in proportion than those of the ostrich and are relatively more stout, the "drumstick" being nearly as large in circumference as the thi^h of a medium sized man.

82 Journal of Comparative Neurology and Psychology.

The spinal cord of such an animal necessarily presents features of unique interest and preparations were made with the intention of describing it. Meanwhile, however, the paper of Streeter ('04) appeared giving a description of the spinal cord of the ostrich and on comparison it was found that the spinal cord of the emu, in its general features, is apparently so similar to that of the ostrich that a full description of it was deemed unnecessary. Streeter's desrcription of the one may be referred to for the most part as a description of the other.

This paper, therefore, may be limited to a few of the features presented in the spinal cord of the emu not touched upon by Streeter for the ostrich and to some further observations upon one or two points less fully treated by him.

In the lumbar enlargement alone Streeter mentions having observed a segmented appearance ("neuromeres") in the ostrich cord. He does not refer to such as present in other localities and little more than mentions it in the lumbar region. After noting a few of the general macroscropic features of the emu cord, some attention will be given here to the appearance of evident segmental enlargements in other localities, as well as in the lumbar region. In this respect it may differ from the spinal cord of the ostrich.

Lying in the vertebral canal, the spinal cord of the emu is surrounded by an epidural cavity which is somewhat larger in proportion than is usual in the more commonly studied mammals. The dura mater has an unusually smooth outer surface. Between the nerve pairs of the adjacent segments it appears thicker along the ventral and ventro-lateral aspects of the cord than along the dorsal aspect, while at the levels at which the nerves are attached it appears thicker along the lateral surfaces. From a number of measurements, made later from stained sections, the thickness of the dura mater was found to vary from 87 ix to 146 [X, withan average thickness of 113/^. The arachnoidea is similar in proportional thicknesss and arrangement to that usually found in the mammals. The pia, however, is relatively thick as compared with that of the mammalian cord. It is well defined as to its outer surface and so thick as to be easily stripped from

Hardesty, spinal Cord of the Emu. 83

the specimen hardened in formahn. Measurements from the sections show its thickness to vary from 36 ix to 65 fi, with an average thickness of 53 n. Completely fused with and continuous with the pia mater are its three ligaments. These were described by Stkeeter as occurring in the pia of the ostrich cord and they are remarkably well developed in that of the emu. The liganienta iongitiidiualia lateralia course along the lateral surfaces of the cord proper in a line ventral to that along which the dorsal and ventral nerve roots approach each other (/,//, Figs. 2. 3 and 4). The hga7nentinn longitiidinale ventrale {Lh, Figs. 3 and 4) is triangular or dentate in section, its apex projecting into the relatively wide fissnra uicdiana ventralis. Neither of these ligaments disturbs the outer contour of the pia and can only be detected in the fresh by a slight difference in the color of the lines it occupies. The lateral ligaments appear grooved into the white substance of the cord and increase in thickness with tiie increase in the diameter of the cord. i. e. , from the neck caudad. They appear fusiform in transverse section, thinning at the edges to become continuou;; with the pia. Measurements at their thickest part give an average thickness of 1 98 ii. All three are true ligaments, being composed almost entirely of large, well defined elastic tissue fibers with the characteristic stellate cells in transverse section, while the pia proper is composed mostly of white fibrous tissue.

In addition to the usual filaments bearing blood vessels and connecting the dura with the periosteum of the vertebral canal, there occurs along each side of the specimen and a short distance caudad to the points at which the spinal nerves penetrate the dura, a series of special supporting filaments. These liganienta, epiduralia occur in groups of 4 to 6 each, approximately in line„ and one of them, that nearest the nerve, is always larger than, the others [^Lep, Figs, i and 2) In the formalin material this, larger one appears whiter than the surrounding tissue, cylindrical and well defined, and microscopic examination reveals it to be a bundle of elastic fibers. Upon opening the dura these bundles along each side are found to be opposite and connected: with the strongest of the ligamenta denticiilata .

84 Journal of Comparative Neurology and Psychology.

Each of the specimens possessed 48 pairs of spinal nerves, or 48 segments in its spinal cord.

In the cervical divison and for about one-third the length of the spinal cord of the emu, the spinal nerves are attached at angles inclined toward the head, or in other words, the respective intervertebral foramena throu^^h which the nerves pass are slightly more cephalad than the levels at which their dorsal and ventral roots come together on the periphery of the cord. Along the middle third of its length the nerves are attached more approximately at right angles to the axis of the cord. As in the ostrich, the spinal cord occupies the entire vertebral canal and there is no Cauda equina. The sacral and coccygeal nerves are arranged so that their respective foramena are but little caudad to the points of attachment of their dorsal and ventral roots.

In consequence of the angles at which the nerves are attached to the cord, the fila radicularia of the dorsal and ventral roots of all the spinal nerves fray along the surface of the cord in both directions. The fila of each root form cephalad and caudad divisions of the root, which divisions, increasing as they course along their respective lines of insertion, approach each other and fuse to forni the root ( see Fig. 2, ventral root).

As to the arrangement of their fila radicularia and their angles of attachment, the spinal nerves of the emu

/•V's'- I ■ The 23rd and 24th segments of tlie spinal cord of the emu, dorsal aspect. Dura mater intact. Xr. — nerve roots ; Lcp. — Ligamenta epiduralia. Natural size after formalin.

differ from those

Hardesty, spinal Cord of tJie Eiini. 85

of the mammals and especially man. In the human, with the exception of the first two or three cervical nerves, all the spinal nerves of the adult pass through their intervertebral foramena caudad t<> the levels of their attachment to the cord and all the fila of their dorsal and ventral roots are distributed cephalad. In the iiuman loetus the cord occupies the entire vertebral canal and the foramen of each nerve is then opposite the segment or level of the cord at which the nerve is attached, but, owing to the fact that the vertebral column grows more rapidly than the spinal cord and continues to grow after the cord has attained its adult length, the foramena become displaced caudad. The cord being attached to the encephalon. the effects of the increase in the length of the vertebrae are superimposed so that the nerves become pulled downward and the nerve roots are necessarily drawn caudad into progressively increasing lengths till the final effect is that the conus meduUaris seldom reaches lower than the upper end of the first lumbar vertebra and the whole lumbar and sacral portion of the vertebral canal contains only the resulting Cauda equina. Like the ostrich, the emu may be said to have two brains, the encephalon and the " lumbar brain ", or the remarkably large intumesccntia hnnbalis. The latter, necessary for the control of the chief mass of ihe body musculature, lies in an especially enlarged portion of the vertebral canal. Thus the two ends of the central nervous system must become mechanically fixed long before the growth of the animal is complete, and as the vertebrae increase in length the intervening portion of the spinal cord must also increase in length. This results in remarkablv long segments in the cervical and thoracic regions of the spinal cord of the emu, and as compensation, explains the fact that the ner\es and nerve roots are not drawn caudad instead, as they are in man. Further, between foetal life and maturity the lower thoracic and the lumbar vertebras of man undergo a proportionately greater gain in length than do the cervical vertebrae,^ thus accen

' Measurements of the length of the first five cervical vertebrae and of the lumbar vertebrae of several human skeletons at birth and several adults show

86 Journal of Comparative Neurology and Psychology.

tuating the drawing downward of the more caudal pairs of nerves. The young ostriches, as I hav^e observed them on the ostrich farms of this State, appear to have necks very much shorter in proportion to the length of the body than do the adult birds, and in all probability the same is true for the emu. If this is true, then during growth the cervical vertebrae must undergo a proportionately greater gain in length than an}' below them, whicli would explain the fact that the nerves of the upper third of the cord are attached at angles slightly cephalad instead of caudad. Another feature to be noted in the spinal nerves of the emu is that the roots and beginning portions of the lower cervical and the thoracic nerves pass from the cord dorsal ward to their respective intervertebral foramena (see Fig. i ). This differs from the course of most of these spinal nerves of the mammals and means that this portion of the cord lies deeper in the bony framework about it. The small neck nerves proper are similar in this respect to the neck nerves of the mammals. As the natural result of the dorsal direction of the nerves, the nerve roots approach each other along the dorso-lateral aspect of the cord, instead of the lateral or ventro-lateral aspect (see Figs. 3 and 4).

J^Yg: 2. Lateral longitudinal or horizontal section of 20th .segment pas.sing at level of central canal. D — dura mater; P — pia mater; Zty^ — Ligamentum epidurale; Z//— ligamentum longitudinale laterale; Lh — ligamentum longitudinale ventrale; /'/' — ventral root; Fi' — fila radicularia of ventral root; .S^'— substantia grisea; Cc— canalis centralis; 7\ww — nucleus marginalis minor; Se — segmental enlargement; J/— level midway between nerves. X A- The parts were outlined by camera.

Fig. J. Sc and M. Transverse sections of 22nd segment. Se — taken through segmental enlargement corresponding to level similarly indicated in Fig. 2; ./I/— taken midway between the enlargements of the adjacent segments at level similarly indicated in Fig. 2. Lh' — ligamentum longitudinale ventrale; _fr — fila radicularia of both roots; other reference letters same as in Fig. 2. Outlined by camera. X 4 Fig. ^. Sc and A/ — Transverse sections of 17th segment. Taken at levels corresponding to those of Fig. 3. Reference letters, drawing and magnification same as in Fig. 3.

that between birth and the adult stages the cervical vertebra; undergo an average gain in length of about 2^0 %, while the lumbar vertebra; gain about 450 «/,, in length.

Haruesty, spinal Cord of the Emu.



88 Jounial of Comparative Neurology and Psychology.

The nerves were clipped within the vertebral canal and the brain and spinal cord were removed unseparated, with the dura mater intact except that of the ventral surface of the brain* The whole was placed in io% formalin which was renewed on the following day. No measurements were made till after the material had .been in this fluid for several months and, unfortunately, no weights were taken in the fresh condition. At the time the study began some weighings and measurements were made of the formalin material. Since the two specimens proved to be quite similar, the following record for one of them will serve for both:

Length ot entire nervous system ....

Length of spinal cord from tip of calamus scriptorius to end of conus medullaris .....

Lateral diameters of spinal cord:

1st segment, near calamus scriptorius 8th segment 17th segment, 22nd segment, 32nd segment, 42nd segment. Weight of brain and spinal cord with dura except

that of basal surface of brain Weight of brain severed at tip of calamus

scriptorius .......

Weight of spinal cord from tip of calamus

scriptorius to end of conus medullaris Weight, of lumbar region, 28th to 38th

segments inclusive .....

Owing to the action of the formalin upon^the tissue, the weights must be regarded as ranging somewhat higher than the fresh material would have given. They may be considered as approximate however, and as indicating the relations between the part-. In every case the pieces were dried on blotting paper for a given length of time before weighing.

It will be noted ( i ) that the cord is quite slender, being about twice as long as that of man; (2 ) that the cord is relatively nearly twice as heavy as the brain, and ( 3 ) the great size of the lumbar enlargement, it being nearly one-third the weight of the brain, and measurements show it to have more than twice the diameter of any other region of the cord.

96.5 centimeters.


8. millimeters

5 6

69.5 grams.




Hardestv, spinal Cord of the Emzi. 89

The following features should be mentioned in which the spinal cord of the emu differs from that of the ostrich, as described by Streeter:

1. The ostrich is given 51 pairs of spinal nerves or 51 segments in its spinal cord. Each of the specimens of emu here studied had but 48 segments.

2. While the sinus rhomboidalis in the dorsal aspect of the lumbar enlargement is very evident in the emu cord it is not quite so wide relatively as Stkeeter shows it to be in the ostrich.

3. The entire length of the spinal cord of the Emu from the point of the calamus scriptorius to the end of the conus medullaris was in one specimen 87 centimeters and in the other 90 centimeters. Practically the same measurement gave Streeter 81 centimeters for the ostrich. He mentions however that his specimen was a young one.

4. To the unaided eye there was no evidence whatever of a distinct cervical enlargement in the spinal cord of the emu. Streeter indicates an enlargement in the ostrich, evident in the gross appearance.

According to Furbinger ('88) the 17th to the 21st segments of the ostrich cord correspond to the region of the cervical enlargement of the higher veitebrates. In the emu, as in the ostrich, the smallest part of the cord is in the neck proper — about the first twelve segments. This is to be expected from the fact that for the slender neck the inervation required of these segments must be small. At the i6th segment in the emu there begins a perceptible increase in the general diameter, in the middle of the 8th segment both the lateral and dorso-ventral diameters being 5 millimeters, while in the middle of the 17th segment, the diameter was nearly 6 millimeters. From the 1 6th segment a gradual increase continued till at the 22nd segment the general diameter became 7 millimeters. From the 22nd to the 27th segment, instead of a decrease there was an increase though more gradual than before. At the 27th segment the increase toward the lumbar enlargement began to show and at the 29th segment the increase became more sudden. At the 32nd

90 Journal of Comparative Neurology and Psychology.

segment, which is in the height of the lumbar enlargement, the lateral diameter became i6 millimeters. The dorso-ventral diameter, however, taken along the lateral wall of the sinus rhomboidalis and the fissura ventralis, was only 7 millimeters.

From the caudal end of the first segment to the beginning of the lumbar enlargement, the spinal cord of the of the emu appears cylindrical or circular in section. With the lumbar enlargement there begins a slight lateral flattening of its contour which attains its height at the middle of the enlargement. The conus medullaris is relatively much longer than in the mammalian cord. At first the decrease from the enlargement is quite sudden, then becomes more gradual, the conus being drawn out to very slender proportions and comprising fully 10 segments, all of which are shorter than those cephalad to the lumbar region.

In length the segments of the two extremities of the cord were considerably shorter than those of the intervening portion, gradually increasing in both directions but more rapidly from the cephalic end. In the thoracic region they attaiaied a remarkable length. The following are the lengths of some of the segments of one of the specimens obtained by measuring from the center of attachment of one nerve pair to that of the next. The measurements were made upon the formalin preserved material.

1st se







1 2th



. . . 36


. . . 38









35 th





_ . \ J


i ] i... 1 _


Streetek ('04 ) described segmental swellings in the lumbar enlargement of the ostrich and shows them occurring at the level of the attachment of the ventral roots of this region.

Hardest Y, Spinal Cord of the Enm. 91

Such swellings, though discernable, were not so marked in this region of the emu cord as his illustration shows them for the ostrich. However, in the longer segments of the emu cord there are indications of segmental arrangement not described by Streeter for the ostrich. These appear in the specimen as slight enlargements either at or. in the upper segments, just caudad to the levels at which the nerve roots pass through the dura mater. They are sufficiently evident to be easily noted in the fresh specimen both with the dura intact and after it has been removed. They are formed by a somewhat gradual increase in the diameter rather than as bulbous swellings and they are evident from the dorsal aspect as well as from the ventral and lateral views of the specimen ( Se, Fig. i ).

The appearance is more marked before the dura is removed. This is due to the dorsal and ventral root fibers collecting beneath the dura to form the roots, as may be seen in Figure 2, which is a lateral longitudinal section passing- along the line of formation of the ventral roots. It is most apparent in the longest segments where the nerves are considerably apart, while obviously in the shorter segments it becomes less discernable or not at all, owing to the fusion of adjacent swellings.

By comparison of transverse sections through the different levels of a segment the nature of the segmental enlargements maybe observed and their extent determined. Figure 3, Se, represents a stained transverse section taken just below the level at which the 22nd pair of nerves is attached, or the level corresponding to that indicated by Se, figure 2, which is a longitudinal section of the 20th segment. Figure 3, M, is a transverse section of the 22nd segment midway between the nerves (level M, fig. 2). Figure 4, Se and M, represent similar sections from the 17th segment. It is seen at once that the Se sections are of appreciably greater diameter than those designated by J/ and that the increase involves both diameters of the specimen. The following measurements of a {^w of the segments indicate the extent of the segmental enlargements. They are made from stained and mounted transverse sections and are exclusive of the dura mater.

92 Journal of Comparative Neurology and Psychology.

Number of Segment. Diameters in millimeters



Near tip of calamus





1st segment :

near nerve



mid-way between




8th segment :

near nerve



mid-way between




17th segment :

near nerve






22nd segment :

near nerve






32nd segment :

near nerve






40th segment :

at nerve



mid-way 4 3

The dorso-ventral diameter of the 32nd segment includes the depth of the sinus rhomboidahs and the fissura ventrahs. Between the two there is Httle more than the gray commissure of the lumbar cord.

Examination of the stained sections reveals that both the gray and the white substance of the cord take part in producing the segmental enlargements. Figures 3 and 4, which are camera drawings, show that that gray figure is larger at the levels nearer the fusion of the nerve roots than it is mid-way between the nerves of adjacent segments.

From both transverse and longitudinal sections (Fig. 2), it is apparent that the white substance plays a somewhat greater role in forming the swellings than the gray. For quite an extent midway between the adjacent nerve pairs of the longer segments there are no fila radicularia coursing upon the pia or evident beneath it. If ventral and dorsal root axones are connected with this region of the segment, as no doubt they are, they must course within the white substance, making their exit and entrance at levels nearer the levels of the respective nerve roots to which they belong. Just as the fila emerging or entering

Hardest Y, Spinal Cord of the Emit. 93

along the pia collect rapidly in approaching from both directions to form the nerve roots of a segment, so must the axones continued from them within the substance of the cord increase in abundance toward the level at which the nerve roots are formed and thus, together with their divisions and collaterals, must they contribute to the segmental enlargement near this level. As Streeter points out for the ostrich, the emu with its large spinal cord must be an animal, the principal nerve control of whose body is supplied from its spinal cord segments simply — "an animal that works chiefly with its primary apparatus." From the fact that, with the exception of the slender caudal extremity, the most slender region of the cord is in the neck, the great mass of its dorsal root axones must be of short course. Indicated by the segmental swellings, it is very probable that a large proportion of the dorsal root fibers of a segment are distributed in that segment alone, making many of the reactions of the simple reflex character. Accordingly, the descending cerebro-spinal fasciculi would be small, as must be the case from the proportions of the cord.

As the efferent or ventral root fibers are absent along the surface of the cord midway between the nerve pairs and begin to emerge and collect rapidly toward the level at which the root is formed, so the cell-bodies giving origin to these fibers show a tendency to cluster toward the middle of the segment. However, this clustering is not so great as might be expected. The ventral columns maintain their shape and contain large cells throughout the segment. For three of the longer segments counts were made of the cells contained in sections through the segmental enlargements and in sections from midway between adjacent nerve pairs. Each count involved 20 transverse sections of 30 ^ thickness and included all cell-bodies having a mean diameter of 10 // and above. The following gives the average number of cells per section found in each of the two localities of these segments:

94 Journal of Comparative Neurology and Psychology.

, ^ f near nerve

1 7th segment < . , , ^

' *' (^ midway between nerves

,, , f near nerve

iQth segment < •, , .

[. midway between nerves

J ^ f near nerve

22nd segment < • , , ^

° ( midway between nerves

17 cells 15 cells 21 cells

18 cells 29 cells 21 cells

In other words, the numbers of the larger cells found'in the twenty sections (an extent of 0.6 mm.) from each locality of the segments were :

f/ear nerve viidtvay between nerves 17th segment 341 cells 303 tells

19th segment 420 cells 365 cells

2:;nd segment 576 cells 432 cells

There were many cells smaller than \o tx. It is very probable that many of the cells having a mean diameter of 10 /i and above belong to association neurones and therefore do not give origin to ventral root fibers at all. The largest cells and those considered as undoubtedly giving origin to efferent fibers gave mean diameters ranging from 25 to 45 //. These, however, while much inore abundant near the nerve, were scattered throughout the segment indicating perhaps that many of the efferent fibers, being drawn toward the level at which the root is assembled, course some distance through the substance of the cord before emerging from it. In longitudinal section most of the largest cells appear fusiform in shape with their long diameter drawn more or less parallel with the long axis of the cord.

The grouping of the cell-bodies in the gray substance is very indefinite. The largest cells appear more abundant in the regions of the ventrolateral and lateral groups. Occasionally a large cell can be found lying in the beginning white substance about the lateral periphery of the gray figure. Frequently one or two cells of the largest type occur in the base of the dorsal horns (Streeter's dorsal group) and perhaps correspond to the nucleus dorsalis or Clarke's columns in other animals.

Throughout the entire thoracic cord and extending into the cervical region, there is found in the lateral periphery of the spinal cord of the emu a line of scattered cells representing the 7121cletis marginalis vmwr described for the birds and reptiles. In the emu it is situated within and against the ventral border

Hardestv, spinal Cord of the Emu. 95

of the ligamentum longitudinale laterale {Ninni, Figs. 2, 3 and 4). It is practically in line and continuous with the nucleus niarginalis major of the lumbar enlargement. Its largest cells are about as large as those of the ventral horn and differ from them in apparently having fewer processes. They are similar in shape to those described by Kolliker ( '02 ) in the spinal cord of the dove. They are embedded m a thin column of gray substance more or less invaded by the surrounding medullated axones. These cells too are markedly segmental in their distribution. Midway between the nerve pairs of adjacent segments they are very sparsely scattered, being found at the rate of about one cell to every four or five sections. In the regions of the segmental enlargements the nucleus is thicker and the cells often occur two or three deep in it (see Fig. 2). Thus, this longitudinal band of cells also takes part in the segmental arrangement of the specimen.

According to Sherrington ( '00 ) this nucleus appears to have been first observed by Gaskell in the alligator in 1885. Gaskell ( '88 ) mentions it in discussing the segmentation of the spinal cord in the Sauropsida (crocodiles and birds) and refers to it as a lateral-surface-group of cells which is as strictly metameric as the cells of the dorsal root ganglia. The nucleus was independently described by Lachi ('89) and Kolliker ('01 ). Kolliker gave it the name Hofmann' sc/ier Kleinkem in distiction from the nucleus marginalis major, which he called Hofmann'scher Grosskern. In another paper Kolliker ( '02 ) makes a comparative study of these nuclei, finding them in the salamander, proteus, lizard, alligator, dove, and also represented in the spinal cord of young dogs. In the dove and the 15-day dog he shows the nuclei distinctly metameric, arranged, especially in the former, in isolated groups of cells one on either side of each segment. In the spinal cord of the emu the metameric groups are not so isolated and distinct as Gaskell and Kolliker describe them in the alligator and dove, there being occasional cells scattered between the segmental enlargements.

The lumbar enlargement with its nucleus marginalis major and the slender conus meduUaris are apparently quite similar

96 Journal of Comparative Neurology and Psychology.

in section to those of the ostrich as described by Streeter. For this reason the description of these parts of the emu cord may be omitted here.

That the segmental enlargements are morphologically so evident in the emu cord is no doubt to some extent due to the great length of its segments. Were the segments as short relatively as they are in the mammals, the swellings would tend to merge one into the other and be less noticeable. However, it must be conceded that morphological segmentation at least is more marked in the spinal cords of birds and their relatives, the reptiles, than in the higher vertebrates. It is of interest as indicating the position of the birds in the animal scale. It is more marked in the still lower vertebrates, and in the invertebrates, the general rule is that the central nerve axis consists simply of a ganglionic chain, distinct ganglia for each body segment, while swellings opposite each pair ot nerves are indicated in the early development of all vertebrates.

Segmental swellings in the spinal cords of the lower verte,brates have be^n long observed. Luderitz ('81) notes that Collins described such enlargements at the levels of the attachment of the nerves in fishes as early as 1685. Later such conditions in the fishes, especially the Triglidae, have been frequently mentioned. Carus ('14) was the first I think to discuss their functional significance. He also described them for the spinal cord of the snake. Luderitz pictures segmental enlargements in the spinal cord of the ringed snake {Tropidonotiis natnx) and also shows that such are indicated in the cervical segments of the rabbit. Further, by comparing transverse sections taken at different levels of the segments, he thought it indicated in the human spinal cord also. Numerous publications have since appeared supporting in general the views of these earlier writers.

The tendency of the ventral horn cells to cluster toward the levels at which the nerves are atached, also has a rather voluminous literature. It is of course very evident in those animals (fishes, etc.) having very marked segmental swellings. Stieda ('68) may be mentioned among the first to study it in

Hardesty, spinal Cord of the Evni. 97

the fishes and birds and Clarke ('51) and Kolk ('59) were among the first to claim such grouping for the higher vertebrates. BiRGE ('82) calls attention to the clustering or pulling together of the ventral horn cells toward the regions of the attachment of the ventral roots of the frog cord and Waldeyer ('88) and Kaiser ('91) discuss similar conditions in the cord of the gorilla and man.


Birge, '82. Die Zahl der Nervenfasern und der motorischen Ganglienzellen

im Riickenmaik der Frosches. Archiv fih- Anat. unci Physiol., Piiysiol.

Ahtlil. Carus, '14. Versuch einer Darstellung des Nervensystems, etc., Leipzig. Clarke, '51. Researches into the structure of the Spinal Cord. Pliilosopli.

trans. Royal Soc, London. Part II. Fiirbringer, '88. Untersuchungen zur Morphologie und Systematiii der Vogel.

Theil I., Amstcrdatn. Gaskell, '8S. The Cranial Nerves. Jour, of Pliysiology (Foster), 10. Kaiser, '91. Die Functionen der Ganglienzellen des Halsmarkes. Haag. Kolliker, '91. Ueber einen noch unbekannten Nervenzellen in Riickenmark der

Vogel. Aiiad. Anzeiger, IVien, Nr. XXV. Kolliker, '02. Weitere Beobachtungen Uber die Hofmann'schen Kerne am

Mark der Vogel. Anat. Anzeiger, 21, Nr. 3. Kolliker, '02. Ueber die oberflachlichen Nervenkerne im Marke der Vogel und

Reptilien. Zeits.f. 7viss. Zool., 72. Kolk, '59. Bau und Functionen der Med. spin, und oblong. Bi-atoiseinueig,

(Quoted by Liideritz). Lachi, '89. Alcune particolarita anatomische del ringonfiamento sacrale del

midollo degli uccelli. Memorie delta Societa Toscana di Scienze Natiirali,

Pisa, 10. Liideritz, '81. Ueber das Riickenmarkssegment. Archizfiir Anat. und Piiysiol.,

Anat. Ahtlil. Stieda, '68. Studien iiber das centrale Nervensystem der Knochenfische.

Zeits.f. wiss. Zool., ^Q. Sherrington, '00. On the Outlying Nerve Cells in the Mammalian Spinal

Cord. Pliilosopli. trans. Royal Soe., London, Series B. Streeter, '04. The Structure of the Spinal Cord of the Ostrich. Am. Jour.

of Anat., 3, No. i. Waldeyer, '88. Das Gorillarlickenmark. Abhandl. der kgl. preuss. Akad. der Wiss. zu Berlin.


By S. J. Holmes.

Contributions from the Zoological Laboratory of the University of jMichigan, No. gi.

It is customary to explain tiie phototactic movements of organisms as dependent upon a sort of forced orientation which is brought about by the fact that when the two sides of the body are stimulated unequally they give unequal muscular responses which result in turning the animal until both sides are stimulated to the same degree, when the creature moves either towards or away from the light in the direction of the rays. But however satisfactory this explanation may be for the phototactic movements of most organisms, there are several cases where locomotion is directed by the rays of light which cannot be accounted for in this way. It is well known that earthworms, leeches, and the larvae of house-flies, blow-flies and many other insects have a strong tendency to shun the light and collect in the darkest regions they can reach. These forms have been cited as affording typical illustrations of negative phototaxis. The often precise orientation of these organisms to the direction of the rays very naturally disposes one to explain their phototactic movements as taking place according to the scheme just mentioned. No one has attempted to work out in detail the exact mode of response in any of these forms, although the fact of their orientation to the direction of the rays of light has been described by several different observers. A variety of explanations may be offered according to the gener

' The descriptive part of this paper with the exception of a few minor changes was read before the section of Animal Morphology of the International Congress of Arts and Sciences at St. Louis, Mo., Sept. 21, 1904.

Holmes, Random Movements. 99

al theory we have referred to. It might be assumed, in the earthworm, for instance, that light causes a greater contraction of the longitudinal muscles on the less illuminated side, or that it inhibits the action of the longitudinal muscles on the side that receives the greater stimulus, or that both these effects are combined. The attempt was made to get some clue to the method of orientation by carefully watching the movements of the animals under the influence of light. It soon developed that what seemed at first a forced orientation, the result of a direct reflex response, is not really such, but that the orientation which occurs and which is often quite defim'te is brought about in a more indirect manner by a mode of procedure which is in some respects similar to the method of trial and error followed by higher forms.

The response of the earthworm to light has been noted by several observers Hofmeister,^ Darwin,^ Graber,^ Yung,^ Hesse,' Parker and Arkin,*^ Smith, ^ and Adams. ^ While all parts of the surface of the earthworm are sensitive to light the directive influence of light, as Parker and Arkin have shown, is greatest at the anterior end of the body and diminishes towards the posterior end. The negative reaction of the earthworm becomes less as the intensity of the light is diminished, and in very weak light the reaction, according to Adams, becomes slightly positive. The directive influence of light was tested by Parker and Arkin and by Adams by placing the worm at right angles to the direction of the rays and noting the number of times it extended its head towards or away from the

'HoFMElSTER, W. Die bis jetzt bekannten Arten aus der Familie der Regenwiirmer, Braunschweig, 1S45.

'^Darwin, C. The Formation of Vegetable Mould through the Action of Worms, with Observations of their Habits, Netv York, 1900.

Graber, V. Grundlinien zur Erforschung des Helligskeits und Farbensinnes der Thiere, Leipzig, 1884. '

  • YUNG, E. Compt. Rend. Trav. Soc. Helv. Sci. Nat. l8g2, p. 127.,

^Hesse, R. Zeit. 7uiss. Zoo/. 1896, 61, 393.

^Parker and Arkin. A/n. Jour. Physiol. 1901,4, 151.

■'Smith, A. C. Am. Jour. Physiol., 6, 459.

  • Adams, G. p. Am. Jour. Physio!., 9, 26.

lOO Journal of Comparative Neurology and Psychology.

light, or simply straight ahead. In all except very weak light the number of negative reactions was found to exceed the positive ones, the proportion being greater, as a rule, the more intense the illumination (Adams). In these investigations attention was mainly centered upon the direction of the response, the directive influence of light falling on different regions of the body, and the effect of light of different intensities. The mechism of the reaction was not especially considered.

If the earthworm is more likely to turn away from the light than towards it, it might be concluded that the light exercises a direct orienting effect upon its movements, but this does not necessarily follow. When exposed to a very strong light (the beam of a projection lantern passed through an alum cell was used) the earthworms experimented with showed the following type of reaction. The light in the first place stimulates the animal to activity, calling forth the regular movements of locomotion. Waves of contraction of the circular muscles pass from behind forward; the anterior end of the body is extended and set down; waves of contraction of the longitudinal muscles follow those of the circular ones and pull the posterior part of the body forward. As the worm crawls it frequently moves the head from side to side as if feeling its own way along. If a strong light is held in front of the worm it at first responds by a vigorous contraction of the anterior part of the body; it then swings the head from side to side, or draws it back and forth several times, and extends again. If in so doing it encounters a strong stimulus from the light a second time it draws back and tries once more. If it turns away from the light and then extends the head it may follow this up by the regular movements of locomotion. As the worm extends the head in crawling it moves it about from side to side, and if it happens to turn it towards the light it usually withdraws it and bends in a different direction. If it bends away from the light and extends, movements of locomotion follow which bring the animal farther away from the source of stimulus.

The experiments of Pakkek and Akkin show that the head of the worm in crawling is more apt to turn away from the light

Holmes, Random Movements. loi

as it extends than towards it. It is probable, although not expressly stated, that account was taken of the first decided turn that was made. If one watches carefully the movements of a worm that is placed at right angles to a strong beam of light it will be seen that the head frequently moves from side to side before extension takes place. These movements may be very slight and ordinarily would escape attention. There is often a similar movement during the process of extention. Frequently the head is bent over towards the light during the first part of the extension and then bent the other way and extended farther, or again it may be waved back and forth several times. Slight trial movements in all directions are continually being made. The reason why the worm makes more turns of a dicided sort away from the light than towards it is largely because the little trials that bring the worm nearer the light are not followed up. Many of the turnings that would naturally be counted as negative are preceded by a slight positive turn followed by a stronger negative one. In order to ascertain whether the negative reaction was manifested at the very beginning of the response the following experiment was tried. A worm was allowed to crawl on a wet board. When it was crawling in a straight line it was quickly lowered into the beam from a projection lantern so that its body would lie at right angles to the rays. The exposure to the light was made in each case when the worm was contracted, and the first detectible movement of the head to one side noted. In the two specimens employed the first detectible turn was away from the light 27 times and towards the light 23 times. After a few extensions the worm in nearly all cases soon turned and crawled away from the light. The first detectible movement of the earthworm seems, therefore, to be nearly as likely to be towards the light as away from it. Tne slight preponderance of negative turns may be due to the fact that some of the smaller trial movements were overlooked, to a slight direct orienting effect of the rays, or merely to chance. A decided turning such as was probably counted in Parker and Arkin's experiments may represent a first trial, or a result of perhaps several very small trials. It is easy to

I02 Journal of Comparative Neurology and Psychology.

see that in strong illumination extensions towards the light are ■checked while those that bring the worm away are followed up, but this is not so obvious in light of weaker intensity. Whether the negative reaction takes place entirely by the selection of random movements is hard to determine with certainty, but there is no doubt that this factor plays a very large part in the process.

Experiments performed by Miss Rhodes and myself three years ago upon the phototaxis of leeches showed that the method of orientation in these forms is, in principle, the same as that of the earthworm. When specimens of Glossiphonia are placed in strong light their locomotor reflexes are set in action. The mode of progression in Glossiphonia differs from that of the earthworm, although in certain fundamental respects the same. The anterior part of the body is extended, the mouth which acts as a sucker is attached, and then the posterior part of the body is brought forward and fastened by the sucker at its caudal end. Then the anterior part of the body extends again and the other movements follow as before In its progress the leech frequently raises the extended anterior part of the body and waves it from side to side as if feeling its way. If the animal turns it in the direction of a strong light it is quickly withdrawn and extended again, usually in anotlier direction. If the light is less strong it waves its head back and forth several times and sets it down away from the light; then the caudal end is brought forward, the anterior end extended and swayed about and set down still farther from the light than before. When the leech becomes negatively oriented it may crawl away from the light, like the earthworm, in a nearly straight line. The extension, withdrawal and swaying about of the anterior part of the body enable the animal to locate the direction of least stimulation, and when that is found it begins its regular movements of locomotion. Of a number of random movements in all directions only those are followed up which bring the animal out of the the undesirable situation.

The phototactic reactions of the larva; of the J/uscidc? ha.\'e

Holmes, Random Movements. 103

been studied by Dr. Loeb' who found that these forms travel away from the hght approximately in the direction of the rays. If an object was held so as to cast a shadow close to their line of locomotion the larvae would crawl close to the margin of the shadow almost directly away from the light. Loeb remarks upon the significant fact that the oral end of these larvse is the region of especial sensitiveness. Larvae which crawled out of the shadow into the direct sunlight were often found to respond vigorously as soon as the anterior half or third of the body was exposed to the stronger illumination. "Das Thier hemmte seine Bewegung und bogden Kopf um, ca. 90-130° nach rechts und links. Kam dabei die Spitze des Kopfes wieder in den Schatten. so kehrte das Thier in den Schatten zuriick. Geschah das aber, wie es meist der Fall war, nicht, so setzte das Thier seine Bewegung in das Sonnenlicht fort. Die Hemmung der Bevvegung aber beim Uebergang aus dem Schatten in die Sonne war keine allgemeine Erscheinung. Meist gingen die Thiere ohne Versogerung aus dem Schatten in die Sonne. Dass die Lichtstrahlen, weche den Kopf treffen, wesenlich massgebend fur die Orientirung sind, geht aus folgender Beobachtung hervor: Hatte ich ein ausgewachsenes Thier auf einem Brett und schob ich das letztere so aus dem Schatten in die Sonne, dass nur der Kopf des Thieres von Sonnenlichte getroffen wurde, so stellte das Thier sofort seine Medianebene in die Richtung der Sonnenstrahlen. Brachte ich den aboralen Pol allein in's Sonnenlicht, so trat diese Orientirung nicht ein. Thiere, denen ich die vordersten Segmente am oralen Pol abgeschnitten hatte, fiihrten auch keine Orientirungsbewegungen mehr gegen Licht aus. Auf solche vivisectorische Versuche indessen, die die Hemmung einer Reizwirkung zur Folge haben, ist wenig Gewicht ze legen" ( /. c. p. 71). I have repeated the experiment of cutting off the heads of the larvae but found that so much of the soft contents of the body flowed out after this operation that the larvae made only indefinite movements and soon died, so the experiment threw no hght upon the problem.

^LoEB J. Der Heliotropismus der Thiere und seine Uebereinstimmung mit dem Heliotropismus der Pflanzen, iVurzburg, 1890.

I04 Journal of Comparative Neurology and Psychology.

Dr. LoEB does not discuss just how the rays of light orient the body of the fly larvae, so we are left to suppose that he would adopt the same theory of orientation in this case that he has applied to other forms, although from his account of the movements of these larvae it would almost seem that he had in mind the method of orientation that we have found to be followed by the angleworm and leech. Observations which I have made upon the phototaxis of blow-fly larvae with the problem of orientation especially in mind soon convinced me that the movements of these forms are directed by light through following up those random movements which bring them away from the stimulus. In normal locomotion the fly larva raises the anterior third of its body, extends it and places it down upon the surface to which it adheres; then the posterior end of the body is pulled up, and the anterior end extended again, and so on. At each contraction the ventral part of the posterior portion of the body is brought forward more than the dorsal and is held in place by means of small projections on the surface while the anterior end is being extended, thus giving the larva a forward impetus. Successive waves of contraction pass from the posterior end and as they reach the anterior part of the body they extend it at the same time that it is raised. The locomotion is a sort of combination of ordinary vermicular movements and the looping movement which is found in the crawling of the larvae of the geometrid moths. As the larva progresses the anterior part of the body is often swayed from side to side. When the head is put down, say to the right, the rest of the body is pulled along in the same direction. The source of progressive locomotion is mainly in the posterior half of the body, the anterior portion being used mainly to set the direction of movement.

When strong light is thrown on a fly larva from in front, the anterior end of the creature is drawn back, turned towards one side, and extended again. Often the head is moved back and forth several times before it is set down. Then it may set the head down when it is turned away from the light and pull the body around. If the head in moving to and fro comes into strong light it is often retracted and then extended again in some

Holmes, Random Movements. 105

other direction, or it may be swung back without being withdrawn. If a strong Hght is thrown upon a larva from one side it may swing the head either towards or away from the Hght. If the head is swung towards the Hght, it may be withdrawn or flexed in the opposite direction, or, more rarely, moved towards the light still more. If it is turned away from the light the larva usually follows up the movement by locomotion. Frequently the larva deviates considerably from a straight path, but as it continually throws the anterior end of the body about and most frequently follows up the movement which brings it away from the stimulus, its general direction of locomotion is away from the light. In very strong illumination the extension of the anterior part of the body away from the light is followed by a retraction, since in whatever direction it may extend it receives a strong stimulus and the larva writhes about helplessly for some time. Sooner or later, however, it follows up the right movement. Occasionally the larva may crawl for some distance directly towards the light, but after a time its movements carry it in the opposite direction. When once oriented the direction of locomotion of the larvae is comparatively straight.

In the animals here described there is, so far I can discover, no forced orientation brought about by the unequal stimulation of the two sides of the body, but an orientation is produced indirectly by following up those chance movements which bring respite from the stimulus. I do not deny that there may be an orientating tendency of the usual kind, but if there is it plays only a subordinate role in directing the movements of the animal. The orientation of these forms is essentially a selection of favorable chance variations of action and following them up. It is a type of reaction differing from photataxis in its typical form. It does not come under the head of photopathy and it differs from Jennings' "motor reflex" by which many of the so-called tropic reactions arc producted in the Protozoa, although in common with the two latter modes of response it may be considered as a form of "selection of over-produced movements. " The light reactions of the forms studied may be inter

io6 Journal of Comparative Neurology and Psychology.

preted as a resultant of two motor responses; first, the activities of locomotion which are set up by the stimulus of the light, and second, the act of jerking back and bending the body from side in response to a strong stimulus from in front. Here are two instincts or reflexes, however we may be pleased to call them, which are in a measure antagonistic in that the first is frequently overcome by the second. The direction of the external stimulus determines which of these two instinctive tendencies predominates. We do not need to assume that the animal consciously selects and follows up the movements that bring it out of a disagreeable situation, and we need not assume that any association is established between stimulus and reaction, even of the most fleeting sort. I do not wish to state dogmatically that such an association is never formed in these organisms, since observations on other forms lead me to speak guardedly on this point; but it cannot, I believe, be more than a minor factor in their phototactic response, if it exists, In so far as the effect of previous experience is not involved, the type of reaction in question differs from the trial and error method of primitive animal intelligence. At the same time, it resembles that method in that a multitude of movements are made of which only certain ones are followed up. It may be said to be a form of the trial and error method minus the element of learning by experience.

The type of reaction we have described is one which is very widespread among the lower animals, and it doubtless enters as one element into many of the tropic reactions which we commonly explain as the result of forced reflexes. The element of spontaneous, undirected activity is one of vast if not essential importance in the life of nearly all animals. The simpler animals profit by their varied experience, although they may not learn, and thus secure some of the advantages which it is generally considered the special function of intelligence to confer.

In a valuable paper which appeared after the preceding portion of this article was written Dr. Jennings' points out the

'Jennings, H. S. Contributions to the Study of the Behavior of Lower Organisms, Washington, 1904.

Holmes, Random Movements. 107

important role played by the method of trial and error in the behavior of the lower organisms, especially the protozoa In the work that has been done on the instincts and reactions of animals in recent years too much stress has doubtless been laid upon the action of the environment on the organism and too little upon the internally initiated actions of the organism itself. Animals are frequently regarded as if they were more or less passive instruments played upon by external agencies and responding in the right way because they are so constructed that they cannot do otherwise. Rather, they are like instruments running by their own inherent energy, like a music box that is wound up and so regulated as to produce a variety of melodies. External agencies press the stops here and there and change the tune. If one tune does not suit, the environment is heard from and the instrument shifts to another.

The method of trial and error in its widest sense is one of those very large categories under which a multitude of varied activities may be subsumed. Even the process of natural selection may be considered a form of it, since all variations may be regarded as trials, and the unsuccessful ones errors. By a little squeezing we might also include many of the phenomena of development and regeneration. In psychology it is comnionly recognized not only as the method of primitive animal intelligence, but as forming an essential element of the process of reasoning in its more abstract forms. Now it is coming to be read back into the realm of instinct and tropisms. In all these fields it is, par excellence, the method of adaptation. Instinctive behavior is either a direct expression of it, or, so far as instincts are stereotyped, indirectly the outcome of it through the principle of natural selection.

The role played by the trial and error method in the behavior of the lower organisms has, as yet, elicited but little com'ment, owing probably to the fact that attention has been centered more upon other features of their behavior. It may have been considered by some investigators as too obvious for remark since anyone who attentively observes the conduct of almost any of the lower animals for ten minutes can scarcely

io8 Journal of Comparative Neurology and Psychology.

fail to see the method exemplified. If he were watching a chick pecking at a variety of objects and giving signs of disgust when it had seized a nauseous substance he would doubtless regard the process as one of trial and error whatever name he might apply to it. A study of the conduct of much lower organisms would disclose many cases almost equally evident. The lives of most insects, crustaceans, worms, and hosts of lower invertebrate forms including even the protozoa show an amount of busy exploration that in many cases far exceeds that made by any higher animal. Throughout the animal kingdom there is obedience to the Pauline injunction, "Prove all things; hold fast that which is good."

The trial and error method is set off by Jennings in sharp contrast to the usual scheme of tropic reaction. "The tropism," he says, "leads nowhere; it is a fixed, final thing, like a crystal." And elsewhere: "This method of trial and error, which forms the most essential feature of the behavior of these lower organisms, is in complete contrast with the tropism schema, which has long been supposed to express the essential characteristics of their behavior. The tropism was conceived as a fixed way of acting, forced upon the organism by the direct action of external agents upon its motor organs. There was no trial of the conditions; no indication of anything like what we call choice in the higher organism; the behavior was stereotyped." (p. 250). If the term trial and error is used in the wide sense here employed I cannot but think that the distinction between the method so designated and the orthodox scheme of tropisms is not, after all, so wide as it at first appears. The motor reaction of Parameciuin is certainly a fixed way of acting brought about almost inevitably by certain factors of the environment. Its behavior is certainly as stereotyped as that of any organism whose reactions are definitely known. Its reactions are "forced movements" in the untechnical sense of this expression, and there is no more evidence of choice in its conduct than in the contraction of a muscle; for we can scarcely speak of choice in a creature that reacts in oneway to all sorts of stimuli. On the other hand the trial and error method may be ex

Holmes, Random Movements. 109

tended to cover the reactions of an animal which orients itself according to the usual scheme. If an earthworm when illuminated more on one side of the body than another simply turned away, by a direct reflex, from the greater stimulus and kept on doing so until the body was brought parallel with the rays we would have a case of orientation according to the commonly accepted theory of tropisms. The creature is forced into line by unequal stimulation of the two sides of its body. When in crawling one side of the body comes to be presented to the light more than the other deviation is corrected by a direct movement away from the stimulus. Getting out of line may be regarded as an error which brings about a certain reflex that sets the animal again upon a straight course. The worm keeps in an approximately straight path because those movements (errors) which bring it out of a straight course are counteracted by a definite reflex, while those movements (successful trials) which bring the creature away from the light are followed up. It is because these errors are corrected that the orientation even according to the ordinary theory, is maintained. These deviations are as much errors as the random movements towards the light in the method of orientation that is actually followed. Tlie stimuli in ordinary tropisms may, however, serve to prevent errors as well as to correct them after they are made. Take an organism that orients itself by the direct method. As it swings out of line it is doing so against the influence of the unilateral stimulus which tends to turn it back during, as well as after, the trial. But the mechanism of preventing the movement and of reversing after it is made is, in this case, the same; only in proportion as the checking predominates does the random character of the movements becomes reduced. Between the behavior of an organism like an earthworm that jerks back and turns to one side when the anterior end is stimulated and that of one which turns directly when the side is presented to the stimulus there is to be sure a maked difference in behavior, but there is an underlying basis of similarity in the two cases in that in both errors are made, although they are corrected in different ways. The end result of both methods is the same, i. e.

no Journal of Comparative Neurology and Psychology.

to get the organism away from the stimulus. In the one case it is accompHshed by a direct reflex, without more ado; in the other only after a considerable waste of energy in inconsequential vermiculations. In orientation according to the usual theory of tropisms errors are made in abundance ; but they are corrected in a more direct and efficient way than in the more or less haphazard method so frequently followed.

By a careful analysis of the phototaxis of Stentor, Euglena, and some other protozoans Jennings has conduced that the oritation of these forms to light takes place according to the trial and error method, and not by the method of simple forced reflexes. The reactions of Euglena are of especial interest since this form apparently shows a combination of both direct and indirect methods of orientation to the direction of the rays. Eiiglena may react to a strong or sudden stimulus from the light by backing off and starting ahead in anew direction. Several trials of this kind may be made until finally the creature becomes oriented when it swims to or from the light according to the intensity of the stimulus. Euglena is also capable of orienting itself by gradually bending its course until it comes to be approximately parallel with the rays. Ordinarily this form swims in a straight spiral path. Should light shine on the body from one side the sensitive anterior end would be stimulated differently in different parts of its spiral course. According to Jennings, it is the diminution of light as the animal turns the anterior end away from the stimulus that causes the motor response. When the Euglena turns so that the anterior end is less illuminated, it is stimulated to swerve back further towards the light, and, by a succession of such responses, it finally becomes oriented to the direction of the rays. Swimming through that portion of its spiral course that causes the diminution of light at the anterior end is that part of the creatures activities that must be looked upon as error, if we go so far as to regard the passing through different sections of a continuous spiral course as trials. But to view the matter in this way is to go far towards obliterating the distinction between orientation through trial and error and orientation by the direct method. In the mode of photo

Holmes, Random Movements. iii

tactic response here considered Etiglena does not react by a number of indiscriminate movements until the right one is accidentally hit upon, but by a direct reflex whose effect is to bring the organism more nearly parallel to the direction of the rays. The phototaxis of Euglena is not so manifestly the outcome of the trial and error method as that of the earthworm. In the latter case light does not cause directly a movement which makes for orientation. The direct response may or may not have that effect. The successful movement is accidentally hit upon, but one can scarcely say this in the case of Euglena in which the orientation takes place more nearly in accordance with the usual scheme.

It is perhaps difficult to decide where best to draw the line as regards the employment of the expression trial and error. If it is extended to include the phototaxis of Euglena and other protozoa where there is a gradual adjustment of the path by appropriate direct responses until it coincides with the direction of the rays, we can hardly stop short of including, at least to a considerable degree, the cases of phototaxis that take place according to the commonly accepted theory. We may regard all departures from the straight and narrow path as errors according to whatever theory of phototaxis we may choose to adopt, and we can look upon all movements in that path as successful experiments. I would suggest that if the term trial and error is widened, as seems desirable, so as to include such reactions as are described in the first part of this paper where there is no discernible element of learning involved, its application be limited to those cases in which the adapted movements may be regarded as chance successes. This would exclude the tropisms of the orthodox kind; it would exclude the gradual orientation of such forms as Euglena where oblique stimulation causes a direct response which brings the body more nearly parallel to the rays. It would include many of the reactions of the protozoa where, as in the phototaxis of the blue Stentor, the right direction of movement is hit upon by chance, and a large part of the actions of higher forms. y\ll organisms make errors. In some cases these errors are rectified by an appropriate direct reflex,

1 1 2 Journal of Comparative Neurology and Psychology.

in others by the chance success of a random movement. There will doubtless occur many cases difficult to classify where trials are not perfectly random movements but where the stimulus may have a certain directive effect which is in large measure obscured. A tropism of the 'direct sort is not necessarily a perfectly fixed and rigid affair. It may be a tendency more or les obscured by a lot of random movements arising from internal causes. An organism may be drawn to a certain point through a direct orienting reflex, but if there be at the same time a large element of random activity in its behavior it may seem to reach that point by the method of trial and error. In the trial and error method the random character of the movements impresses us most; in the tropisms, the element of direct determination by the environment. Both of these factors run through the behavior of all animals, but they are mingled in various proportions in different forms. In the lives of most, if not all, animals both are essential elements in the adjustment of the organism to its conditions of existence.

University of Michigan, Ann Arbor, Mich., Dec. 6, 1^04.


By Walter C. Jones, M. D.

With twelve figures.

The work which forms the basis of this paper was done in 1898-1900, at the Zoological Laboratory of Northwestern University, Evanston, Illinois, and was presented in part fulfillment for the degree of Master of Arts. Circumstances prevented the immediate preparation of the results for publication. In the considerable interval to the present, only one paper, as far as I know, has been published, dealing with the development of the sympathetic nervous system; this paper, by Hoffmann, 1902, is briefly noticed in my review of the literature. The writer wishes to thank Professor Wm. A. Locv, Director of the Laboratory, for invaluable help in the work and in the revision of the manuscript.

The earlier view in regard to the origin of the sympathetic nervous system was that advanced by Remak, to the effect that it arose i?i situ from the mesoblast. Balfour, after his researches on elasmobrauch fishes ('78), brought forward a new view, namely, that the sym pathetic nervous system arises from the epiblast iu connection with the spinal and with certain of the cranial nerves. He claimed that the sympathetic ganglia of the trunk region "are at first simply swellings on the main branches of the spinal nerves." Subsequently, these swellings are removed each from its respective nerve, retaining, however, fibrous connections with the nerve through a short branch, which forms a ramus comnninicatis . They appear at first to be independent, becoming united later by commissures, and forming a continuous cord on either side.

1 Contribution from the Zoological Laboratory of Northwestern University, William A. Locy, Director.

114 Journal of Comparative Neurology and Psychology.

ScHENK and Birdsell published, in 1879, '^he results of their observations on certain mammalia, which, as Balfour says, "seem to indicate that the main parts of the sympathetic system arise in continuity with the posterior spinal ganglia; they also show that in the neck and other parts the sympathetic cords arise as a continuous ganglionic chain."' Onodi ('86), working on elasmobranchs, agrees essentially with Balfour, and gives excellently clear figures intended to show that the cells which form the sympathetic cord arise as outgrowths from the spinal ganglia.

Paterson, in 1891, revived the idea of the mesoblastic origin of the sympathetic nervous system. According to his researches on mouse, rat, and human embryos, the earliest traces of the sympathetic are seen as a cellular cord lying in the mesoblast between the aorta and the cardinal vein, in the anterior dorsal region. This cord is bilaterally symmetrical, and is composed of cells which he claims are differentiated mesoblastic cells. At the time of its appearance, the cord has no connection whatever with the spinal nerves nor ganglia, and constitutes the anlage of the sympathetic. The next step is the formation of the rami communicantes, which arise as fibrous outgrowths passing from the spinal .nerves to the sympathetic cord. The ganglia appear next, and are formed at the points where the rami join the cord, resulting presumably from the growth of both sympathetic cells and nerve fibers. The collateral sympathetic is developed by the outgrowth from the sympathetic cord of cellular branches, which later give rise to the ganglia, nerves, and plexuses. In this category are placed the cervical and sacral portions of the sympathetic chain and also rather doubtfully, the grey rami communicantes. This view is repeated by Paterson, in 1903, in Cunningham's Anatomy.

Marshall ('93), on the other hand, agrees very closely with the theory advanced by Balfour. In frog and chick embryos, Marshall finds that the sympatetic nervous system arises "as a series of outgrowths from certain of the cranial and from all of the spinal nerves. These develop ganglionic swellings,""' which later become connected by fibrous commissures, thus forming the gangliated chain of the adult.

His, Jr. ('97), tracing in the chick the history of an anlage similar to that described by Paterson, finds, in very early stages, that the cells forming it come from the ganglia of the spinal nerves, thus confirming

'^^^ Comparative Embryology.' Vol. 2, p. ^84. In yertebrate Embryology.' Page IJ4.

Jones, Development of the Sympathetic. 1 15

the researches of Onodi ('86), 011 elasmobranchs. In the human embryo, His, Jr. finds that the development of the sympathetic begins (in a 10 mm. embryo) with the outgrowth of the white rami communicantes from the spinal nerves. A little later, the sympathetic cord appears and is joined by the rami. These findings contrast with those of PatER8ON ('91), who observed that the structure first to appear is the sympathetic cord, the rami developing later.

Hoffmann's observations on the sympathetic system of selachians ('99) agree, as far as essential features of development are concerned, with those of Balfour and Onodi. Like them, he shows that the cells forming the sympathetic aulage arise in connection with the spinal nerves, thus favoring the view that the sympathetic is epiblastic in origin. In urodeles, he finds ('02) the first trace of the sympathetic occurring as scattered cells connected by slender rami to the ventral branches of the spinal nerves. He does not express an opinion as to whether these cells are of epiblastic or mesoblastic origin. In 30 mm. salamanders, the sympathetic has come to be a continuous chain, in some places fibrous, in other places cellular in character, extending from the first spinal nerve back to the tail region and connected by rami to all of the spinal nerves.

Another question relates to the connection, during embryonic stages, between the sympathetic anlage and the adrenals. These two structures come into very intimate relation with each other during their development. In reptiles, there arises on each side of the vena cava soon after its formation a longitudinal cord of cells, which Minot calls the "mesenchymal anlage" of the adrenals. On the dorsal side of this anlage and somewhnt toward the median line, appear clusters of cells, which are derived from the sympathetic ganglia. They constitute the "sympathetic anlage'" of the adrenals.

These two portions come in contact, and at first, in amniota at least, the sympathetic part grows more rapidly, and partially surrounds the mesenchymal portion ; but soon the relations become reversed, and the mesenchymal portion gradually surrounds the sympathetic ; what finally becomes of the latter is not known. Balfour ('78) found in elasmobranchs, in the case of the posterior adrenals, that each of these bodies shows a small sympathetic ganglion attached to either end of it, the whole structure being attached to a spinal nerve by a ramus. This mass of cells gradually becomes divided into a ganglionic and a glandular portion ; the latter acquiring a mesoblastic investment becomes adrenal, while the former develops into sympathetic tissue. Hoff ^" Human Embryology." Pages 48 j and 4.86.

ii6 Journal of Comparative Neurology and Psychology.

MANN ('99), working on selachians, confirms the findings of Balfoi'^r, in their essential features.

A brief summary of the literature embodies the following points: (i) The cells from which the sympathetic arises, probably come originally from the spinal ganglia, as shown particularly by Balfour ('78), ScHENK and Birdsell ('79), Onodi ('86), and His, Jr. ('97), Paterson ('91) being the most recent investigator to hold to the mesoblastic theory. {2) In elasmobranchs, the sympathetic begins with the development of the ganglia, which arise as cellular outgrowths on the spinal nerves. These gradually are removed from the nerves, retaining, however, fibrous connections, which constitute the rami The commissures arise later, probably as outgrowths from the sympathetic ganglia. (3) In mammalia and in aves, the sympathetic appears first as a continuous longitudinal cord, which later is joined by rami communicantes, arising as outgrowths from the spinal nerves. At the points where the rami join the cord, ganglia develop, while the intervening portions of the cord remain as commissures. (4) The collateral sympathetic arises as outgrowths from the sympathetic cord. (5) A portion of the sympathetic anlage has, during its development, very intimate relation with the adrenals. (6) From an embryological as well as from an adult morphological and physiological stand-point, the sympathetic system is best considered as an integral part of the whole nervous system and not as a structure that is to any considerable extent distinct or separable from the rest of the nervous system.

My own observations upon the development of the sympathetic nervous system have been confined mainly to the derivations of the ganglia, their commissures, and the rami communicantes. The observations were made upon toad and frog embryos, and were limited to the region posterior to the vagus ganglion. Graphic reconstructions were made from serial transverse parafifine sections, supplemented by actual dissections. For the sake of clearness, a general statement is given before proceeding to the description of particular stages of development:

(i) The sympathetic system between the vagus ganglion and the second spinal nerve, in the animals studied, appears to arise from cells scattered in the mesoblast of the region where the future sympathetic cord is to lie. A portion of these cells gradually becomes aggregated to form a cord (Fig's. 2, 7, and

Jones, Development of the Sympathetic. 1 1 7

8, Sy.^ which is the beginning of the sympathetic as an anatomically distinct structure in this region. Scattered cells are to be seen among the fibers of the first and second spinal nerves, all the way up to the spinal ganglia. These cells become continuous below with the cells of the forming sympathetic cord. Also, there is a connection between the sympathetic cord in front of the second nerve and the group of cells which give rise to the sympathetic in the trunk region. Tiiese facts of observation favor the view that the cells which enter into the formation of the sympathetic cord anterior to the second spinal nerve migrate either (a) downward from the ganglia of the first and second spinal nerves or (b) forward from the structure which gives rise to the sympathetic of the trunk region, posterior to the second spinal nerve. However, the question of the origin of these cells requires further investigation.

(2) My preparations show that all of the sympathetic posterior to the second spinal nerve arises from an antecedent structure, which is bilaterally symmetrical, and lies closely applied to the dorsal and external side of the aorta, where the latter is double, and to the outer side of the median aorta, after the union of the two trunks. This antecedent structure consists of an irregular ridge of cells (Fig. i, Ri.) extending back to the region between the ninth and tenth nerves (Fig. 7, Ri.^, suffering interruption, however, in the region of the third. Throughout its entire extent, this ridge is continuous inferiorly with a mass of cells (Fig. i. Ma.), which, between the second nerve and the anterior end of the kidney, lies between the aorta and the Wolffian duct (Fig's, i, 3, 4, and 5, Ma.) and, in the region of the kidney, between this organ and the aorta (Fig. 6, Ma ). It extends some distance anterior to the second nerve, where, however, its connection with the sympathetic is uncertain. This ridge (Fig. 8, Ri.) is continuous anteriorly with the sympathetic cord between the vagus ganglion and the second spinal nerve, and arises at the same time as this portion of the cord. The division of the forming sympathetic into a portion anterior to the second spinal nerve and one posterior to it is observed throughout the following descriptions.

ii8 Journal of Comparative Neurology and Psychology.

Nine Millimeter Stage. — Between the vagus ganglion and the second spinal nerve, at this stage, sympathetic cells are seen scattered in the mesoblast dorsal and internal to the aorta (double). These cells are round or slightly elliptical in shape, have a granular appearance, and are devoid of processes; they stain a little more deeply than connective tissue cells, and usually are very easily distinguished from them. At places, espe

Fig. I. — Transverse section of 9 mm. toad, through the ganglion of the sixth spinal nerve (see Fig. 7), showing the sympathetic ridge, Ri. (The section passes through the extreme anterior end of the kidney on one side and just anterior to the kidney on the opposite side.) Ao. /., Aorta (median dorsal). B. P^., blood-vessel. C/i., notochord. A'., kidney. Afti., mass of cells between aorta and kidney or aorta and Wolffian duct. Jlfns., muscle. /?t., sympathetic ridge. JV. D., Wolffian duct. X 246. Camera Lucida. Reichert, oc. 2, obj. 7 a.

cially in the immediate region of the first and second nerves, these cells are becoming aggregated to form a cord (Fig. 7, Sy.^, which, in cross section, appears very irregular, and shows from three to twelve cells in one section. (Fig. 2. Sy. shows a little later stage.) This cordis the beginning of the differentiation of the sympatJietic i?t this region. It becomes continuous back of the second nerve with the ridge already mentioned.

The mass of cells between the aorta and the Wolffian duct

Jones, Development of the Sympathetic. 1 19

(Fig's. I and 7, Ma.^ with which the sympathetic ridge is closely connected, is rather small and ill-defined anteriorly. In the region of the fourth nerve, however, it becomes more distinct (Fig. 7), and graduall}' increases in size from here back to the anterior end of the kidney, where it lies in contact externally with this organ instead of the Wolffian duct. Fig. i, Ma. shows this mass of cells. Between the aorta [Ao. /.) and the Wolffian duct {W. D.), on the left side, it represents the condition anterior to the kidney; between the aorta {Ao. i.) and the kidney {K.), on the right side, it shows the condition in the region of the kidney.

Midway between the third and fourth nerves, a few cells lying close against the side of the aorta are seen to form a rather poorly defined ridge (Fig. i, Rt.), which can be traced back without interruption to the region between the ninth and tenth nerves (Fig. 7, Ri.). It has not yet made its appearance in the region of the third nerve, but traces of it are seen immediately behind the second. The cells of this ridge-like elevation constitute the earliest structure to ivJiich I have traced the sympathetic system in the region posterior to the second spinal nerve. The source of the cells constituting the sympathetic back of the second spinal nerve is still uncertain. I have no definite clue as to their origin, outside of the fact that these cells mingle freely with the nerve fibers as far back as the ninth. This close connection between the nerves and the sympathetic ridge suggests, at any rate, that the cells are of epiblastic origin.

At this stage, then, the following points are especially to be noted: (i) Between the vagus ganglion and the second nerve, scattered cells are becoming aggregated to form the sympathetic cord (Fig. 8, Sy.^. (2) In the region posterior to the second nerve, the ridge-like structure from which the sympathetic arises (Fig's, i and 7, Ri.^ extends from the second nerve back to the region between the ninth and the tenth, being interrupted in the region of the third. (3) The cells of the forming sympathetic mingle freely with the fibers of the first nine spinal nerves.

Twelve Millimeter Stage. — In this embryo, between the vagus ganglion and the second spinal nerve, we find the sympa

120 Journal of Couiparative Neurology and Psychology.

thetic occurring as an almost unbroken rod of cells (Fig's. 2 and 8. Sy.).

Behind the second nerve, the sympathetic ridge is much Vviffber and much better defined than in the preceding stage ;

/^«^. 2. — Transverse section through a 12.5 mm. toad, midway between the first and second spinal nerve (see Fig. 8), showing cells becoming aggregated to form the sympathetic cord, {Sy.). Ao. 2., aorta (of one side, where double). Ch., notochord. In., intestine. Sp. Ca., spinal canal. X 246. Camera lucida. Reichart, oc. 2, obj. 7 a.

the outer cells of it are becoming more closely compacted and, in some places, a definitely limiting membrane has been differentiated. This group of cells is the beginning of the diffcrcntiaiio?i of the sympathetic cord in the body region. In some places, this cord includes a part or all of the ridge, this being especially noticeable anteriorly, just in front of the point where the median dorsal aorta begins (Fig. 8, *). Fig. 3 is taken in this region, and shows nearly the whole of the ridge being formed into the sympathetic cord iSy.).

The cellular structure with which the sympathetic ridge is connected, remains the same as in the preceding stage, except for a general increase in size and an enlargement, in the region of the fourth nerve (cf. Fig's. 7 and 8, Ma.). This enlarge

Jones, Development of the SympatJietic.


ment persists in later stages, being distinctly noticeable in i8 mm. embryos, and is due to relationship with another structure, this relation suggesting that between the sympathetic and the adrenals described by Minot.^ The relation of the spinal nerves to the developing sympathetic system is practically the same as in the 9 mm. stage, except in the case of the sixth, where the cord is somewhat removed from the nerve, retaining distinctly, however, short fibrous connections. These fibers constitute the beginning of a ramus communicans (Fig. 8, 7?.).

Fig. J. — Transverse section of a 12.5 mm. toad, between the third and fourth spinal nerves (see Fig. 8), showing nearly the whole of the sympathetic ridge being transformed into the sympathetic cord, Sy. Pg., pigment. Cf. neck of cells marked t with Fig. 5, Rt. Other abbreviations, same as in ffig's. i and 2. X 246. Camera lucida. Reichert, oc. 2, obj. 7 a.

Fourteen Millimeter Stage. — At this stage, the sympathetic cord is easily traced between the second spinal nerve and the vagus ganglion, except immediately behind the first nerve (Fig. 9, Sy.y., connection with the vagus ganglion is uncertain.

Posterior to the second nerve, the ridge now has become much more prominent. In many places, it becomes high and narrow (Fig's. 6 and 10, Ri.) before the sympathetic cord is differentiated at the free border of it (Fig's. 5 and 6, Ri. and Sy.^.

'"Human Embryology." Pages 485-489.

122 Journal of Comparative Neurology and Psychology.

In other regions, the cord is seen very early, including the greater part of the ridge (Fig. 3, Sy.), and appears to be pushed outward and upward by the ridge, as the latter becomes better developed; while more frequently the cord is not seen distinctly till the ridge has grown to about half the size to which it finally attains. Between the second nerve and the point where the median dorsal aorta begins (Fig. 9, *), the ridge lies against the aorta (Fig. 5, Rt.). In the region of the kidneys, it extends outward and upward, lying between this organ and

/T;^. ^. — Transverse section of an 18.5 mm. toad, through the third spinal nerve (see Fig. 10), showing the relation between the sympathetic cord, Sy., and the third spinal nerve, S/>. N. j. The dotted parallel lines indicate the nerve as reconstructed from sections just anterior to the one here represented. For development of ramus, cf. Fig. 12. Other reference marks, same as in preceding figures. X246. Camera lucida. Reichert, oc. 2, obj. 7 a.

the notochord (Fig. 6, Rt.). In the structure with which the sympathetic ridge is connected (Fig. 9, Ma.), two changes need to be noted ; first, a general increase in size and second, the obscuring of the sympathetic ridge in the region of the fourth nerve, due to relationship with a structure referred to in the preceding stage as being possibly adrenal.

In regard to the relation between the spinal nerves and the sympathetic, we find short rami in connection with the fourth.

Jones, Development of the Syjnpathetic.


fifth, sixth, and seventh nerves and an incipient one on the eighth (Fig. 9, R.). The relations of the other nerves are the same as in the preceding stage, except that the cord is better defined now, and, in section, has the appearance of lying against the nerve, there being fewer cells scattered among the nerve fibers. (Fig. 4, Sy. shows a little later stage.) At this stage, a membrane is found in many places, entirely surrounding the cord and making it perfectly distinct. We can distinguish now clearly for the first time two regions where the develop

Fig. ^. — Transverse section through an 18.5 mm. toad, just posterior to the third spinal nerve (see Fig. 10), showing the sympathetic ridge, Ri., and the sympathetic cord, Sy., well differentiated along the free border of the ridge. Reference marks, same as in preceding figures. X 246. Camera lucida. Reichert, oc. 2, obj. 7 a.

ment of the sympathetic as a whole has proceeded father than elsewhere. These places are in the region of the brachial and the lumbar plexuses. Also, the sympathetic is usually better developed in the immediate region of the nerves than it is between them (Fig. 9).

Eigliteen Millimeter Stage. — Between the vagus ganglion and the second nerve, the sympathetic is easily traced as a large, distinct, and continuous cord (Fig. 10, Sy.). Posterior to the second spinal nerve, the ridge (Fig's. 5 and 6, Ri.^ now is well

124 Journal of Comparative Neurology and Psychology.

marked, especially in the region of the lumbar plexus, where it has reached the height of its development (Fig. lo, Rt.). The angle which it makes with the horizontal plane varies in different regions. Anteriorly, it extends upward and inward (Fig. 5, Rt.), approaching even a horizontal position in some places. Following it backward, in the region where the median aorta begins (Fig's. loand li, *), it assumes rather abruptly a vertical position, and then immediately begins to lean outward, till, in the region of the kidney, it points upward and outward (Fig. 6, Rl), making an angle with the horizontal plane varying from 30° to 60°.

fig, <5. — Transverse section of 18.5 mm. toad, between the sixth and seventh spinal nerves (see Fig. lO), shovi^ing the sympathetic ridge, A*/., and the sympathetic cord, Sy., well differentiated at the top of the ridge. Abbreviations, same as in preceding figures. X 246. Camera lucida. Reichert, oc. 2, obj. 7 a.

The sympathetic cord (Fig. 6, Sy.) is larger and more definite, and is lound surrounded by a membrane more frequently than it is seen without one. The cord lies at the top of the ridge, including in a few places the greater part of the latter. The rami communicantes, at this stage, have increased in size and distinctness. Besides those in connection with the fourth,

Jones, Development of the Sympathetic. 125

fifth, sixth, seventh, and eighth nerves, there is also an incipient one on the ninth.

Twenty-one Millimeter Stage. — In a toad 21 mm. long, the sympathetic cord is continuous from the vagus ganglion back to the region between the ninth and tenth nerves. But it is to be borne in mind that the development back of the second nerve is different from that in front of this point. The ridge has mostly disappeared. A considerable portion of it remains, however, between the fourth and fifth nerves and smaller parts, posteriorly. Compare Fig. 10, Ri., where the ridge is well developed, with Fig. II, Ri., in the region of the eighth and ninth nerves, where the ridge has almost completely disappeared.

Between the vagus ganglion and the second spinal nerve, the sympathetic cord is removed a considerable distance dorsally from the aorta. (Cf. Fig. 2, Sy., where it lies almost upon the aorta, Ao. 2.) The mass of cells anterior to the kidney, with which the ridge was connected in earlier stages (Fig's. I and 7, Ma.), is still present, but is flattened posteriorly, where it occurs as a horizontal plate-like mass of cells extending laterally from the aorta. As the ridge disappears, the sympathetic cord is left alone, large and distinct, lying most of the way upon this mass of cells but not continuous with it and removed laterally a considerable distance from the aorta. In the region of the kidney, the cords lies on the dorsal side of this organ. The sympathetic cords of the two sides can be traced back as far as a point midway between the ninth and tenth nerves, where they unite underneath the aorta.

Beginnings of rami are found in connection with the first and third nerves ; a small one is seen on the second, while rather distinct ones appear on the fourth, fifth, sixth, seventh, eighth, and ninth nerves (Fig. ii, R.). The sympathetic cord is surrounded by a well-developed membrane, throughout most of its extent. The cord is composed now not of cells alone, for most of the way it shows fibrous structure also. Furthermore, it is better developed in the immediate region of the nerves than it is between them (Fig. 11, Sy.), a condition only slightly noticeable in earlier stages. Between the nerves, the cord is rela

126 Journal of Comparative Neurology and Psychology.


Fig. 8

^tg- 7- — Diagrammatic reconstruction of the sympathetic nervous system of a 9 mm. toad, dorsal view. The kidney and the sympathetic with its related structures are represented as having been shifted from a position inferiorly along side the aorta (see Fig. i) to a position on a level with the spinal cord, but removed laterally a distance equal to that between the sympathetic and the spinal cord before the change of position. The more closely dotted portion represents sympathetic tissue; posterior to the second spinal nei>ve, the 'a'idtli of this closely dotted portion represents the height of the sympathetic ridge. The lightly dotted area represents the structures with which the sympathetic is connected during its development. X marks the anterior end of the adult kidney. The connection between the sympathetic and the spinal nerves is not shown. * marks the anterior end of the median dorsal aorta. H. A'., head kidney. V. IV., fourth ventricle. A'., vagus ganglion. /., 2., etc., ganglia of first, second, etc. spinal nerves. For other abbreviations, see Fig's. i-6. X 21.

Fig 8. — Diagrammatic reconstruction of the sympathetic nervous system of a 12.5 mm. toad. The representation of the rami communicantes here and in succeeding figures is exceedingly diagrammatic. For abbreviations and other explanations, see Fig. 7. x 28.

Jones, Development of the Sympathetic.


Fig. g. — Diagrammatic reconstruction of the sympathetic nervous system of a I4 mm. toad, i^i'., ramus communicans ; dotted line lettered 7?., ramus only partially developed. Other abbreviations, same as in Fig's. 7 and 8. X 28.

Fig. 10. — Diagrammatic reconstruction of the sympathetic nervous system of an 18.5 mm. toad. The ridge, Ri., which is figured as extending inward horizontally throughout its entire extent has, in the actual specimen, an upward and inward d\x&c^\oxi anteriorly (see Fig. 5, Ri.) and an upward and outward direction posteriorly (see Fig. 6, Ri.). Reference marks, same as in Fig's. 7, 8, and 9. X 24.

128 Journal of Comparative Neurology and Psychology.

tively slender, this condition being especial!}- marked in the region of the kidney. The cellular enlargements of the sympathetic cord in connection with the nerves now may be called ganglia, and the constricted portions between the nerves, which are becoming distinctly fibrous in structure, may be called commissures.


Fig. II. — Diagrammatic reconstruction of the sympathetic nervous system of a 21 mm. toad. Reference marks, same as in previous figures. X 24.

Jones, Development of the Sympathetic.


Condition after the Metamorphosis. — By the time the tail has almost disappeared, the sympathetic cord not only is completely separated from its antecedent structures, but is removed dorsally and somewhat laterally from the aorta. In the mid-trunk region, it lies as high as the upper border of the notochord (Fig. 12, Sy. G. 2.), while in the preceding stage it was on a level with the lower border of it.

Fig. 12. — Transverse section through the second spinal nerve, from a toad in which the tail has almost completely disappeared. Col. Sy., collateral sympathetic. R., ramus. Sp. G. 2., ganglion second spinal nerve. Sp. N. 2., second spinal nerve. Other abbreviations, same as in preceding figures. For development of ramus, cf. Fig. 4. X 246. Camera lucida. Reichert, oc. 4, obj. 3.

The ridge and the structure with which it was connected, have now atrophied almost completely. Nearly all of the rami communicantes have become much longer, and the collateral sympathetic also is well developed (Fig. 12, Col. Sv.), while the differentiation of ganglia and commissures is now almost complete.

130 Journal of Comparative Neurology and Psychology.


(i) In the region between the vagus ganglion and the second spinal nerve, the sympathetic arises in a comparatively simple and direct manner : cells, probably of epiblastic origin, scattered in the mesoblast gradually become aggregated (Fig. 2, Sy. ) to form a cellular cord. This is similar to the process in mammals described by Paterson, except that he finds the cord is at first entirely independent of any other structure, while in the toad the cells of the sympathetic cord lie in contact with the fibers of the first and second spinal nerves, from the earliest stages.

(2) In the region posterior to the second spinal nerve, the origin of the sympathetic system is more complex. It appears first as a small ridge of cells (Fig. i, Ri.) lying close along side the aorta. The cells at the top of this ridge, as it becomes higher, are differentiated to form the sympathetic cord (Fig's. 5 and 6, Ri. and Sy.). Later, the ridge disappears entirely, leaving the cord free, save its connections with the collateral sympathetic and with the spinal nerves (Fig. 12, Co/. Sy. and R.).

(3) The sympathetic ganglia and commissures arise directly from the sympathetic cord. The latter becomes enlarged in the region of the nerves, forming the ganglia, while the portions between the nerves become reduced, forming commissures, which, immediately after metamorphosis, are composed largely of nerve fibers. (Fig. 1 1 shows the enlargements and the constricted portions.) These findings contrast somewhat with the observations of Balfouk on elasmobranchs. where he finds that the ganglia arise entirely independent of each other, the commissures appearing later as outgrowths from the ganglia.

(4) The rami communicantes arise in the toad in the same manner as in elasmobranchs (Balfour). The cord (the ganglion in elasmobranchs) lies in contact with the nerve from the very first (Fig. 4, Sy.). Later it gradually is removed from the nerve, retaining, however, fibrous connections, which constitute the ramus. The rami appear earliest in the mid-trunk region, my preparations showing the first one in connection with the

Jones, Development of the Sympathetic. 1 3 1

sixth nerve .(12 mm. stage). The first and the ninth are the last nerves to develop rami.


For publications prior to 1878, seeONODi, Arck.f. Anat. u. Enhvickeliingsch. '85, and Patersox, Philosophical Transactions, '91. Balfour, F. M.

'78. Monograph on the Development of Elasmobranch Fishes. Pp. 172 173 and 239-244. '81. Comparative Embryology. Vol. 2, pp. 358-386 and 548. Hertwig, Oscar.

'92. Text-book of Embryology. P. 462. His, W., Jr.

'97. Ueber die Entwickelung Bauchsympatheticus beim Hiihnchen und Menschen. Arch. Anat. iind Physiol. A7iatom. Abteihing, pp. 137171. (Supplement.) Hoffmann, C. K.

'99. Die Entwickelungsgeschichte des Sympathicus bei den Selachiern (Acanthias vulgaris). Verhand. Akad. Wetcnsch. Amsterdam. Tweede Sectie, Deel 7, No. 4. '02. Die Entwickelungsgeschichte des Sympathicus bei den Urodelen. Verhand. Akad. Wetensch. Amsterdam. Tweede Sectie, Deel 8, No. 3. Huber, Carl G.

'97. Lectures on the Sympathetic Nervous System. Jour. Comp. Neurol., Vol. 7, pp. 73-145Marshall, A. M.

■9',. Vertebrate Embryology. Pp. 134, 274, 386-387, and 538. McMurrich, J. P.

'03. The Development of the Human Body. Pp. 441-452. Minot, C. S.

' 92. Human Embryology. Pp. 630-63? and 485-489.

'03. Laboratory Text-Book of Embryology. Pp. 194-195, 324, 242244, and 267-268. Onodl.

' 85. Arch. f. Anat. 11. Entwickelungsgesch.

' 86. Ueber die Entwickelung des sympathischen Nervensystems. Arch. f. Mik. Anat., Vol. 26, Erster Theil, pp. 553-';9i. Paterson, A. M.

'91. Philosophical Transactions, Vol. 181, pp. 159-186. '03. In Text-Book of Anatomy, edited by D. J. Cunningham. Pp. 673-674. Schenk and Birdsell.

' 79. Ueber die Lehre von der Entwickelung der Ganglien des Sympa.. theticus. Mittheil. aus d. cmhrvolosc. Instit. Wi-en.



Even a superficial and casual examination suffices to convince one that the commonly recognized varieties of activity — reflex, instinctive, voluntary, habitual, etc., — are intergrading types, not sharply separated classes. Concerning the fact of the existence of clearly definable types of action there is no dispute ; concerning their genetic relations there is surprising lack of agreement. Some maintain, with Spencer, that activity developes from the simple to the complex, from reflex action, through instinctive, to voluntary action. Others, especially those who have attended more carefully to the development of action in the human subject than in the animal kingdorh generally, with equal assurance insist that the true and primary course of development is from the relatively complex, variable, and apparently voluntary act to the reflex, automatic and habitual. It is my purpose in this discussion of the subject to try to show that these two courses of development are supplementary rather than contradictory, that phylogeny presents us with facts which favor the former view, ontogeny with facts which favor the latter.

For present purposes it will suffice if we select as essentially important types of action the reflex, the instinctive, and the voluntary. An examination of typical acts to which we should unhesitatingly apply these terms indicates that the types may be defined in terms of the complexity and variableness of action. The reflex act is simple and uniform ; the instinctive act is more complex as well as more variable ; the voluntary act is either simple or complex and extremely variable — unique. These facts are conveniently expressed in the following simple classification :

Editorial. 133

( Uniform (Reflex, Automatic, Habitual) Simple Acts ■' Variable (Instinctive)

( Unique (Voluntary)

( Uniform (Habitual, Automatic) Complex Acts < Variable (Habitual)

( Unique (Voluntary) Furthermore, it will be convenient for us to consider that single uniform response to a simple external or internal stimulus which is called a reflex as the unit of activity. Then, according to the degree of variableness, the same relatively simple act may be classed as reflex, instinctive, or voluntary. The frog when it responds to a touch near the nostril by raising its fore leg as if to brush away the stimulating object, for all that we can tell from a single reaction, may be reacting in any one of the three ways. Only by careful observation of a series of reactions to the same situation can that information which is necessary for the classification of the act be obtained. If with each repetition of the stimulus the same reaction occurs immediately and with no perceivable variations, we class it as a reflex. If the series presents acts which, although essentially the same, vary slightly from one another, so that only the general form of the reaction is predictable, we apply the term instinctive. And, finally, if from time to time new acts appear in the series, so that the reaction can never be predicted, the acts are voluntary. As a usual thing, however, instinctive and voluntary actions are more complex than reflexes, and analytic study reveals that they are really made up of a number of simple reflexes closely associated. It was the discovery of this that led Loeb' to speak of the instinctive act as a chain-reflex.

Possibly it is worth while to present an illustration of these types of action and their differences as exhibited by the frog. Suppose the conditions of reaction to be the following. To the stimulus of a touch on the nose the animal reacts by, i) a leg movement, or 2) ahead movement, or 3) locomotion, or 4) a combination of the three. Then, according to the degree of variableness of the reaction series, the reaction would be classed as:

'Comparative Physiology of the Brain and Comparative Psychology, p. 178.

134 Journal of Comparative Neurology and Psychology.

I. Reflex, if, for example, the reaction series read i-i-i-i-i-i or 2-2-2-2-2-2 or 3-3-33-3-3 etc. II. Instinctive if the reaction series read i, 2— 1, 2—1, 2—1, 2 etc. In this series of numerals differences in weight indicate slight variations in the emphasis of the various units of the action. III. Voluntary if the reaction series read i, 2, 3 — 2, 3-3-2-1, 3 etc. Leaving difference in complexity out of consideration for the moment, the reflex act is uniform, the instinctive variable, the voluntary unique.

'Having now defined our reaction types, we may note that comparative studies show that animals differ with respect to their activities in two fundamentally import mt directions : first, in the number and variety of the reactions which they are -capable of giving in response to a particular stimulus-complex (situation); and second, in the rapidity with which they are able to narrow down the trial and error series of reactions to those acts which are appropriate to the situation. From one point of view this is learning, from another it is the mechanizing of activity. The former of these differences may be known as the scope of reaction; the latter, as ability to profit by experience. Organisms in markedly differing degrees, according to their complexity of form and environmental relations, follow the injunction "prove all things ; hold fast that which is good."

Of chief interest in this connection is the recently discovered fact that even the simplest organisms are not exceptions to this rule, for Jennings has noticed that Paramecium, Vorticella, Stentor and many other unicellular forms exhibit the trial and error method of reaction ; and more recently Holmes^ has described the same type of reaction in the earthworm and other invertebrates. These discoveries prove the falsity of the prevalent notion that simple organisms exhibit single simple stereotyped reactions which are suitable in the long run, but show no signs of adaptation or even of

' The selection of Random Movements as a Factor in Photota.vis. This Journal, 15, loo.

Editorial. 135

such trial and error reactions as are common among more complex animals. It would appear rather that all animals more or less thoroughly prove or test their environment by the method of trial and error. Sometimes we find a wide range of reactions, as in the cat, dog and monkey ; and again we find a narrow range, as in the unicellular organisms. Organisms differ, then, not in the essential method of reacting, but, first, in the scope or variety of reactions exhibited in the proving of things, that is, in the thoroughness with which they are capable of testing a situation ; second, in the influence which the tests have upon subsequent reactions. According to Jennings and other students of simple forms of animal behavior, the unicellular organisms go through much the same series of trial and error reactions each successive time a situation is presented to them ; they do not in any marked degree profit by their experience. Higher organisms, on the contrary, give unmistakable evidence of profiting by their proving of things, for from the chance stumbling upon the right act which is characteristic of the trial and error method at first, they progress to the point at which the right reaction is given as soon as the situation is presented : they learn, by experience, the appropriate form of reaction. Some animals quickly lay hold upon that which is good and tenaciously hold to it, others profit scarcely at all by the repetition of a reaction. Certain animals continue to stumble upon the right reaction from the beginning to the end of life, others become able to select the profitable reaction. In general the statement is true that those organisms which mosc nearly prove all things are also best able to hold fast that which is good, but as will be shown later there are certain striking and significant exceptions to this rule.

If, now, we seek for the developmental relations of the two aspects of activity which have been selected as primarily significant — scope of the reaction series and ability to learn, with its resultant mechanization of action — we note that the scope of reaction is primarily a phylogenetic aspect, ability to learn an ontogenetic aspect. For in the race, generally speaking, activity developes from the simple uniform series of trial reactions to

136 Journal of Comparative Neurology and Psychology.

the complex variable series ; there is change from generation to generation in the number and variety of reactions exhibited in the trial and error series. At times these changes are not in the direction of an increase in the scope of reaction, for in certain cases regress in one direction conditions progress in another : the developmental process is not simple and continuous. In the individual, on the other hand, repetition of a reaction brings about mechanization ; habits are formed ; the number and variety of reactions which were exhibited early in the life of the organism are diminished later by the avoidance of those which are useless or harmful Plasticity in activity is an aspect of phylogenetic development ; fixity in activity, of ontogenetic development.

Precisely such a point of view with reference to aspects of activity as is here taken necessitates a justification of the separation of phylogeny and ontogeny. Why should we speak ot one aspect of actionas characteristic of phylogenetic development, another of ontogenetic ? The answer to this question is found in the history of activity. For if we choose a type of organism which is capable of associative learning we find that any individual of the species during its life exhibits two kinds of action: first, reflex, instinctive and voluntary acts which are not shaped by experience ; second, instinctive and voluntary acts which are modified by experience, and which in case of the latter give rise to habits. The first represents the hereditary ability of the individual ; the second, the acquired ability. In the offspring of the individual there appear hereditary activities similar to those of the parent, but not exactly like them, for no two individuals have precisely the same form. There are more or less marked variations, and for the development of activity these variations are of great importance. The reflexes of a species appear in generation after generation with but slight and usually unnoticed changes ; the instinctive acts show greater differences, the old ones are modified and new ones appear, possibly as the result of the inheritance of certain of the tendencies to action which have been established as habits in the parent ; the voluntary acts are strikingly different, unique. It is here that variation

Editorial. 137

brings about those changes which determine the course of the development of activity in the race, changes in the scope of reaction. The trial and error series may be longer in the offspring than in the parent; if so we speak of the animals greater scope of reaction and of its greater originality or initiative. The connecting link between ontogenetic and phylogenetic development has to be sought in the value of habits to the race. If the results of individual experience are transmitted, it is clear that a species may gain a larger fund of instincts while at the same time becoming capable of a wider range of reactions. Variations form the basis of changes in the scope of reaction; the inheritance of acquired tendencies to action accounts for changes in instinctive action. Each new generation may possess new instinctive acts, in addition to new possibilities of voluntary action.

But the animal kingdom presents, instead of uniform increase of plasticity and fixity (scope of reaction and mechanization) with increase in complexity of structure, divergent lines of development. Certain animals are markedly plastic or voluntary in their behavior, others are as markedly fixed or instinctive. In the primates plasticity has reached its highest known stage of development ; in the insects fixity has triumphed, instinctive action is predominant. The ant has apparently sacrificed adaptability to the development of ability to react quickly, accurately and uniformly in a certain way. Roughly, animals might be separated into two classes : those which are in high degree capable of immediate adaptation to their conditions, and those which are apparently automatic since they depend upon instinctive tendencies to action instead of upon rapid adaptation.

Attention to these and other striking differences in behavior should enable us to make a valuable classificatian of animals on the basis of activity. Certainly it is apparent, and it can not be better exhibited than by reference to such facts as those under consideration, that there has been no one direct course of development in animal behavior ; rather, there seem to be several directions of development, of which the present highest stage of one is marked by the insecta, that of another by the primates. Robert m. yerkes.

138 Journal of Comparative Neurology and Psychology.


In the directed reactions of lower organisms we find repeated attempts by various authors to distinguish in one way or another two great classes of phenomena, and for these two classes different terms have been used. Thus Rothert^ distinguishes two sorts of taxis, strophic and apobatic, depending on the manner in which they are brought about. Strophic taxis corresponds to the typical taxis or tropism of Verworn or Loeb. It is conceived as brought about by a turning of the organism toward or from the source of stimulation, due to a difference in the intensity of action of the stimulus on the two sides or ends of the organism. Apobatic taxis is defined as produced by a backward movement at passing from one concentration to another.'* The typical case for this is the reaction method of bacteria. Rothert includes under apobatic taxis the reaction of Paramecium and other infusoria to chemicals, etc., though for this his definition is inadequate, since the backward movement is only a comparatively unimportant detail of the behavior, and is often omitted.^

Pfeffer^ rejects Rothert's terms, substituting for strophic taxis the word topotaxis, for apobatic taxis the word phobotaxis. Massart^ applies the term taxis only to the phenomenon called by Rothert strophic taxis, while the word phobism is proposed for the backward movement called by Rothert apobatic taxis.

Davenport" makes in his discussion of the reactions to light a distinction corresponding to a certain extent with those already mentioned. He distinguishes reactions whose direction is determined by the direction of the rays of light, from re

^ Flora, 88, 1901, p. 393.

"^Rothert, I. c. p. 393.

'See the account by the present author, this Journal, 14, 1904, pp. 458-460.

  • Pflanzenphysiologie, Bd. 2, 1904, p. 755.

^Annaks de r Institute Pasteur, 1901, p. 25. '

^Experimental Morphology, vol. 1, 1897, p. iSo.

Editorial. 1 39

actions whose direction is determined by differences in intensity of illumination. The former he calls phototaxis, the latter photopathy. Photopathy is held to be due to the fact "that increased brightness causes a movement forwards, that a diminution of brightness causes a movement backwards, or vice versa, thus resulting in an accumulation of the organisms in the darker •or lighter parts of the field." (1. c. p. 210.) Phototaxis on the •other hand "includes that active migration of organisms whose direction is determined by that of the rays of light" (p. 180). In elongated organisms Davenport thinks that phototaxis is •due to the difference in intensity of illumination of the two sides of the organisms (p. 209), while in Amoeba he thinks it possible that the direction taken by the ray in passing through the organisms is the determining factor (p. 210). Yerkes' holds, with Holt and Lee, that both phenomena are due in some way to the intensity of the light, and classes as taxis those reactions in which the direction of movement is determined by orientation, while -pathy includes reactions in which the movement "is not definitely directed through the orientation of the organism." This distinction doubtless indicates the real observational source •of the attempt to separate two great classes of reactions, and •does not attempt to make the definition outrun the known facts. Kinesis is a term which seems to have been first used by £ngelmann^ for the increase or decrease of activity produced by •certain agencies. The fact that certain bacteria increase or de•crease movement in the light he called photokinesis. Rothert (/. c, p. 374) accepted the term kinesis for such changes in the amount of activity produced by chemicals, calling this chemokinesis. Loeb^ had observed similar phenomena and applied to them the German word Unterschiedsempfindlichkeii. Gar key,* Nworking with Loeb, later substituted the term kinesis for the German word. The term UnterscJiiedsempfindliclikeit, signifying literally sensitiveness to differences, might well include much more ihan a mere increase or decrease of activity ; it could properly

'Mark Anniversary Volume, 1903, p. 361. '^Arch.f. d.ges. Physiol., 30, 1882. ^Arch.f.d.ges. Physiol., 54, 1893, P- '^l^^Am. Journ. Physiol., 3, 1900, p. 29.

140 Journal of Comparative Neurology and Psychology.

embrace all reactions due to a change in the intensity of action of a stimulating agent. Gakrev's use of the term kinesis seems a little uncertain, since he seems in one place (p. 292) inclined to consider it equivalent to Davenport's -pathy. thus including directed reactions to changes in intensity, while in general he seems to limit it to increase or decrease of activity, in the sense in which it was employed by Engelmann and Rothert. GarREY uses the term in contradistinction to tropism, which is defined in accordance with the well known views of Loeb.

Perms ending in -metry have been contrasted with taxis, especially by Strasburger' and Oltmanns,^ to distinguish those features of the reactions that are determined by differences in the intensity of the stimulating agent. Photometry is employed by Strasburger to signify especially the phenomenon that some organisms move toward tiie source of light in a certain intensity of illumination, while they m^ove away from it at a higher intensity. By Oltmanns photometry is used of any reactions due to differences in the intensity of light ; "movements produced by light of different intensity could fittingly receive the name photometric movement" (Oltmanns, /. c, p. 190).

Thus we find various attempts to distinguish by one criterion or another two great classes of reactions. On the one hand we have apobatic taxis (Rothert), phobotaxis (Pfeffer), phobism (Massart), kinesis (Engelmann, Garry and others), -pathy (Davenport), -metry (Strasburger and Oltmanns); in these orientation is not a marked feature. On the other hand we have tropism or taxis proper (Loeb, Davenport, Massart, Verworn and others), strophic taxis (Rothert), topotaxis (Pfeffer); in these orientation is a marked feature. What is the precise basis for this distinction into two classes?

In the behavior of infusoria (ciliates and flagellates) it is possible to distinguish clearly the two classes of phenomena on which this distinction is based, and to determine the real nature of the difference.'^ In both classes the cause of reaction is. in

^Jena. Zeitschr. f. iVaturw., 12. 1878. Flora, 75, 1892, p. 183. •'In the effects of the electric current the usual reaction method is in the Ciliata interfered with by the forced cathodic reversal of the cilia, — a factor not par

Editorial. 1 4 1

the infusoria, some change in the relation of the organism to the environment. It is usually, or perhaps universally, a change in the intensity of action of a stimulating agent on the body as a whole or on its most sensitive portion. The organism, having been in one condition, passes to another, and it is the transition that acts as the effective stimulation. In one of the two classes of reactions to be distinguished, the effective stimulation is due to the fact that as the organism progresses from one region to another it meets different conditions, and to the changes thus caused it reacts. This is the case under natural conditions with the reactions to mechanical stimuli, to heat and cold, to chemicals, to osmotic pressure, and to light when passing to a region of greater or less intensity of illumination. The cause of reaction is analogous to that in our own case in passing from a region of moderate temperature to a hot or cold region. The differences determining reaction are here arranged along the axis of the course, so that it is the backward or forward movement that brings them into action. The reaction is a change of movement such as to carry the organism successively in many directions — a series of trial movements. As soon as one of these movements carries the organism away from the stimulating agent — that is, in such a direction that the changes to which it subjects the organism lead toward the optimum instead of away from it — the reaction ceases, since the cause for it- has ceased. The organism therefore continues in that direction. The position of the body has little or no effect on the production of the change that causes stimulation, or on the release from stimulation. The organism might retreat from the stimulating agent backward or forward or sideways ; in the one case as in the other it would be relieved of the stimulating changes. The different individuals may swim away directly or obliquely, their

alleled under any other conditions. As a result of this interference the reaction of the ciliates to the electric current takes in many respects a different character from the rest of their behavior. I wish therefore to emphasize the fact that the general relations set forth in the text do not apply to the reactions of the ciliates to this agent. Save for the forced cathodic reversal of cilia the response to the electric current would fall in our second class of reactions, as it actually does in the flagellates and rotifers.

142 Journal of Comparative Neurology and Psychology.

various paths crossing at all sorts of angles ; the only requirement is that the path shall on the whole lead away from the greatest (or least) intensity of the stimulating agent. Hence when many specimens react in this way their paths need not be parallel, and as a rule no marked orientation results. When a common orientation of many individuals does occur, it is produced through "exclusion," — through the fact that under the given conditions movement in any direction but one causes the changes which act as effective stimuli, so that all are forced to move in that direction. The behavior of Oxytricha in reacting to heat, as shown in Fig. 7 of the author's paper on Reaction to Heat and Cold* will serve as a type for this reaction method. It is, so far as the infusoria are concerned, the basis for apobatic taxis (Rothert), phobotaxis (Pfeffer), and -pathy (Davenport), and might farther be classed, from certain points of view, as kinesis or -metry.

The second class of reactions includes those in which the changes that act as stimuli are brought about by a swerving toward one side or the other, while the movements in the axis of progression have no such effect. In the infusoria movements from side to side are of course a regular part of the locomotion. In the reactions to the effects of water currents, of centrifugal force, of gravity, and of light rays coming from one side, the lateral movements of unoriented animals induce marked changes. In a water current or under the action of a centrifugal force, or of gravity, lateral movements meet with less resistance in one direction, greater resistance in another direction, and these changes of resistance act as stimuli. In light coming from one side the sensitive anterior end is more illuminated as the organism swerves towards one side; less illuminated as it swerves toward the other, and these changes act as stimuli. The reactioa is the same as in the first class ; the organism changes its course by a series of trial movements. It continues these movements, till it comes into a position in which it is no longer subjected to the changes that act as stimuli. Such a position is found only when the axis of the course coincides with the direction of ac

'Carnegie Institution of Washington, Publication i6, 1904, p. 16.



tion of the impinging force ; in otlier word, when the organism is oriented. The anterior end will be directed toward the source of the stimulating agent or away from it, according as it is the decrease or increase in intensity that is the effective stimulus^ The transference of the organism along the axis of progress has in these cases no effect on the relation of the organism to the stimulating agent, so that so far as this reaction is concerned backward progression might take place as well as forward. Under these stimuli a certain orientation of the organism, not a certain direction of progress, is the essential result of the reaction. This class of reactions forms the basis, so far as the infusoria are concerned, for taxis proper (Massart, Davenport, etc.), strophic taxis (Rothert), and topotaxis (Pfeffer).

Thus in the first class of reactions the essential result is a certain direction of progression — toward a region of greater or of less intensity — while in the second class the essential result is the orientation. But in both classes the nature of the stimulus is the same and the reaction is the same. The stimulus is some change in the relation of the organism to the surrounding conditions — a change in the intensity of action of some agent. The response is a motor reaction that consists of a series of trial movements. The response continues in each case until the effective stimulus comes to an end, then the usual motion is resumed. In the first class this does not produce orientation (save sometimes by "exclusion"), because stimulation comes to an end without it. In the second class it does produce orientation, because stimulation does not come to an end without it. The difference between the two classes depends on the peculiar difference in the distribution of the stimulating agent, not upon different ways of reacting on the part of the organism. The reaction is by "trial" movements continued till a cessation of the effective stimulation is brought about ; this will lead to orientation or not, just as the conditions require.


'For details as to this relation, see this Journal, 14, 1904, pp. 470, 472, and 478 ; Carnegie Institution, Publication 16, 1904, p. 60.

F44 Journal of Comparative Neurology and Psychology.


In the whole field now awaiting the student of mental life in animals, there is nothing of more cardinal interest and importance, both for animal psychology in itself and for general psychology, than the group of problems centering about the development of the conative life. And of these problems none promise results so significant in themselves or so generally illuminative of the mechanism of mind, as the problem of instinct. The question of the origin of instincts ranks as a classic. Many prominent naturalists and psychologists have taken a distinctly anti-evolutional view of the matter, claiming intelligence and volition as the source of instincts. They make a good case ; at least there are facts of introspection which point unmistakably in this direction. On the other hand, the facts at our command point to an evolution of mental processes as well as of physical structure. Instinct is clearly the conative life of lower forms. They seem to have no other conative life. Higher forms have both instinctive and intelligent conation. But it is still an open question whether the instinct is the result of the automatization of volitional processes, or whether it is the unconscious reaction which has been selected by the environment and is thus the precursor which prepares the way for volition.

The adaptation to the environment, through complex and yet very fundamental instincts, as well as the fact of habit formation in the individual life, make for the view that c/iotcc is a factor in the formation of instinct. On the other hand, the general order of evolution and development would place the simple, the instinct, before the more complicated volition. Volitions may be comparatively simple and instincts may be quite complex. But the general plan of the instinct marks it as an earlier form than the volition. We certainly have the indubitable evidence of introspection for this order in ontogenesis. And we see the same order in the mental development of the child. The same is true in the phylogenetic series; but the further we go from a ner

Editorial. 145

vous system like our own, the greater becomes the merely analogical element, and the less trust we can place in the inference.

But indisputable factual evidence can be brought to bear for the definite answering of this question of the origin of instinct — as to whether instinct in its simplest lines originates through the automatization of volitional acts or not. That is, we can obtain definite data concerning the functional and genetic relations of instinctive and volitional processes in animals which have both. The method we propose is simply this : — Take an animal with well defined but comparatively simple instincts. Make a careful study of its instincts, both as they come forward in the growing animal and as they function in the adult animal. Then devise a set of experiments, for the express purpose of testing the intelligence of the adult animal, its power and capacity for dealing with new conditions. In this way it will be quite definitely ascertained whether the present intelligence is such as would have sufficed for the determination of the behavior of the same animal which now constitutes its instinct. For it is a reasonable and necessary assumption that if instinct arose thus through the automatization of volitional acts, the present intelligence of the animal should be equal to dealing as efficiently with similarly complex situations. For the function of automatization is not to do away with intelligence by rendering it useless, but to make possible its further advance by providing for the ready performance of that which has been learned already. This is the way it does work in our own development.

Of course great care and skill will be required in devising the experimental means of this test and in carrying them out. Such criticisms as those made by Professor Mills {Science, N.S., 191, 745.) of Dr. Thorndike in particular and of the "laborators," as he calls them, in general, must be kept carefully in mind. We must assure ourselves in every possible way, as by checks and counter-checks and all possible variations, that the stimulus is acting upon the animal in the way that we think it is, that it is really attending in animal fashion to the situation as we are thinking the situation, and that it is really feeling the

146 Journal of Comparative Neurology and Psychology.

impulses which we think it is feeHng. In other words we must use every means to ascertain the mental factors of the situation before offering any interpretation of the meaning of the behavior, or using the observed behavior to strengthen any particular theory of psychic functional relations We cannot hope for absolute certainty of the processes in other minds by any known method, and the certainty is less the further we recede from ourselves. But the careful student of the habits of the animal in its natural condition, by exercising care to have the animal in as nearly natural conditions as the experimental necessities admit of, and by being careful to isolate the experimental conditions for the animal, will be able to interpret the movements and incipient movements in mental terms and thus have a close approximation to accurac)' in grading the intelligence of the animal as shown in the particular situation. Accurate and detailed knowledge of the natural life of the animal, it cannot be insisted too strongly, is of the highest importance both for the proper planning of experiments and for the interpretation of the results.

If under these conditions it can be shown that an animal now has a grade of intelligence sufficiently high to serve as the guide in shaping any and all of the instinctive modes of behavior with which it is equipped, it has not been shown thereby, of course, that intelligence was the means of its development. The development of mind may have brought fourth this grade of intelligence at a later stage in the race history than that at which the instincts took their rise. In this case the results would be merely negative. If, though, it can be shown that a given animal has, as its mental equipment, an intelligence too low to deal with a situation as complex as that dealt with by the instinct, the inference is then clearthatthe instinct has come by some other way than by intelligence. For our major premise, as before, is that intelligence as a whole never retrogrades, but when released through automatization proceeds to greater complexity and higher organization through dealing with more complex situations. This case would presuppose a retrograding and would therefore be impossible. And so it was not con

Editorial. 147

scious (I do not mean self-conscious) choice which determined originally the mode of reaction which is now instinctive.

This position being established, the way would then be open for the psychologist to construct and maintain a theory of the origin of volition and cognition as developing out of a system of organic behavior. This negative evidence, if it can be found, will contribute to a constructive theory of the systematic development of mind as a progress from the simpler and less organic to the higher and more organic. For intelligence is both of these as compared with instinct.




Worcester Insane Hospital.

The last few years have witnessed a marvelous impetus and great evolution in our ideas of anatomical and chemical facts concerning the central nervous system. The cerebro-spinal fluid has also shared in this advance, partly as the result of independent study, but chiefly because of its intimate relation to the nerve structures. The literature on the subject is scattered through the various journals and reports, but outside of Thudichum's book on the chemical constitution of the brain of man and lower animals, and of Sicard's recent work on the cerebro-spinal fluid, there has been no serious attempt to summarize the recent publications on the chemistry of the nervous system, both in normal and pathological states. It seems opportune, therefore, to review the most available facts along these lines, and point out in what manner these may be utilized to stimulate further research. Of course, in the limits imposed by a review, only the most important papers can be abstracted. The work on neuro-chemistry has been pushed in several directions, of which the principal ones merit the following subdivisions :

1 . More systematic attempts at classifying the various substances of the brain.

2. The isolation of nerve products by newer methods.

3. The further ultimate analysis of these products.

4. The chemistry of the finer structure of the nerve cells, including the micro-chemistry of various stains.

5. The autolysis of brain tissue.

6. Analysis ot the brain and cord in various normal and pathological states.

7. The chemistry of the cerebro-spinal fluid, particularly as a medium for the various products of nerve degeneration.

8. The physiological action of various normal and pathological nerve products.

The substances so far isolated from the brain follow. The classijication is merely a tentative one and must not be looked upon as

CoRiAT, CJionistry of the ?■«'€} vous System. 149

final. It is partly adopted from Thudichum and Koch and partly the result of my own observations.

A. Water.

B. Proteids.

1. A globulin coagulating at |7° — 50° C.

2. A globulin coagulating at 70° C.

3. Xeuroglobulin, coagulating at 45° — 50° C. (According to Halli BrRTON it is the coagulation of this substance in the nerve cells, which is the physico-chemical cause of death from hyperpyrexia).

4. Neurostromin (in traces).

5. Nucleo-proteid (Levexe, Halliburton, Schkarin).

6. Neurokeratin (KtJHNE and CHITTENDEN).

7. Albumoses and Peptones.

C. Extractives.

1. The purin bases of the nucleo-proteids (Adenin, guanin, hypo xanthin) .

2. Pyramidin bases (thymin and cytosin).

3. Amido-acids (leucin and tyrosin).

4. Urea (also a normal constituent of the cerebro-spinal fluid).

D. Carbohydrates.

1. Inosite.

2. An unknown carbohydrate which gives the furfurol reaction and is

probably related to the pentoses.

3. Galactose (a split-product of cerebrin).

4. Dextrose (in the cerebro-spinal fluid) — CoRlAT.

5. Pyrocatechin (in the cerebro-spinal fluid Pj—Hallburton.

6. Glycogen (?),

E. Acids.

1. Nucleic acid.

2. .Sarcolactic acid.

3. Lactic acid (optically inactive ethylidene lactic acid. This is the

result both of post-mortem decomposition and of long continued activity in the central nervous system, as in convulsive seizures. In both cases, the acid passes into the cerebro-spinal fluid).

4. Glycerophosphoric acid (a decomposition product of lecithin).

5. Formic acid.

6. Acetic acid.

7. Succinic acid.

8. Stearic, oleic acid, palmitic acids (decomposition products of lecithin).

9. Litho- and Butophosphoric acid (Thudichum).

F. In Pathological Conditions.

1. Cholin (a decomposition product of lecithin — Mott and Hallibur ton, Donath, Wilson, Coriat).

2. Neurin.

3. Uric acid.

4. Kreatin.

5. Trimethylamin (a post-mortem product).

150 Journal of Comparative Neurology and Psychology.

6. Neuridin.

G. Inorganic bodies. Sodum, potassium, ammonium, calcium and iron, present either as dissociated ions, or in organic combination. The radicles CI, SO^, PO4, and CO3 are also present as disociated ions or with the cations above mentioned.

H. Lecithins (Monophosphatids containing one nitrogen molecule in the proportion N:P^i:l). The various lecithins from eggs, barley, malt, yeast and brain show variations in the amounts of phosphorus and in the methyl content.

1. Lecithins (Stearylolely, margeryloleyl, palmityloleyl).

2. Amido-lecithins (amido-myelin). Contains two nitrogen molecules


3. Kephalins (P:N=i:l).







Assurin (N:P:=2:2).

4. Amido — Kephalins (amido-kephalin P:N^i:2). I. Cerebrins.

1 . Phrenosin

2. Kerasin (Thudichum)

3. Amidocerebrinic acid — glycosid (Bethe)

4. Phrenin

. Cerebrin acid

Cerebrin acids


Cerebrin phosphoric acid (Bethe) J. Cholesterin (occurs both free and in the form of an ester). K. Sulphur compounds (Cerebusulphatides and sulphatids of Thudichum). L. Amidofats (Krinosin and Bregenin).

As a contrast, Thudichuim's classification alone should he consulted (Die chemische Konstitution des Gehirn.s des Menscheu und der Tiere, pp. 89-91), and the reader will also find there a minute discussion of these various products, with particular relation to their chemical reactions.

It is unnecessary to enter here into the details of the newer methods for the isolation of the various brain .substances. The original papers must be consulted, of which reference is given in the bibliography afifi.xed to this review. It suffices to mention Thudichum's complete isolation methods, the acetone method of Koch for the separation of lecithin, kephalin and cerebrin, Levkxe's methods for the isolation of nucleic acid and the uucleo i)roteids, Koch's modification

Cor I AT, Chemistry of the Neri'oiis System.


of the method of Herzig and Meyer for the quantitative determination of lecithin by an estimation of the methyl groups, Halliburton's method for the detection of cholin in the cerebro-spinal fluid and blood, and Coriat's method for the detection and quantitative estimation of cholin in nerve tissue.

The first serious attempt at the quantitative analysis of the brain in normal and diseased individuals, was that of Gutinkov, who estimated water, fixed substances, nitrogen, sulphur and phosphorus, both in the white and grey matter. His series comprised 15 human foetal brains at about the third month, the brains of seven people who died suddenly without any previous illness, and of various mental and somatic diseases. These latter included three cases each of carcinoma and syphilis, two each of pulmonary tuberculosis, neurasthenia and senile dementia, four cases of acute alcohol poisoning, seven of chronic alcoholism, and one case each of post-epileptic dementia, idiocy, postmaniacal dementia, general paralysis, epilepsy, syphilitic pseudo-paralysis, mania and stuporous melancholia. A summary of his various findings as compared with the averages for normal brains, is readily seen from the following table. For the normal human brain, as the average of seven analyses, the following figures are obtained, and it is with these that the pathological material is compared.

Water (gray matter) 84.62

Water (white matter) 69.73

Phosphorus (gray matter) 1.2128 (dry) 0.1877 (moist)

Phosphorus (white



37S (dry)



Nitrogen (gray matter)

8.9082 (dry)


•3692 1


Nitrogen (white matter)

5.580S (dry)




Sulphur (gray matter)


6821 (dry)

.0927 1


Sulphur (white matter)


5250 (dry)

.1589 1


"^ a, <A

c« -^






3 3 cu



3 « 

43 V


51) ■


a. >,






















Acute alcoholic poisoning



Chronic alcoholism





Post-epileptic dementia



















Post-maniacal dementia









General paralysis






Senile dementia











Syphilitic pseudo







Mania tranquilla










Senile dementia





(N ^ normal. -(

= increased,

— = decreased).

152 Journal of Comparative Neurology and Psychology.

The figures themselves, for which the original article must be consulted, are certainly far from convincing, the pathological material especially showing some glaring con traditions; for instance, the difference in the two analyses of neurasthenia and senile dementia. Furthermore, the figures on the percentage of water in general paralysis, are not as high as the marked pial oedema of this disease would lead us to believe.

Thudichum's figures are the result of the working over of a large amount of material and of complex elaborated methods. For the gray matter of the human brain he gives the following results :







Ether extract (with kephalin, lecithin and cholesterin) i-95o

Cerebrosides, cerebrinic acid and myelin


Lecithin, kephalin and myelin (out of the



material) 0.7S0



Lactic acid



Sulphuric acid and neuroplastin


Sulphuric acid in extract

not estimated

Phosphoric acid






Watery extract


^or the white matter of human brain







Ether extract (kephalin, lecithin and cholesterin) 1 1 .497 Cerebroside and myelin 6.910

Lactic acid 0.0456

Inosite 0.2171

Alkalies (as carbonates) 0.1717

The following figures were established for the entire human brain, as the sum of results from separate analyses of the two hemispheres, cerebellum, mid-brain and medulla, gray and white matter. The figures are given in grammes, and relate to the entire weight of the brain without membranes (1296 grammes).

CoRiAT, CJicniistry of the iXeJi'ous System. 155



White matter


White matter

(insoluble in ether)


White matter

(soluble in ether)


Buttery material









24 2(



Lactic acid










Koch's figures were obtained from the brain of an epileptic. The analysis was conducted with great care and with the aid of more recent and accurate methods.



us callosuDi.






Simple proteids












Inorganic salts






Kephalin and m







Phrenosin and Kerasin



Cerebrin acids






Sulphur compounds



The wide differences in the figures of these two investigators show how far we are from establishing a comple quantitative composition for the normal human brain. Such figures can only be obtained as the average of a large number of analyses, and they must be firmly established and fully available before we can hope to make any comparison with pathological brain material.

The recent studies that have been made on pathological brains and cords, have relate! mainly to diseases where there is extensive general or focal destruction of nerve tissue, as in general paralysis and hemiplegia, comparisons with microscopical findings being made at the same time. W. Barrett, as the result of the analyses of the brains and cords in general paralysis, found an invariable increase in

154 Journal of Comparative Neurology and Psychology.

the water while phosphorus was always slightly decreased, but this decrease seemed to bear no relation to the extent of fiber degeneration or cell chromatolysis. Noll, in an experimentally degenerated sciatic nerve, found a decrease of phosphorus and of the alcoholic extract. C0RL\T found choliu in the peripheral nerves, brain and cord of a case of alcoholic polyneuritis, the amounts being parallel to the extent of the Marchi reaction. The presence of the alkaloid here must be looked upon as the result of autolytie processes in the lecithin.

MoTT and Barrett, as the result ot analysis of two cords from cases of hemiplegia, concluded, (i) that on the degenerated side of the cord a breaking up of the phosphorized fat occurs, (2) the amount of lecithin present is diminished, (3) fat is present in excess, (4) the amount of extractives soluble in ether is increased, (5) the proteid residue is diminished, (6) the phosphorus is diminished, (7) the ether extract has a buttery apjjearance instead of being crystalline. Halliburton has shown that the cause of death from hyperpyrexia is a physico-chemical one, being due to the coagulation of the cell-globulin, and when this constituent of cell protoplasm is coagulated, the protoplasm as such is destroyed. The temperature for this reaction is 47° C, but temperatures as low as 42° C will have the same effect if continued sufficiently long.

So far there have been published only two studies on the important question of the autolysis of brain tissue, those of Levene and C0RLA.T. The former established the presence of an intracellular proteolytic enzyme in fresh brain tissue, and which acted only in an acid medium. Corla.t, working from the standpoint of the genesis of cholin in the central nervous system, on account of its importance as a toxic agent in producing the convulsive seizures of epilepsy and general paralysis, concluded that there exists in the brain an enzyme capable of splitting lecithin into glycero-phosphoric acid, stearic acid and cholin, the latter being detected and quantitated by nieans of its platinum compound. This enzyme acted only in neutral and slightly alkaline media, and like all enzymes it could be destroyed by heat.

Hatai, in studying the effects of starvation on the brains of white rats, found that the ab.solute weight in his starved groups was diminished, and there was an actual diminution of the water and an increase in the relative amount of extractives, from which he infers that the protein substance must have been mostly affected. It was therefore expected that changes in the cell bodies would be found, but no marked alteration of the Nissl bodies could be demonstrated. The same writer, as the result of injection and feeding experiments with

CoRiAT, Chemistry of the Nervous System. i 5 5

lecithin on white rats, to show its effect upon their growth, coucludes as follows : White rats which receive! the lecithin either by injection or feeding, gained in body weight more rapidly than those which did not receive it, the gain in the experimented rats being on the average ■60 % greater than in the controls. The relative weight of the central nervous system in the lecithin rats was normal and contained the same proportion of water and solids as in the controls, this being another indication of the normal character of the growth The peripheral nerves in both groups also grew normally, as shown by the relative area of the axis cylinder to its sheath. The action of the lecithin appeared to be that of a stimulating agent for normal growth, and the rats which received it showed a greater power of resistance to unfavorable changes in their surroundings.

Koch's work on the action of electrolytes on colloidal solutions of kephalin and lecithin is also of interest, as showing the possibilities of the study of metallic poisons on the various constituents of the central nervous system, from the standpoint of physical chemistry. These colloidal solutions, which form perhaps the nearest artificial approach to the substratum of the living cell, are precipitated l)y divalent katious, while precipitation is prevented by univalent and trivalent ones.

The most important constituent of the individual cell is the Nissl bodies, which consist principally of nucleic acid. This acid is not found pure, but in combination with proteids forming nucleo-proteids which are probably the most complex bodies in a living organism. These compounds have the property of acids, contain phosphorus and iron and also the purin and pyrimidin bases. There is also a carbohydrate, the exact nature of which is unknown, but which gives the furfurol reaction and probably belongs to the pentose group. The purin bases of cerebronucleo-proteid are adenin and guanin, and the pyrimidin bases, according to Levene, thymin and cytosin. The amount of phosphorus is lower than in the nucleo-proteids from other tissues. Under pathological conditions, especially in general paralysis, this nucleo-proteid passes into the cei-ebro-spinal fluid ; the proteid in the normal fluid consists almost entirely of serum-globulin, nucleo-proteid being completely absent.

During the last few years the cerebro-spinal fluid has called forth an immense literature, partly because it may contain the products of nerve degeneration (cholin, glycerophosphoric and lactic acids, lecithin ?), and partly because in conditions of acute or chronic inflammatory processes in the brain, certain cells may appear in the fluid, the finding of which is a valuable diagnostic aid. The chromo- and cyto-diagno

156 Journal of Comparative Neurology and Psychology.

sis of the fluid do not concern us here. The chemistry of the fluid will be touched upon, only so far as it relates to the products of myelin decay. The physical chemistry of the fluid (freezing point, etc.), the various proteids and the reducing body are all of interest, but their consideration would exceed the scope of this review.

The most important chemical substance which passes into the cerebrospinal fluid is cholin, which is the nitrogen-containing methyl group of lecithin, and can be looked upon as one of its autolytic decomposition products. In the normal fluid it is found in mere traces, but in the fluid fiom cases of general paralysis, as hrst shown by Mott and Halliburton, large amounts are present. Since Halliburton first demonstrated cholin in general paralysis, in his Croonian lectures on the chemical side of nervous activity, it has been found in a multi- ■ tude of other pathological conditions, but principally where recent and active nerve degeneration is taking place. The ilecithin of the myelin sheaths of nerves, is decomposed, probably through the agency of some enzyme, into glycerophosphoric acid, stearic acid, and cholin. This latter is found in the blood and cerebro-spinal fluid, while it is absent from the urine, except in experiments on animals, where the alkaloid is given subcutaneously or intracerebrally. When cholin appears in the fluid, there is also an increase of phosphoric acid therein (DoNath), this being another proof of its derivation from lecithin. In long continued activity of the central nervous system, as in convulsive seizures, lactic acid in large quantities appears in the cerebro-spinal fluid.

Cholin is detected by means of its platinum combination, with which it forms a double salt, and crystallizes in single large yellow octahedral and occasionally twin crystals, and easily differentiated from the platinum salts of neurin and potassium, by reason of its solubility, percentage of platinum and large size and grouping of the crystals. Since Halliburton first pointed out its occurrence in general paralysis, the work has been further elaborated by Mott, Wilson, Donath and CoRiAT. As a result it is found in many pathological states, where there is an axonal reaction with corresponding fiber degeneration (central neuritis), general paralysis, brain tumor, tabes, Huntington's chorea, delirium tremens, polyneuritis, epilepsy, senile dementia (especially where there is great wasting of the cortex), cerebral hemorrhage, beri-beri, Pott's disease, disseminated sclerosis, amyotrophic lateral sclerosis, focal myelitis, organic dementia, brain syphilis, chronic hydrocephalus (?) spina-bifida, neurasthenia ( ? ) and hystero-epilepsy. It has also been found in the blood of animals after experimental sec

CoRiAT, Chemistry of the Nervous System. 157

tion of the sciatic nerve, and in the peripheral nerves, brain and cord in polyneuritis and in all these latter, the amount present was parallel with the extent of the Marchi reaction. This is easily explained if we remember that the Marchi reaction depends on the preciptation of metallic osmium by stearic and oleic acids which are decomposition products of lecithin. The action of cholin on blood pressure has been carefully studied by Mott and Halliburton, while Donath has shown that the injection or intracerebral application of cholin produces in animals, strong convulsive phenomena, and he concludes from these experiments that the seizures in general paralysis and epilepsy are due to intoxication by this alkaloid.

In closing, we should like to say a few words on Donath's latest, and in many respects, his most pretentious paper, on the significance of cholin in epilepsy. He examined for cholin in sixty-four cases, detecting it by means of the platinum compound. In addition to the organic cases, where its presence would be expected, he finds it in those which we look upon as being functional in nature (neurasthenia). In addition, he finds the myelin forms of lecithin in two of his cases. This seems to us extremely doubtful and altogether inexplicable. Although using the platinum compound as a means of detecting cholin, yet according to the drawings and photomicrographs appended to his paper, he does not confine himself to octahedral or hexagonal crystals for a positive reaction, but labels anything and everything of crystalline nature as cholin," while we have always hesitated, and then put down as negative, what Donath in a large number of cases calls a positive cholin finding.

The bibliography appended makes no claim to completeness ; it merely gives those papers directly mentioned in this review, or those which have been consulted for the discussion of the material.

158 Journal of Comparative Neurology and Psychology.


Barrett, W.

On the Amounts of Water and Phosphorus contained in the Cerebral Hemispheres and Spinal Cords in General Paralysis of the Insane and in other Conditions. Archives of Neurology, 1899. Bethe, A.

Archiv fiir exp. Pathologie und Pharmakologie, 1902, 48, 73Coriat, I. H.

A Contribution to the Chemistry of Nerve Degeneration in General Paralysis and other Mental Disorders. Am. Jour, of Insanity,

1903, 59, No. 3.

The Chemical Findings in the Cerebro-Spinal Fluid and Central Nervous System in various Mental Diseases. Adi. Jour, of Insanity,

1904, 59, No. 4.

The Cerebro-Spinal Fluid in Hydrocephalus. Adi. Jour, of Physiology,

1903, 10, No. 3. The Production of Cholin from Lecithin and Brain Tissue. A>n. Jour. of Physiology, 1904, 12, No. 4. Donath, J.

Die Behandlung der progressiven Paralyse. Allg. Zeitschrift fi'ir Psy chiatrie, 1903, 60, H. 4. Zeitschrift fi'ir physiol. Chemic, Oct., 1903.

Der Phosphorsauregehalt der Cerebro-spinalfliissigkeit bei verschiedenen insbesondere Nervenkrankheiten. Zcit. fiir physiol. Chemie, 1904, 42, H. I and 2. Die Bedeutung des Cholins in Epilepsie. Dcut. Zcit. fiir N erz'enheilk uncle, 1904, 27, H. I and 2. Gutinkov, S.

Zu Lehre von der chemischen Zusammensetzung des menschlichen Gehirns. Zcit.f. Psychiatric, 1897, 53, 270-329. Halliburton, W. D.

The Croonian Lectures on the Chemical Side of Nervous Activity.

1 90 1 . The Coagulation Temperature of Cell-Globulin and its bearing on Hyperpyexia. Archives of Neurology, 1903, 2. Halliburton, W. D. and Mott, F. W.

The Physiological Action of Cholin and Neurin. Philosophical Transactions, No. 7, 1899. Hatai, S.

The Effect of Lecithin on the Growth of the White Rat. American

Journal of Physiology, 1903, 10, No. i. The Effect of Partial Starvation on the Brain of the White Rat, Jour, of Physiology, 1904, 12, No. I.

Cor I AT, Chemistry of the A'eii>otis System. 159

Koch, W.

Zur Kenntniss des Lecithins, Kephalins und Cerebrins aus Nerven substanz. Zeit. fi'ir physiol. Cheinie, 36, H. 2 and 3. The Lecithins. Decennial Publications, University of Chicago, 1902. Die Lecithine und ihre Bedeutung fiir die lebende Zelle. Zeit. fir

physiol. Cheinie, 1903, 37, H. 3. Methods for the Quantitative Cliemical Analysis of the Brain and Cord. Am, Jour, of Physiology, 1904, 11. Xo. 3. Levene, P. A.

On the nucleo-proteid of the Brain. Archives of N'em-ology and Psychopathology, 1899, 2, Nos. 1-2. On the Chemistry of the Chromatin Substance of the Nerve Cell. /our.

of Medical Research, 1903, 10, No. 2. On the Autolysis of Brain Tissue. Journal of Medical Research, 1903, 10, No. 2. Mainzer.

Von der Beziehungen des Tigroids zu Kern und Plasma. Monat. fiir Psvchiatrie u. A^eurologie, 1903, 12, H. 3. Mott, F. W.

Note upon the Cholin Test for Active Degeneration of the Nervous System. .Archives of N'eiirology, 1903, 2. Mott, F. W. and Barratt, W.

Observations on the Chemistry of Nerve Degeneration, .-irchives of Neurology, 1899, 1. Noil, A.

Ueber die quantitativen Beziehungen des Protagon Zum Nervenmark. Zeit. fir Physiol. Chemie, 1899, 27, H, 4 and 5. Scott, F. H.

On the Structure, Microchemistry and Development of Nerve cells; with special reference to their Nuclein compounds. Trans. Canadian Inst. 1898 and 1899, 6. Sicard, J. A.

Le Liquide Cephalo-Rachidien. Paris, 1902. Thudichum, J. L. W.

Die chemische Konstitution des Gehirns des Menschen und der Tiere. 1 90 1. yA^ilson, M. S.

La Choline dans le Liquide Cephalo-Rachidien comme une signe de Degeneration Nerveuse. Rei'. Neurologique, April 30, 1904. Wood, H. C. Jr.,

On the Toxic Action of the , Decomposition Products of Lecithin, Univ. of Penn. Med. Bull., May, 1902.


Burkholder, J. F. The Anatomy of the Brain. A Study of the Human Brain from the Brain of the Sheep. A Manual for Students in Medicine, Biology and Psychology. With an introduction by Professor Henry H. Donaldson. 175 Pages, 32 Full Page Plates. Chicago : G. P. Engelhard & Co., 1904. This is a laboratory manual designed to meet the needs of those for whom human material is not accessible in sufficient amount for thorough demonstration purposes. It is well adapted to serve this 'purpose in medical schools and colleges, the text being clear, the figures good and the arrangement logical. There are unfortunately several misprints, some of serious nature, such as the confusion in the designations of Plates VIII and IX. Plate IX is defective also in that it does not show as clearly as such a dissection should show the ventricular boundaries. On the whole, however, the work is very well done and the book deserves a wide circulation. It should do much toward moderizing the instruction on the nervous system, both in medical schools and colleges. c. j. h.

Smith, G. Elliot. Studies in the Morphology of the Human Brain with special Reference to that of the Egyptians. No. i. The Occipital Region. Records of the Egyptian Gov't. School of Medicine, Vol. 2, pp. 125-170, 2 plates and 47 text-figures. Cairo, 1904. The ultimate purpose of this series of studies the author states to be anthropological. And he adds, "The aim of these preliminary morphological studies is to enable us to discriminate between important and valueless features, between the kiud of information that is worth collecting and that which it would be a mere waste of time to seek." This, then, is the motive underlying the present elaborate comparative study of the occipital region: "to learn the relative value of the •data upon which our conclusions are to be based." In reaching his conclusions regarding the homologies in this region (for which the original memoir must be consulted) the author studied more than 600 hemispheres of various Primates of every genus and about 2000 human hemispheres. The importance of the paper, however, lies not so much in the fixing of particular homologies as in the sifting out of the essential landmarks from the unimportant variations. c. j. h.

Literary Notices. i6i

Sterzi, G. Die Blutgefasse des Riickenraarks. Untersuchungen iiber ihre

vergleichende Anatomic und Entwickelungsgeschichte. Anaton. Hefte

{Merkel u. Bonnet), I Abt., 24, 1904, pp. 364, 4 plates and 37 text-figures.

This is an important contribution to a field which has been but

little cultivated — the comparative anatomy of the blood supply of the

spinal cord. The work is divided into two parts, anatomical and

embryological. In the first, representative types of the several animal

classes are successively described and in the second the development

of typical species is followed. Each section is followed bj a summary

and a brief general review closes each part. The general conclusions

are that the blood vessels of the spinal cord divide in such a manner

as to bring about more effective nutrition of the nervous matter the

higher we ascend in the animal series, and that the ontogenetic stages

traced out represent stages in the phylogenies of their species.

c. J. H.

Prince, Morton. The Course of the Sensory Fibers in the Spinal Cord and Some Points in Spinal Localization Based on a Case of Section of the Cord. Journ. Nerv. Ment. Dis., 32, 81-100.

A stab wound in the neck severed both dorsal funicles completely and the lateral funicles on one side more deeply than on the other. Tactile sensation was preserved on one side, showing clearly that such sensation is not transmitted exclusively by the dorsal funicles. A study of the hemianaesthesia shows that tactile sensation is transmitted by the lateral funicles of the opposite side. The case is important as presenting an unusually clearly defined lesion with excellent opportunities for careful study. The paper is accompanied by a digest of the literature. c. j h.

Gushing, Harvey. The Sensory Distribution of the Fifth Cranial Nerve. Bull. Johns Hopkins Hospital, 15, 213-232, 1904.

An exhaustive and fully illustrated study based chiefly on clinical examination of patients previously operated upon for complete extirpation of the Gasserian ganglion and of the cervical nerves. The cutaneous distribution is mapped more accurately than has hitherto been done and also the mucous membrane field within the mouth. In the latter field the observations show complete anaesthesia to all forms of sensation, including such irritants as ammonia fumes, on the whole of the operated side, with two exceptions, (i) The area at the base of the tongue supplied by the glossopharyngeus retains not only its sense of taste, but also general sensation unimpaired. (2) The second exception is the preservation of sensation in the field supplied by the chorda tympani of the facialis. The sensation of taste is not affected

1 62 Journal of Comparative Neurology and Psychology.

in the slightest degree in any of these cases, and, \>hile ordinary tactile thermal and pain sensations are abohshed, there remains in this second field a sensibility to a moving contact which leads the author to suspect the existence of fibers of general sensation in the chorda tympani. This peculiar sensation was found to be totally absent, as well as gustatory sensation, in a case in which the operation had involved the facial as well as the trigeminus nerve.

The clinical data are conti-oUed by a discussion of the adult anatomy, the comparative anatomy and the embryological relations, the results of which clear up several vexed questions of peripheral nerve distribution, as well as strengthen to the point of demonstration some of the homologies proposed by recent students of nerve components in the lowliest vertebrates. c. j. h.

Pinkus, Felix. Ueber ein dera mensclllichen Haar benachbartes Sinnesorgan. Vcr/il. Ges. Deiitscher Natiirf. u. Aci-ztc. 75 Versam., Cassel. Zweiter Teil II, Hefte, Med. Abt., pp. 344-346, 1904.

Dr. Pinkus describes the hair and its accessory structures and considers that in position and structure these parts are analogous to the scales of reptiles. The human hair preserves the triple-hair structure characteristic of all mammals though the schema is often reduced to a single hair. A rudimentary scale lies immediately in front of the hair, while the sensory organ under consideration occupies a position behind (i. e. in the acute angle of) the hair. Histologically this knob-like body consists of a broad cutico-papilla with thickened epidermis. This structure, the under surface of whose epithelium is composed of pallisade-like cells, is highly innervated. The theoretical conclusions, which seem perhaps considerably to overbalance the meagre descriptive portions, are as follows : "We have to do, therefore, with an organ of great antiquity the fundamental form of which is perhaps preserved to us in the form of a genuine epithelial sensory ridge, in Hattcria punctata presumably the oldest reptile.' Our author agrees with Keibel that the hair is "homologous with a specially modified part of a scale." c. l. h.

Goldstein, Kurt. Zur Frage der Existenzberechtigung der sogenannten Bo genfurchen des embryonalen mensclllichen Gehirnes, nebst einigen weiter.

en Bemerkungen zur Entwickelung des Balkens und der Cap.sula interna.

Anat. Auz., 24, 579-595' I904 The author finds no cerebral fissures in the human embryo of 3^

to 4 months. The so-called Fissura prima (vordere Bogenfurche of

His) is not a true fissure, but it might properly be called a "sulcus

olfactorius" or "fovea olfactoria." Contrary to the idea of His, also.

Literary Notices. 1 63

the corpus callosum, beginning in the lamina terminalis, expands into its adult relations by intussusception of new fibers, and not by actual coalescence of the hemisphere walls. o. e. c.

Lewis, Warren Hermon, Experimental Studies on the Development of the Eye in Amphibia. I. On the Origin of the Lens. T/ie American Journal of Anatomy, 3, 505-536, 1904.

In this communication Lewis reports a series of experiments upon the embryos of Rami palustris and R. syivatica, in which the optic cup was severed from the brain and removed from its normal position before any fundament of the lens had appeared. The author interprets his experiments as proving that contact between the optic cup and the skin is necessary for formation of the lens, and that the power to take part in lens-formation is not restricted to any particular region of the skin or optic cup. These facts, he believes, are "entirely in opposition to Weismann's theory ot determinants." g. e. c.

Levi, Guiseppe. Ueber die Entwickelung und Histogenese der Ammonshorn. formation. Arch. f. mik. Anat., 64, Heft 3, 389-404, I PL, 1904.

The chief cause of the rolling in of the cortex in the Ammonshorn is the atrophy of the medial wall of the hemisphere in the region of the plexus chorideus. j. b. j.

Scaffldi, Vittorio. Ueber den feineren Ban und die Funktion der Hypophysis des menschen. Arch. f. mik. Anat., 64, Heft 2, 235-255, i PI., 1904.

Two kinds of cells are present, distinguished by their staining properties, each of which probably produces a definite substance which helps to form the secretion of the gland. j. b. j.

Kamon, K. Ueber die "Geruchsknospen." Arch. f. mik. Anat., 64, Heft 4, 653-664, I pl-, 1904.

Esox, Trigla and calf. Sublimate ci aJ.\ iron haematoxyliu; Golgi and methylene blue; maceration with 0.05% chromic acid. Figures from haematoxyliu sections support the conclusion, "Es giebt weder in der Ceruchschleimhaut der Fische noch der Sanger Bildungen. mit den Geschmacksknospen verglichen werden konnen."

J. B. J.

Heath, Harold. The Nervous System and Subradular Organ in two Genera of Solenogastres. Zool. Jahrb., Ahth. f. Anat. u. Ontog., 10, Heft 3, 399408, I pl., 1904.

General description nf the nervous system. j. b. j.

Rubaschl<in, W. Studien iiber Neuroglia. Arch. f. mik. Anat., 64, Heft 4, 575-626, 4 pis., 1904.

Cat. Staining with methyl violet after special method of fixation. Description of types of glia elements with classification accord

164 Journal of Comparative Neurology and Psychology.

ing to their genetic relations: glia-genetic cells, astrocytes of young form, astrocytes of definitive form, cells without processes. Relation of neuroglia to the vessels, to the nerve cells and fibers. Structure and relations of the ependyma. Many of the figures show lack of care and exactness, and some are inadequate to their purpose.

J. B. J.

Pinkus, F. Ueber Hautsinnesorgane neben den menschlichen Haar (Haar scheiben) und ihre vergleichend-anatomische Bedeutung. A>xh. f. mik. Anat., 65, Heft i, 121-179, 4 pis., 1904. Man, ape, monotremes, reptiles. Histological methods and intravitam methylene blue. Deals with the homology of the organ in man with the Tastfleck in reptiles and of the hair with an undifferentiated part of the reptilian scale. j. b. j.

Ballowitz, E. Die Riechzellen des Flussneunauges (Petromyzon fluviatilis L.)

Arch.f. mik. Aiiat., 65, Heft i, 68-85, ' ?'■> I904Ewing, Henry Z. The Functions of the Nervous System, with Special Regard

to Respiration, in Acrididae. A'aiisns Unh'. Sci. Bull., 2, 305-319, 1904.

The main object of Mr. Ewing's investigations on several species of grasshoppers has been the study of the nervous control of respiration. In addition he has made a considerable number of observations on other functions of the nervous system. His results in the main corroborate those of Bethe, and point to the segmental character of many of the reactions of the insect.

The conclusions reached are based on effects observed after the removal of the various ganglia. While the supraoesophageal and suboesophageal ganglia are found to exercise no control over respiratory movements, each ganglion of the thoracic and abdominal ventral nerve cord appears to act as the center for the respiratory and general reflex movements of the segment in which it lies. Neither the supraoesophageal nor suboesophageal ganglion is the center for coordinated movements. These can be carried out by the insect after the entire head has been cut off. Removal of the supraoesophageal ganglion destroys the animal's power to inhibit reflexes, and it no longer executes swallowing movements. There is also a loss of tonus of the muscles and general weakness. The suboesophageal ganglion appears to have some control over equilibration.

F. W. C Lukas, Franz. Psychologic der niedersten Tiere. Eine Untersuchung ilber die

ersten Spuren psychischen Lebens im Tierreiche. Branmiiller, Wein und

Leipzig, 1905, Vni + 276.

This book, although complete in a certain sense in itself, is issued

.as the first part of a general aniiual psychology which the author tells

Literary Notices. 165

us in his preface he may sometime finish. The present volume is concerned with the Protozoa, the Coeleuterata, the Echinodermata and the Vermes.

Alter an excellent discussion of criteria of the psychic and an examination of various forms of activity in an introductory chapter on the problems and methods of animal psychology, the author proceeds to consider the materials for a psychology of the above named phyla in accordance with the following plan :

A. Description of anatomy, especially of the nervous system.

B. Description of activities.

I. Changes of substance: I. Taking food, 2. Assimilation and dissimilation, 3. Excretion. II. Changes of form: i. Adaptation and transmission, 2. Growth and reproduction. III. Changes of energy: i. Apparent spontaneous manifestations of energy-change, movement, light, electricity, 2. Action of mechanical, photic, chemical, thermic, electric and acoustic stimuli.

C. Consideration of the question of the existence of consciousness.

I. Bearing of anatomical conditions, especially of the nervous system. II. Bearing of activities, especially of movements. III. Bearing of the question of the value of consciousness for the organism.

As this outline indicates, the author follows in general the scheme of classification of organic processes offered by Verworn. He gives brief, clear, structural descriptions, and satisfactory accounts of well chosen typical modes of behavior. The work is not thorough and exhaustive in its reviews of the literature of its topics, since it is intended for the general biologist rather than for the specialist in animal behavior or comparative psychology.

So far as materials are concerned it presents nothing new, for the author has taken all his structural facts directly from the works of other authors, and the only original work he claims within the field of activity is the verification of the reports of other investigators. Consequently the only portion of the book of much value to the specialist in animal psychology is that which is concerned with the author's conclusions, on the basis of the facts collected, concerning the existence of consciousness in the animals of the phyla mentioned, and its values.

In brief, the conclusions stated are that the Protozoa give no evidence of consciousness, that certain of the Coelenterata, the Hyroids, exhibit the first signs of consciousness in the animal kingdom, while the Echinodermata and Vermes give evidences of somewhat more complex forms of consciousness. The author is guarded in his

i66 Journal of Comparative Neurology and Psychology.

statements and places his conclusions upon clearly defined bases of inference, which he discusses at length in his introductory and concludino; chapters.

This book is a valuable contribution to the text-book and student literature of animal psychology. R. M. y.

Parker, G. H. Hearing and Allied .Senses in Fishes. Bull. U. S. Fish Commission, 45-62, for 1902 (published in 1904). Professor Parker after an admirably clear and concise historical sketch of his subject describes a series of experiments in which by nerve cutting he succeeded in demonstrating the specific functions of the ear and lateral line organs of Fundulus, mackerel and menhaden .

Detailed work was done with Fundulus hctovcUtus. According to the author this animal when normal responds to vibrations transmitted to the water of the aquarium by a tuning fork of 128 V. by movements of the pectoral fins and increased respiratory rate. After the eighth nerve has been cut these responses no longer appear, hence the author concludes that the fish possesses a sense of hearing.

Similarly it is shown that the reactions to slight mass movements of the water, which are characteristic of the normal Fundulus, are not given by individuals in which the nerves to the lateral line organs have been cut.

This work is of special interest and importance, first, because the experiments were simple in plan, carefully executed, and definite in their results;' second, because the results themselves are in large part contradictory of much that has heretofore been published on jthe subject.

It is Parker's belief that the cutaneous organs, the lateral line organs and the ear form a series of sensory organs which respond to mechanical stimuli of various degrees of delicacy. Gross stimuli perhaps effect the cutaneous organs chiefly, less violent stimuli act primarily upon the lateral line organs, and the most delicate mechanical stimuli affect the ear. r. m. \ .

Jennings, Herbert S. Contributions to the Study of the Behavior of Lower Organisms. Carnegie Institution of Washington, 1904, Publication No. 16, pp. 256.

First Paper. Reactions to Heat and Cold in the Ciliate Infusoria. As Paramecium swims it draws towards its anterior end a vortex of water which passes in a slender stream along its oral groove to the mouth. Thus any stimulus that can be transmitted by such a stream, as for instance a chemical or heat and cold, comes into contact with the

Literary Notices. 167

anterior end and oral groove earlier than with any other part of the body. These are, moreover, the most sensitive parts of the body. Now when Paramecium thus encounters a stimulating agent, it appears that these are the parts of the body that react. The animal at first swims backward, circling at the same time toward the aboral side. After a moment the animal swims forward again, still turning toward the aboral side. The direction of its course is thus changed, and the animal continues in this new direction if it does not again encounter the stimulating agent ; but if it does, it goes through the same motor reaction again, by swimming backward, turning toward the aboral side and then swimming forward once more. This process is repeated until the animal is finally brought beyond the range of the stimulus. The animal invariably turns toward the aboral side, even when this at first brings it more immediately into the region of the stimulus than it had been before. The author aptly names this, orientation by trial and error. And he further finds that this is the only manner of orientation observable in Paramecia, and that this fact is utterly at variance with the current theories of thermotaxis.

The conclusions are based on extended observations of Paramecium, and they are confirmed by entirely concordant observations of the other ciliate infusoria, Oxytricha fallax, O. aeruginosa, Stylonychia myiilus, Stentor caendcus, Spuvstomum ambigiium and Bursaria triincatella. Both heat and cold were tried as stimuli in nearly all the cases.

Second Paper. Reactions to Light in Ciliate Flagellates.

The reaction of the ciliate, Stentor, to light is the same as its retion to heat and cold. It stops or swims backward a short distance, then turns toward the aboral side, and resumes its forward motion. "This is the reaction which is produced by strong mechanical stimuli, by heat, and by chemical stimuli, acting upon the anterior end or upon the body as a whole." And here too, the reaction is the same, although at first the animal may be brought directly into the stimulus which it is seeking to avoid. "-The direction of turning is thus determined by aji interfial factor — the structure of the body." The orientation is not governed by the direction of the light-rays (contra Loeb), nor yet by the part of the body that receives the stimulation (contra Verworn and others). Indeed the response is not altered if the stimulus acts on all parts of the body at once.

The flagellates normally swim in a spiral path, continually swerving toward that side which bears the larger "lip" and the eye, the socalled dorsal side. While swimming toward a source of light Euglena viridis will, if suddenly the illumination is decreased or greatly in

1 68 Journal of Comparative Neurology and Psychology.

creased, swerve more rapidly toward the dorsal side and swim less rapidly forward, and then proceed as before but in a new direction. In course of this the anterior end of the body describes a circle while the posterior end moves but little and is virtually a pivot for the freer end. "The reaction which occurs when the illumination is changed is really an accentuation of a certain feature of the usual movements." Now if the change in illumination has been slight, the animal will slowly resume its forward movement swerving less noticeably toward the dorsal side, so that presently it will be moving along its former spiral course. But if the modified illumination is now too great, the animal while describing the pivotal movement will find its anterior (and more sensitive) end turned toward the light at some points of its circular course and away from it at others ; since now the light is supraoptimal, the animal will commence the forward movement while its anterior end is turned from the light, so that its new spiral course as it is gradually resumed will be directed at least somewhat away from the source of light. If this change in course is not sufficient to make the animal recede from the source of light, the animal will either tack again in the same way as before, or else change its orientation gradually but in a way very comparable with the previous. That is, the organism will swerve farther than normally toward the dorsal side and this will cause its anterior end to point successively in many different directions. In some of these positions the anterior end is directed more nearly toward the source of light, in other positions farther away from it. In the latter case the swinging toward the dorsal side becomes less marked ; hence the succeeding phase of the swing, which carries the anterior end toward the light, is less pronounced; The anterior end therefore does not swing so far in the direction toward the light as in the preceding phase it swung away from the light. But as a result of such swerving as does occur the anterior end is now directed more away from the source of light than before. Thus this gradual change of orientation is effected by means of the same principles as the former, more sudden change. In either case the direction of the rays of light is not a determining feature, and the orientation is effected by a definite sort of "motor response."

This description applies to the movements of orientation of ciliates and flagellates in general. The ciliate studied was Sfentor oreruleus: the flagellates were Eugkna viridis, Cryptomonas ovata (sometimes called Chilomonas ovatd), and a species of Chlamydomonas .

Third Paper. Reactions to Stimuli in Certain Rotifera. The author studies the Rotatoria, a group of metazoa much re

Literary Notices. 169

sembling the infusoria in their mode of Ufe, in order to find whether the reaction method of the metazoa differs radically in character from that of infusoria. The following free-swimming Rotifera, which progress through the water in the same manner as the ciliate infusoria, were studied; several species of the Rattulidae and of the Cathypnadae, two or three species of Euchlanis, Ploesoma lenticulare, Amiroea cochkaris, and Brachionus pala. The stimuli used were chemicals^ heat, light, and the constant electric current. The reactions of all these Rotifera are essentially the same as those of the ciliate infusoria.

Fourth Paper. The Theory of Tropisms.

This paper forms an admirable summary and theoretical discussion of the undoubtedly careful and exact observations described in the foregoing three papers. And it is to be said, further, that in each paper the author evinces the rare combination of qualities — a lively theoretical interest combined with the power of impartial observation.

The paper aims utterly to destroy the theory of the several "-taxes." Firstly, the theory of Loeb and others that the direction of ray (in light and electricity) is the determining factor in orienentation is untenable, because all the organisms studied react only by swerving toward a structurally defined side of their own bodies, even although this movement causes them to turn directly toward the rays that they seek to avoid. The theory that the intensity of stimulation effects orientation by a local effect on such parts of the body as it reaches is untenable, because in many, if indeed not most, organisms the anterior end is the only part sensitive enough to be effected by the stimuli (that is, the threshold of orientation is the threshold of stimulation of the anterior end); and because, as in the previous case, the response is invariably a perfectly definite "motor reaction" toward an anatomically defined side of the body of the organism. The defenders of the theory of tropisms will hardly call in question Professor Jennings' patently exact and unbiased observations, nor his methods, which are of the most careful, nor yet the range of organisms which he has chosen to study, which is far greater than that of any other individual experimenter. But they may possibly ask Professor Jennings what it is if not intensity of stimulation that causes the organisms, while circling on their anterior ends as pivots, to dart forward more or less suddenly at that point of the circle where the anterior end receives least of the undesirable stimulus ; and if this cause is the intensity, whether his observations have not demonstrated a phototaxis of the anterior end. But even if this were to be the case, Professor Jennings would have effected a most important refinement on the old

I/O Journal of Comparative Neurology and Psychology.

theory, and contributed vastly to our knowledge of the behavior of lower organisms.

Fifth Paper. Physiological States as Determining Factors in the Behavior of Lower Organisms.

In unicellular organisms as in multicellular the forms of reaction to stimuli exhibited clearly indicate the importance of internal factors (physiological states) as determining conditions. "To the same stimuli, under the same external conditions, the same individuals react at different times in radically different ways, (thus) showing the existence of different physiological states of the organism, which determine the nature of the reactions," p. 126. Jennings adduces numerous observations of the behavior of various unicellular organisms and flatworms in proof of the falsity of the prevalent notion that the reactions of these animals are determined almost wholly by external conditions. As a matter of fact all organisms, so far as reactions are concerned, are influenced by both external and internal conditions, and it is only because of the apparent invariableness of the reactions of the lower animals that they are so commonly spoken of as automata which are at the mercy of their surroundings in a quite different way than are the higher animals.

The paper serves the good purpose of calling attention to the fact that reactions are more often determined by the total state of the organism than by the local action of a stimulus.

Sixth Paper. The Movements and Reactions of Amceba.

This paper, like the earlier ones of the volume, rudely shakes the foundations of certain current theories of reaction, and it totally destroys certain explanations of the reactions of simple organisms which have been pretty generally accepted by physiologists. Unfortunately for the peace of mind of certain of the earlier investigators of the movements and reactions of Amoeba, the author was not satisfied to accept what had already been done in this field of research as the basis of his work. He chose instead to begin at the beginning and to attempt to verify the descriptions already given.

Rhumbler and Butschli state that in the normal locomotion of Amoeba there is a forward moving current of endosarc in the middle axis which flows outward at the anterior end of the organism, then backward along the surface. According to these physiologists, the protoplasmic currents within a moving Amoeba are essentially like the movements produced in a drop of water by lowering the surface tension at some point. The axial current moves toward the point of decreased tension. Consequently Rhumbler and Butschli conclude

Literary Notices. 171

that the movements of Amceba are due to changes in surface tension. Jennings, on the basis of careful observation of the movements of several species of iVmoeba, states that there are no side or backward currents, but that the process occurs as follows : "In an advancing Amceba substance flows forward on the upper surface, rolls over at the anterior edge, coming in contact with the substratum, then remains quiet until the body of the Amoeba has passed over it. It then moves upward at the posterior end, and forward again on the upper surface, continuing in rotation as long as the Amoeba continues to progress. The motion of the upper surface is congruent with that of the endosarc, the two forming a single stream " p. 148.

Accurate and detailed descriptions are given of the processes of formation and withdrawal of pseudopodia, and it is shown that contractility is a capacity of the ectosarc. "Altogether, then, our results lead us to look upon Amceba as an elastic and contractile sac, containing fluid. In locomotion one side actively stretches out, becomes attached to the substratum, and draws the remainder of the sac after it in a rolling movement. The primary phenomena are the stretching out of one side, the elasticity, and the contractility of the outer layer" p. 172.

Under the subject of reactions, the author presents the results of extensive studies of the influence of different stimuli upon Amoeba. He distinguishes three classes of reaction : the positive, the negative, and a complex feeding reaction. Several interesting instances of the pursuit of food are cited in support of the author's opinion that the behavior is by no means so simple and uniform as is commonly supposed.

After redescribing the protoplasmic movements of Amoeba in the light of his own observations, the author discusses the physics of locomotion with results which may most fitly be expressed in his own words. "Putting all our results together we must conclude that the movements and reactions of Amoeba have as yet by no means been resolved into their physical components. Amoeba is a drop of fluid which moves in its usual locomotion in much the same way as inorganic drops move under the influence of similarly directed forces. But what these forces are is by no means clear. When we take into consideration the currents as they actually exist, local decrease in surface tension breaks down completely as an explanation for the locomotion and other movements. The locomotion taken by itself might be explained as due to the adhesion of the fluid protoplasm to solids.

172 Journal of Comparative Neurology and Psychology.

taken in connection with the surface tension of the fluids, but this explanation fails when we consider the formation of free pseudopodia, and (lit^cover that all the processes concerned in locomotion can take place without adhesion to the substratum." 225.

Seventh Paper. The Method op Trial and Error in the Behavior OF Lower Organisms.

In higher animals we speak of karm'ng hy the method of trial and error, and in unicellular organisms we may speak of the trial and error method of reaction. In the former instance the animal when first confronted by a certain situation exhibits a number of reactions ; one of these reactions brings satisfaction and therewith the trial reactions cease. The next time the same situation is presented the animal, if it has learned, omits the useless or harmful trials and performs only the appropriate reaction. In the latter instance, namely, in case of unicellular organisms, each re-presentation of the situation brings about, so far as yet observed, the same sort of trial and error series of reactions. Whereas the mammals after a number of experiences omit the majority of the trials, the protozoa apparently have to go through the whole gamut of tests each time.

Trial may then be defined as the process of testing different portions of the environment, and error is descriptive of those trials which result in negative reaction. Now, the problem proposed by the author is, How is error distinguished by the organism? "In ourselves the stimuli which induce the negative reaction bring about the subjective state known as pain, and popularly we consider that the drawing back is due to pain. Is there ground for this view?" p. 248. This problem, Jennings insists, is one throughout the animal series.

Finally, the method of trial and error which the author believes to be the prevalent mode of reaction among lower organisms is not in agreement with the principles of reaction which are at the bottom of the tropism schema. For the tropism is a fixed way of reacting in response to a certain stimulus, and it has no place for the trials which have been demonstrated to occur.

Whether this is fair to the advocates of the tropism formulae for reactions would be hard to say. Certainly some of them will at once contend that Jennings by merely breaking up the stereotyped reaction into a number of parts has not altered the nature of the reaction.

The author's insistance throughout this volume upon the importance of internal conditions in lower organisms is certainly justified by the facts he presents.

Literary Notices. i 73

That this series of papers is fundamentally important for the science of animal behavior is obvious.

E. B. H. AND R. M. V.

Hornaday, Wm. T. The American Natural History. Xe7i< York, Scribner's Sons, 1904, XXV -\- 449.

A popular natural history which C(intains much interesting and valuable information concerning American animals. The work is splendidly illustrated and pleasingly written. It should be of considerable value as a means of arousing interest in animals and stimulating to a study of their characteristics. r. m. v.

'Cohn, Paul. Gemiitserregungen und Krankheiten. Berlin, I'ogel und Kreienbrink, 1903, pp. 148.

The book contains a study of the nature and localization of emotions. The first part deals with the physiological accompaniments of feelings and emotions, the second with their pathological effects, and the third gives hints for a prophylaxis. The author adopts the theory that all feelings can be reduced to organic sensations of the body and of the brain. He professes a general disdain for all previous work, which he claims to be too speculative, and he acknowledges Nietzsche as his only predecessor. None of the modern inquiries into the nature of feelings is mentioned, and the experimental studies of the physiological influence of feelings are utterly ignored. f. m. urban.

Putnam, J. J. The Value of the Physiological Principle in the Study of Neurology. Am. Medicine, 1904, 8, No. 25, 10^1-1056.

Dr. Putnam's address before the section of Neurology of the International Congress of Arts and Sciences at St. Louis is published in full in American Medicine, as cited above, and simultaneously in the Boston Medical and Surgical Journal. The contrast is sharply drawn between Virchow's "Anatomical Principle in the Study of Disease" and the attempt to localize morbid processes on the one hand, and on the other hand the morbid process as maladjustment. "No anatomical research can pierce to the secret of broken coordinations, and yet it is in these that a great part of disease begins, or comes eventally to consist." "Every organism, whether we call it diseased or well, presents itself to our view as a web of interwoven 'energies,' which in order to study them by anatomic means, we must break artificially into fragments that have, in reality, no correspondingly separate existence." This is a point of view which has its application in normal function, as well as in pathology. c. j. h.

1/4 Journal of Comparative Neurology and Psychology.

Halben, R. Theoretisches iiber die Bedeutung des Pigmentes fiir den Sehakt

der Wirbellosen, speciell der Protozoen. Biol. Centralbl., 1904, 24, No.

8, 283-288.

An argument, against Hesse, for the importance of pigment in

light perception. A foot-note contains a discussion of the nature of

light, which should be read by all students of phototaxis. G. w.

Weysse, Arthur W- Notes on Animal Behavior. Science, N. S., 1904, 19, No. 495, 955-957Discusses certain actions of a black-and-tan terrier, which by many would be taken to exhibit a high degree of intelligence. Weysse shows how, in these particular cases, much simpler explanationg can be given, based on the ordinary habits of the dog.


Scott, William E. D. The Inheritance of Song in Passerine Birds. Remarks on the Development of Song in the Rose-breasted Grosbeak, Zamelodia ludoviciana (Linnseus), and the Meadowlark, Sturnella magna (Linnreus). Science, N. S., 1904, 19, No. 95, 957-959 ; Further Observations on the Development of Song and Nest-building in Hand-reared Rose-breasted Grosbeaks, Zamelodia ludoviciana (Linneeus). Science, N. S., 1904, 20, No. 504, 282-283.

Rose-breasted grosbeaks were taken from the nest when five days old and after being kept in confinement until about eight months old they sang a song that "could not be referred to the rose-breasted grosbeak." Meadowlarks taken from the nest and kept in confinement, besides having a peculiar song of their own, accompanied by a "parade or dance," also imitated the songs of some German "blackbirds" {Merula ftierula) with which they were associated. The conclusion is "that birds are influenced in their early lives very strongly by any noise that arrests their attention, even in a wild state, and that their propensity to imitate and differentiate their normal methods of song [% greatly exaggerated under the artificial state wherein they live when in confinement."

In the second paper the observations of the first are confirmed. The songs of the grosbeaks came to resemble that of a green bulbul [Chloropsis hardtvickii) which was kept in an adjacent cage. In May the grosbeaks attempted to build a nest in the cage, but although supplied with materials, they were unable to construct a nest for themselves, though they added a lining of feathers and cowhair to artificial nests that were supplied. leon j. cole.

The Journal of Comparative Neurology and Psychology

Volume XV 1905 Number 3


By J. B. Johnston.

With Plates I to IV.


/. Introduction. 176

a. Nature of the unsettled problems ..... 176

b. Nerve components . . . . . . . -177

c. Functional divisions of the nervous system . . . 180

d. Bearing upon the subject of head morphology . . .182 2. Number and relations of mesodermic somites . . . . .184 J. Branchial apparatus and lateral musculature . . . .188

4 . Segmentation of the central nen'ous system . . . . .191

5. Typical sense organs of vertebrates . . . ■ ■ .194

6. The anterior end of the head . . . . . . .197

7. The segment containing N. III. Segment^ ..... 204

8. Segment 6 2o7

9. The absence of nerve roots from one or more hind brain segments and

the determination of the neuromeres concerned . . . 211

10. Segment 7. The acustico-lateral system . . . . .215

11. Segments 8 and follozving . . , . . . . .2^3

12. The occipital region and the posterior limit of the head. The constitu tion of the vagus. Shifting of organs . . . . .224

/J. The morphology of the visual organs ...... 233

14. The nerves of Piatt, Locy and Pinkus ..... 244

/J. Morphology of the olfactory organ. Functional and morphological

relations to the gustatory system ...... 245

16. The first four head segments ....... 249

a. Segment 4 ......... • 249

b. Segment 3 ......... • 250

c. Segment 2 ......... • 250

d. Segment i Prostomium ....... 252

ly. The dorsal commisstires of the brain ...... 253

1 Studies from the Zoological Laboratory of West Virginia University, No. 9.

176 Journal of Comparative Neurology and Psychology.

/S. Tfu sympathetic system . . . . . . . .257

ig. Relation of dorsal ami ventral nene roots to the myotomes . . 2-8 20. Comparison of head and trunk ....... 260

Summary ............ 261

List of papers cited .......... 265

Description of figures ......... 273

/. Introduction. a. Nature of the unsettled problems.

Shall we consider vertebrates as animals possessing a high degree of cephalization from their first appearace ?. The structure of their near relatives, Amphioxus and Ascidians, is against this view. The structural relations of vertebrates and invertebrates indicate that the ancestors of the vertebrates were segmented invertebrates in which the process of cephalization had not gone very far. Even within typical vertebrates evidence is not lacking that the special sense organs of the head were late to appear; that the branchial apparatus was at one time iTiore extensive, reaching into what is now the trunk ; that the nerves of the branchial region once had a more simple segmental arrangement ; and that in the brain itself the several regions were once less highly specialized than at present. If Amphioxus be considered, the presence of true nephridia (41) in the head and the slight specialization in the head region seem to relate this "lowest vertebrate" with invertebrates rather far down the scale.

If, then, the ancestral vertebrate had only a slight head development, it is evident that the interpretation of the special organs of the head of typical vertebrates is to be reached by a study of their structure, function, and phylogenetic history, with a view to tracing them back to their unspecialized beginnings. When each organ has thus been followed back to its ancestral condition we sh ill have reduced the vertebrate head to terms — not of the trunk, but of a more simple condition which underlies both head and trunk. Such is the real problem of head morphology as the writer understands it.

The central difficulty in framing such a conception of the head is the matter of segmentation. Head specialization has

Johnston, Morphology of the Head. 177

gone so far that in known vertebrates few or no segments contain, even in the form of embryonic vestiges, all of the structures which were typically present in the anterior segments of the " ancestral vertebrates. On this account it has been extremely difficult to reconstruct complete segments, even where our information as to existing structures is apparently complete. Not only have certain organs with their nerves and nerve centers undergone reduction, or disappeared, but new organs have arisen and old ones have been modified so that new nerves and centers have replaced or overshadowed old ones. Furthermore it appears that as a result of these changes various organs and nerves have been displaced from their proper segments by the crowding of lately formed highly developed structures. We are met by a present condition which is the result of processes of redaction and disappearance, modification and growth, and shifting of position of organs whose real segmental relations are to be discovered only by tracing them back to their primitive condition. Thus the deciphering of the segmental relations, depending, as it does, on the proper interpretation of the homology and phylogenetic history of all the organs, will carry with it the solution of the mijor problems of head morphology. It must not be hoped that the last word upon any of these problems can be said at once. The data are yet too meager for the solutitDU of this greatest problem of the evolution of animal structure, in spite of the efforts of a large number of workers directed to it during several decades. What the writer hopes to do in the present paper is: first to apply a new method for the interpretation of head segmentation which recent work on the nervous system has made possible ; and second to point the way to some profitable lines of investigation.

b. Nerve components.

The function of the nervous system to coordinate and direct all the organs of the body requires a definite and constant relation of its constituent parts to the several tissues and organs. The structural and functional relationships within the organism become impressed upon the nervous system and the

178 Journal of Comparative Neurology and Psychology.

arrangement of its cells and fibers serves as a guide to these relationships. Further, by a study of its residual or vestigeal structures and its ontogeny in various classes of animals we gam hints of many of the past relationships between the other organs. On account of its function as a go-between, the nervous system becomes for us an interpreter. The truth of this has long been realized but it has become of practical value only since the work of Osborn, Allis and Ewart led the way to the development of the theory of nerve components by Strong, Herrick and others. According to this theory, those fibers in the cranial nerves which supply the same kind of end organs enter the same or comparable regions in the brain. For example, all general cutaneous fibers (free nerve endings in the skin) enter the spinal V tract in the brain and the dorsal tracts in the cord. It happens that a nerve trunk is commonly composed of two or more sets or kinds of fibers. All those fibers in the several nerves which have the same central and peripheral connections are said to belong to one and the same system of nerve components.

Although this theory concerns itself directly with the analysis of peripheral nerve trunks, its value rests upon the existence of functionally distinct types of end organs on the one hand and distinct nerve centers in the brain on the other hand. The analysis of nerve trunks into nerve components necessarily implies the analysis of the nervous system as a whole from the same point of view, that is, on the basis of function. In this new way of looking at the nervous system the brain, which from the standpoint of structure has always been regarded as the most complex and obscure portion, becomes the most illuminating — and this just because of its complex relations. Three considerations are important in this connection. First, the reasoning on which the theory of nerve components rested could not be made good unless it was shown that the fibers of two components were independent in their central relations. For example, unless it were clearly shown that the centers in which the general cutaneous fibers end are structurally and hence functionally distinct from those into which sensory fibers

Johnston, Morphology of the Head. 179

from the mucosae enter, we could not assume that the observed distinct peripheral course of these two sets of fibers had anyfunctional significance. If it had proved that the fibers from the skin and those from the mucosae entered the same centers and had the same primary and secondary connections within the brain, we should have been compelled to conclude that sensory impulses from the skin and those from the mucosae would bring about the same reactions. As a matter of fact it has proved not only that the primary centers of these two sets of fibers are distinct, but that their secondary and tertiary fiber tracts remain separate and distinct. Hence it may be concluded that any fiber bringing impulses into the cutaneous center must set up reactions which characteristically follow cutaneous stimulation. The same reasoning may and must be applied to each system of nerve components. Furthermore, the central relations of any system of nerve components when certainly known serve as a starting point from which one may reason back with perfect validity to the homology and probable function of the organs innervated by that system. To illustrate by a case which is still under some dispute, it was by this method that the writer concluded that the end buds belonged to the visceral sensory system before their gustatory function had been proved by extended experiment {6"] , 49).

Second, the brain sometimes gives clues to primitive functional relationships of which the peripheral nerves no longer bear traces. In certain segments of the head this or that nerve component is wholly lacking owing to the disappearance of the structure which it originally innervated. Yet in these same segments the brain retains in some cases the center or column which this component should enter. A simple example of this is the presence of the somatic motor column throughout the medulla oblongata even in those forms in which several segments are without any other vestige of ventral motor nerves. The cases of the cerebellum and tectum opticum have been considered at some length in previous papers {6'] , 68, 69, 70), and these and other cases of the same sort will be treated further in the present paper. Wherever the brain contains a

i8o Journal of Comparative Neurology and Psychology.

representative of one of the primary columns there is presumptive evidence that in the primitive vertebrate the corresponding peripheral structures were present in the same segment.

Third, in cases of extreme modification of other systems of organs and of the peripheral nervous system the brain retains structural features which serve to point the way to the interpretation of new structures. The changes of structure and function, so far as yet observed, are less fundamental than the primitive functional divisions of the nervous system. "New structures" arise as modifications within one division, not by structures belonging to one functional division taking on functions belonging to another division. It does not happen, for example, that cutaneous structures (sense organs, components, centers, or fiber tracts) are so modified as to serve visceral functions, or vice versa. This conservatism of the brain aids in the interpretation of structures which otherwise would be very difficult to understand. By this method of reasoning the evidence that the ear is a part of the lateral line system has been completed, and it has been shown that both ear and lateral line system have arisen by modification within the primitive general cutaneous division of the nervous system. The principle will be applied in this paper for the interpretation of the eye and other organs. It is perhaps needless to say that this principle has been only slightly tested and it is not yet known whether it is universally true. While it promises to be of great value, it must be used with extreme caution.

c. Functional divisions of the nervous system.

Such considerations as the above led the writer some time since to offer a scheme of the primitive functional divisions of the nervous system. This theory is not only in accord with the theory of nerve components but is an extension of that theory so as to give due recognition to the central organ. It is unnecessary here to enter into explanations or arguments in support of this theory of functional divisions beyond the matter contained in previous papers (69, 70). Here a concise outline of the functional divisions is given, with their subdivisions and

Johnston, Morphology of the Head. 1 8 1

constituent elements. This outline differs from the one previously published chiefly in that the visual and olfactory organs and the sympathetic system are assigned to definite places in the scheme, a thing which was not attempted before.

A. Somatic sensory division.

1. General cutaneous subdivision. Consists of :

free nerve endings in the skin,

general cutaneous system of components,

dorsal tracts of the cord, spinal V tract in the medulla oblongata, together with their accompanying nuclei : the dorsal horn, nucleus funiculi, nucleus trigemini spinalis, acusticum and cerebellum,

secondary tracts and centers : internal and external arcuate fibers forming the tractus spino- and bulbo-tectalis (fillet), tectum opticum, colliculus and other nuclei,

tertiary tracts to motor nuclei and coordinating centers.

2. Special cutaneous subdivision. Consists of :

neuromasts (acustico-lateral sense organs),

neuromast components,

spinal VIII tract and nucleus, nucleus funiculi, acusticum, cerebellum,

secondary tracts and centers and tertiary tracts as in i,

the cochlea, its nerve and centers in higher vertebrates.

3. Special sense organs belonging to the somatic sensory division.

lateral eyes. Consist of:

retina, which includes the equivalent of sensory ganglion, nerve component, and primary brain center,

optic tract and tectum opticum corresponding to the secondary tracts and centers of the cutaneous subdivision.

pineal eyes. (Compare Sec. 13 below).

B. Splanchnic sensory division.

1. General visceral subdivision. Consists of:

free nerve endings in the mucosae,

fasciculus communis system of components,

Clarke's column or its equivalent, nucleus commissuralis Cajal, lobus

vagi, and lobus facialis, secondary vagus tract and its continuation in the cord, = the direct

cerebellar tract in higher forms, secondary vagus nucleus, ^ end nucleus of direct cerebellar tract in

the vermis of higher forms, tertiary tracts not well known ; part probably run to the thalamus or

hypothalamus in lower vertebrates.

2. Special visceral subdivision. Consists of:

end buds (taste buds),

components and central nuclei and tracts not yet distinguished from those of the general visceral subdivision.

3. Special sense organ belonging to the splanchnic sensory division.

Consists of: olfactory epithelium and nerve, bulbus olfactorius, tractus

1 82 Journal of Comparative Neurology and Psychology.

olfactoiius, area olfactoria, and tertiary tracts to coordinating centers in the diencephalon. 4. Sympathetic system, afferent portion. An outgrowth or offshoot from splanchnic sensory ganglia which reaches a high specialization in the vertebrate series. Consists of simple splanchnic sensory component fibers and of ganglion cells which together with the efferent portion form a somewhat complete or independent system.

C. Somatic motor division. Consists of;

ventral horn of the cord, nuclei of origin of Nn. XII, VI, IV, III, and

nucleus of somatic motor fasciculus, motor components in ventral roots supplying musculature derived from

the somites.

D. Splanchnic motor division. Consists of:

1. Motor nuclei in lateral horn or intermediate zone of the cord and corre sponding region of the medulla oblongata, motor component in dorsal and ventral roots supplying musculature derived from lateral mesoderm.

2. Sympathetic system, efferent portion. Consists of simple splanchnic

motor component fibers (?) and of ganglion cells and fibers con. cerned with glandular secretion, etc.

In addition to the structures included in the above outHne there are certain brain centers which with their fiber tracts probably serve functions of coordination between the four primary divisions. It is not at present possible to define exactly each of the functional divisions with its secondary tracts and centers, or to understand fully the relations of the coordinating structures even in the simplest vertebrate brains. Some contribution upon this subject will be found in the following pages. With one important exception (see Sec. i 5), however, sufficient is already known for our present needs in discussing the homology and segmental arrangement of the organs of the head.

d. Bearing upon the subject of head morphology.

The essence of the theory of functional divisions is this : Each system of organs in the vertebrate body is connected by definite fiber paths with certain nerve centers which direct and coordinate the actions of those organs. Those peripheral and central nervous structures which are related to a given set of organs form a structurally distinct portion or division of the nervous system extending through the length of the animal and represented in successive segments by serially homologous

Johnston, MorpJiology of the Head. 183

structures. The coordination of action between systems of organs having different functions is effected through central connections between these functional divisions. To what extent direct, simple or short connections between the centers of different divisions serve this purpose is not yet fully understood. But at least in some cases it appears that the divisions remain separate as far as their tertiary tracts and coordination is effected through special structures. It follows that these longitudinal divisions of the nervous system are fundamental in character and are much more significant than the transverse divisions (fore-brain, hind-brain). Each functional division is represented in successive segments by serially homologous organs, while the transverse divisions of the brain are for the most part merely mass portions which have little segmental value. It remains to show the bearing of this view of the nervous system upon the problems of head morphology.

A serious attempt can no longer be made to determine the segments of the vertebrate head by a study of the structure and development of one system of organs alone or chiefly. The method is limited because the set of organs selected for study is not coextensive with the whole organism. One quasi exception to this exists. Owing to the necessities of the case the nervous system is represented in every segment of the organism. Each system of organs has its representation in the nervous system ; if we can read aright, in the segmentation of the nervous system we read the segmentation of the organism.

Reasoning in some such way as this, some authors have attempted to determine the segmentation of the head by means of the central nervous system alone. While the central nervous system at an early stage of development may show the number of segments of the whole head, it can show little as to the relations and arrangement of other organs, what organs are present in given segments, etc. Further, any study of segmentation from the nervous system alone is met by the difficulty that, although some part of the nervous system must be present in each segment, one or more of its main divisions may be absent. What part of the nervous system is represented in

184 Journal of Comparative Neurology and Psychology.

a given segment depends upon what other organs are present in that segment. This very obstacle suggests the way in which the nervous system may be used with special profit, namely, as a key to the interpretation of the facts secured by the study of the other organs. From the point of view given by the theory of functional divisions we get an advantage in the study of head segmentation which may be stated as follows. The scheme of functional divisions serves to define a complete or typical segment and so enables us to determine whether all the proper parts are present in a given segment. A complete segment of head or trunk of a typical vertebrate should consist of : (a) derivatives from the ectoderm (skin), from the entoderm (gut), from the dorsal mesoderm (somatic musculature), and from the lateral mesoderm (visceral musculature), and (b) the corresponding four main divisions of the ner\'ous system : somatic sensory, splanchnic sensory, somatic motor, splanchnic motor. With this clearly in mind we shall be able to say when a segment is complete, when incomplete, what constituent has disappeared, and for what probable reason. When we can say this for each segment of the head the problem of segmentation will be solved.

In order to proceed by this method it will be necessary first to give a summary of what has been established by previous workers on the various systems of organs. It is not the intention to give a critical review of the literature of head segmentation or to undertake the discussion of disputed points ; but only to give a brief statement of the facts that may be used in attempting to reconstruct the primitive vertebrate head. It is believed that a constructive effort aiming to correlate all the well established facts from a new point of view is more valuable than criticism or the weighing of one set of facts over against another.

2. Number and relations of mesoilennic somites.

Since the work of later authors has given strong confirmation to the main results of van Wijhe (119) we may use his nomenclature in referring to the somites.

Johnston, Morphology of the Head. 185

a. A continuous series of mesodermic somites appears in early embryos of lower vertebrates. The earliest somite to appear lies in the neck region and others are formed progressively forward and backward from this point. The order of their appearance and the continuity of the series justifies the conclusion that the somites of head and trunk belong to a single series.

b. That somite of this series which van VVijhe numbered somite 2 always lies dorsal to and is continuous with the mandibular arch. The relation of the succeeding somites to the branchial arches is more or less complicated. See Sec. 3 below.

c. In front of this lie two somites, the praemandibular, or somite i of van Wijhe, and the anterior head cavity of Platt (119, 124, 106, 57).

d. Somites i, 2 and 3 give rise to the eye muscles in all vertebrates.

e. The anterior head cavity of Platt never persists.

The question whether somite i is composed wholly of dorsal mesoderm and whether the mesoderm anterior to it is dorsal or lateral may be discussed here. It is a question which can scarcely be answered by direct observation. The exact condition of the mesoderm in this region in cyclostomes is uncertain. KuPFFER has described and figured for Ammocoetes (78) well developed vestiges of branchial and aortic arches surrounding the hypophysial canal, the relationships being the same as in the pharyngeal region ; but this has not been confirmed. His argument from the position of these arches that the (whole of the) praemandibular mesoderm constitutes a lateral or branchiomeric segment is negatived by the fact (stated by himself also) that the dorsal part of this mesoderm forms an eye muscle which is innervated by a true ventral nerve in all vertebrates. In selachians also, wiiere the mesoderm is better preserved, the embryonic history fails to give any positive evidence of the existence of branchial arch mesoderm in front of the mandibular arch.

One important consideration is to noticed here, however,, which has a bearing on the whole problem of segmentation. It is that in the anterior part of the head in all lower verte

1 86 Journal of Comparative Neurology and Psychology.

brates the mesoderm is completely segmented transversely at an early stage and that the dorsal part (somite) is not clearly separated from the lateral (branchiomere) at any time. This is always true of somite 2 and the mandibular arch. In selachians the transverse segmentation precedes the separation of somite 3 from the hyoid arch. Even in amphibia (Amblystoma)^ the hyoid arch is connected with somite 3 and is separately split off from the entoderm, having had no connection with the following lateral mesoderm. Further, the anterior head somites develop as gut pouches in cyclostomes in reminiscence of the mode of mesoderm formation in Amphioxus. The late connection of somite i with the entoderm is a slight indication of the same thing in selachians (57, 17). In all these respects the head mesoderm presents primitive conditions. Now if it be thought that one or more gill slits and arches have ever existed in front of the present mouth, it is in no way unreasonable to suppose that some vestige of the branchial arches (lateral mesoderm) may be retained indistinguishably connected with somite i and the anterior head cavity, just as the mandibular arch is attached to somite 2 without definite boundary line. The fate of such lateral mesoderm other than mesenchyme formation is not known.

f. Somite 4 lies mesial to or slightly caudal to the auditory vesicle during its formation and immediately in front of the ganglion of the glossopharyngeus. This is consequently often designated as the first postauditory somite.

g. More or fewer of the postauditory somites become reduced or disappear during development. Petromyzon forms an exception to this and the conditions there as described by KoLTZOFF {76) are sufficiently important to be reported in this connection. Somites 4 and 5 become divided into mesial and lateral parts which are completely separated in somite 4, but in somite 5 remain connected by a slender bridge. The mesial part of both of these somites lies close to the chorda and later

' References to Amblystoma are based on my own observations not before published.

Johnston, Morphology of the Head. 187

disappears, that of somite 5 the more slowly. The ganglion of the glossopharyngeus crowds outward, downward and forward between the mesial and lateral parts of somite 4, while the vagus ganglion assumes the same relation to somite 5. The lateral part of somite 4 sends dorsal and ventral buds forward above and below the ear. Somite 5 sends a dorsal bud above the ear and similar buds follow from somite 6 and perhaps somite 7. These dorsal buds of somites 4, 5, 6 (and 7) form the continuation of the body muscles on the top of the head as far forward as the olfactory pit. The great extent anteriorly of this musculature in Petromyzon is connected \\\X.\\ the buccal apparatus. Such dorso-lateral portions are found only in somewhat more caudal somites in selachians: somite 7 in Acanthias (58), and Pristiurus (119); somite 8 in Spinax(i2) somite 12 in Torpedo (112). The absence of this lateral part in the region of the cranial nerves accounts (Sewertzoff, Koltzoff) for the position of these nerves outside the myotomes. In Petromyzon the lateral portions of the myotomes cover the cranial ganglia just as the trunk myotomes cover the spinal ganglia in all vertebrates. P'rom this it seems that no somite completely disappears in Petromyzon but the myotomes of the adult represent the full quota of muscle segments.

h. Some of the myotomes in the caudal part of the head region send ventral muscle buds down behind the gills and forward ventrally to form the subbranchial musculature. In selachians this begins with myotome 8 or 9, in Petromyzon with myotome 10. It seems that the number of gills determines which myotome shall be the first to contribute to these muscles. As the number of gills decreases more anterior myotomes are able to send buds into the subbranchial region.

z. Each of the myotomes is supplied by a true ventral nerve root. The number of these roots in the postauditory region of the adult depends on the number of persistent myotomes, and in the embryo transient roots appear corresponding to transient myotomes. A more detailed treatment of this relation will be given farther on (Sec. 12).

j. There appears in all gnathostomes a variable amount of

1 88 Journal of Comparative Neurology and Psychology.

shifting forward of myotomes in the neck region following upon the reduction of postauditory myotomes.

k. The number of somites in the head can not be stated for vertebrates in general, since the caudal limit of the head varies from group to group. This will be further discussed below (Sec. 12).

J. Branchial apparatus and lateral musculature .

The position of the gill slits with reference to the m}-otomes has been one of the most vexed and vexing questions of head morphology. The most direct and natural assumption is that dorsal and lateral mesoderm had primitively a common segmentation and that the gill slits were situated between the mesodermic segments. Each branchial arch would thus be continuous dorsally with a myotome. This condition, if it ever existed throughout the branchial region, is now preserved only in the case of the mandibular arch and somite 2, and in some forms in the case of somite 3 and the hyoid arch. Caudally to this the branchial arches bear no definite relation to the somites, in most vertebrates. This is certainly due in part at least to the reduction and shitting of postauditory myotomes mentioned above and to the growth and expansion of the branchial apparatus itself. These two processes result in a shifting forward of myotomes and backward of gill pouches and arches, not only in relation to one another but also in relation to other organs (58, 60, 12). Since such shifting is observed in the ontogeny it may be expected that the primitive condition would be found in very young embryos. In most vertebrates, however, such a direct correspondence between somites and branchial arches as was supposed above does not exist at the time when the structures in question are formed. These facts have led some authors to the conclusion that the branchial segmentation is independent of that of the dorsal mesoderm.

The writer is unable to harmonize this conclusion with the arrangement of the cranial nerves. Also, this conception implies a distinct and divergent dorsal and ventral segmentation for the typical vertebrate : the mesodermic segmentation proper

Johnston, Morphology of the Head. 189

extends only to the dorsal mesoderm, the segmentation of the lateral mesoderm is determined or dominated by the arrangement of the gills. Aside from the fact that this would scarcely be expected when we consider the character of invertebrate segmentation, the implications are inconsistent with the facts of segmentation in Amphioxus. Here the whole mesoderm is at first segmented and the lateral segments secondarily fuse into continuous lateral plates. A reminiscence of this is preserved, as we have seen above, in the head mesoderm of cyclostomes, selachians, and amphibia. If this be taken to indicate the primitive condition in vertebrates, then the idea of an independent branchial segmentation is hard to accept. It seems much more probable that the position of the entodermal gill pouches and of the ectodermal furrows was at first determined by the law of least resistance, the clefts between mesodermic segments offering favorable places for the breaking through of gill slits. The fact should not be lost sight of that in larval Amphioxus the branchial segmentation agrees with the mesodermic.

Fortunately we have in the work of Koltzoff on Petromyzon direct evidence bearing on this question. The primitive character of Petromyzon gives greater weight to this evidence. The mandibular arch is connected with somite 2 and the hyoid arch corresponds exactly to somite 3. The third, fourth and fifth visceral sacs on account of the thickness of the gut wall, are longer than the corresponding somites so that when five visceral sacs are formed the fifth lies beneath somite 9. While further visceral sacs are forming the gut wall is thinning out and the visceral sacs occupy less space. No visceral sacs are formed behind somite 9, but when the whole number, eight, are formed they occupy the space of eight somites. Immediately behind them somite (myotome) 10 sends down a ventral bud to enter the subbranchial musculature. Then at a later time the development and expansion of the gills carry them back until the last sac lies beneath somites 15-16, and by this movement the subbranchial muscle buds are cut apart from the myotomes 10, II, etc. A complete correspondence between the branchial

1 90 Journal of Comparatwe Neurology and Psychology.

segments and the somites is prevented in Petromyzon by the thickness of the gut wall in early stages. (It is possible that in ancestral vertebrates with less yolk a complete correspondence might have occurred.) With the using of the yolk and thinning out of the gut wall a numerical correspondence becomes possible, and it is important that the gill apparatus never extends beyond the point at which the first subbranchial muscle bud is to form ; — the point, that is to say, which should mark the caudal limit of the gills on the theory of complete correspondence of the two segmentations.

The investigation of other cyclostomes should throw further light on this important question. For the present, the manysided indirect evidence which comes to impress oi,ie who attempts to frame a consistent account of head segmentation as a whole impel the writer to consider that the visceral apparatus and the dorsal mesoderm have a common segmentation. In accordance with this view the following statements may express the most probable conclusions regarding branchial segmentation and musculature.

a. There is essential correspondence between branchiomerism and mesomerism, the mandibular arch corresponding to somite 2.

b. This correspondence probably rests on a primitive complete segmentation of dorsal and lateral mesoderm, tlie gill slits being formed in the clefts between the mesodermic segments. Such a condition persists in the case of the mandibular and hyoid arches and the first and second visceral sacs in some vertebrates.

c. The lateral mesoderm forms visceral muscles moving the gill arches and the lower jaw. The nerves supplying muscles derived from the lateral mesoderm arise from the lateral motor nucleus in the medulla oblongata and are properly known as viscero-motor nerves.

d. Similarly the trapezius group of muscles are derived from the lateral mesoderm and are innervated by viscero-motor nerves. On this ground the pectoral girdle is perhaps to be regarded as a modified branchial arch. This is further supported

Johnston, Morphology of the Head. 191

by the fact that in urodeles (Necturus, 108 ; Amblystoma) there are a large number of pouches in the wall of the gut in the trunk, segmentally arranged and resembling the gill pouches at the time of their formation. These are probably to be interpreted as vestiges of gill slits which formerly existed through a considerable portion of the trunk. Such seems to be the case in Bdellostoma stoiiti, where Price (109) has described thirtyfive gill slits in the larva, the last 10-14 o^ which persist as the adult gills. The last gill slit lies in the same segment with the twenty-ninth spinal ganglion. The writer does not think that the great number of gills in Bdellostoma all belonged originally to the head and that their extension into the trunk is due to the shifting forward of trunk myotomes into the occipital region. The simpler explanation is much more probable ; namely, that there has been a reduction of the anterior gills on account of the parasitic life of the animal. The persistent gills really belong to the trunk and are homologous with the gut pouches seen in the trunk in urodeles.

e. The number of gill slits in existing vertebrates varies. At the so-called critical stage in Amphioxus there are 14, and this is usually regarded as the primary number for this animal. In Bdellostoma stouti 35 are formed, 10-14 persist ; in other species of Bdellostoma there are C-j, 11-12, and 14 gill sacs; in Petromyzon 8 larval, 7 adult ; in Heptanchus 7 ; in Hexanchus 6 ; in most selachians 5 ; in amphibia 4. It is not possible to draw from these figures any conclusion as to the typical number of gill slits for vertebrates. The writer thinks it very probable that the number of gills in primitive vertebrates was large and variable, even within the same species.

f. The lining of the pharynx is innervated by visceral sensory (fasciculus communis) components found in the VII, IX, and X cranial nerves.

4. Segmentation of the central nef^vous system.

Owing to numerous modifications of organs in the specialized head, the cranial nerves have become so changed from their primitive condition that the study of their segmental ar

192 Journal of Comparative Neurology and. Psychology.

rangement has led to the most various results. Since the brain is a single continuous organ, the study of its segmentation would appear to be a comparatively simple matter. It has long been known that the spinal cord and medulla oblongata show a segmentation in early embryos which corresponds to the mesodermic segmentation. Indeed, the segments of the cord and brain were early regarded as merely the effect of mechanical pressure from the mesodermic segments. It has been clearly shown, however, that the transverse divisions are true segments, due to a definite arrangement of elements in the brain wall and often appearing before the mesodermic segmentation.

In spite of the apparent simplicity of the case, the study of the brain has led to scarcely better agreement as to the 7iumber of segments than the study of the cranial nerves. A summary of the findings of various workers is given in the papers by LocY (125) and Neal (97). In the hind brain, where the neuromeres are relativel}' distinct, they have been studied by numerous authors and in all classes of vertebrates. The disagreement among authors is due to the lack of a definite caudal limit to the brain and to differences in fixing upon the first neuromere to be assigned to the hind brain. The differences <:an be harmonized, however, by reference to the position of the nerve roots and ear, and by considering the shifting of nerve roots which has taken place. This discussion properly belongs to a future section (Sec. 9), but certain facts may be stated here. Accepting Hoffmann's enumeration which later work shows to be correct, rather than that of Okr and McClure, there are in the hind brain of reptiles (Orr, Hoffmann, McClure), birds (McClure, Hill), and mammals (Zimmermann) seven neuromeres ; with the 2d, 4th, 6th and 7th of these are connected the Nn. V, VII-VHI, IX and X respectively. There are thus two neuromeres, one between the Nn. V and VII and one between the Nn. VII and IX, whichare without nerve roots. For convenience such neuromeres will be referred to in the following pages as "blank" neuromeres. In fishes and amphibia there is but one blank neuromere, that between Nn. V and VII, and there is less agreement among authors as to the posi

Johnston, Morphology of the Head. 193

tion of nerve roots connected with the hind brain. This will be discussed below (Sec. 9).

The segments lying in front of the cerebellum are variously interpreted owing to the comparatively early appearance of a segmentation corresponding to the permanent brain vesicles. The segmentation of the brain tube in very early stages, before the formation of these vesicles, has been studied by ZiMMERMANN ( 1 24, rabbit, chick, selachians), Waters (ii8, teleost), LocY (125, selachians, with supplementry observations on the chick and amphibia), and Hill (52, teleost and chick). All these authors studied the neuromeres in all regions of the brain before the closure of the neural tube, in which stage those of the hind brain had already been seen by Kupffer and Froriep ; and Locy and Hill have traced the continuous history of the neuromeres to the time of their merging into the permanent brain vesicles. The work of these last two authors is evidently most painstaking and their results are so complete and so far in agreement that they may be taken to represent the present state of knowledge of the neuromeres. They agree in assigning three neuromeres to the fore-brain and two to the midbrain. The writer may say that at first he was strongly inclined against the conclusions of these authors on account of the unexpectedly large number of segments recognized in the anterior part of the brain. However, he has lately made some observations on amphibian embryos {Amblystoma pjuictatwn), repeating and extending those made by Locy, and although these observations are not yet complete they seem thus far to confirm Locy's work. Further, it will appear in the course of this paper that the point of view from which the subject is approached leads to the acceptance of the number of neuromeres described by Locy, as necessary to account for all the nerves and sense organs connected with the brain.

Neal (96, 97) has studied the neural segmentation in selachian embryos and does not accept Locy's interpretation of the early segments in the fore part of the brain as true neuromeres. He has called especial attention to the existence of a blank neuromere between Nn. V and VH and his description enables

194 Journal of Comparative Neurology and Psychology.

us to determine that this blank neuromere is the one following that which gives rise to the cerebellum, while in higher vertebrates it is the second one behind cerebellum.

5. Typical sense organs of vertebrates.

According to the recent work of Dogiel (24), Amphioxus possesses sense organs which in structure are comparable to the sense cells of invertebrates on the one hand and on the other to the olfactory organ of craniates. Sense cells of similar kind have since been described by Marenghi (88) for Petromyzon. From his own study of the epidermis of Petromyzon the writer is led to think that Marenghi's interpretation of these elements is more or less doubtful. The discovery, however, if it should prov^e correct, is most interesting and important and confirmation of Marenghi's description is very much to be desired. All other sense organs of craniates are apparently distinct from any found in acraniates or invertebrates ; namely, eye. pineal org'ans, ear and lateral line organs, and end buds or taste buds. The specialized sense organs found in higher vetebrates need not be considered here.

The literature on the morphology of the sense organs in the skin and mucosae has been reviewed elsewhere (70), with the result of showing that they fall into two categories : the end buds and taste buds belonging to the visceral sensory division, and the neuromast organs (ear, canal and pit organs, amp\illae, Savi's vesicles) belonging to the cutaneous or somatic sensory division. A reprint of the evidence on which this analysis is based would be out of place here, but as some new facts and discussions of this subject have since appeared it will be of use to print a tabular comparison of all cutaneus organs, their functions and nerve connections. The table is on an adjoining page (Table A).

The suggestion has been made by Allis (4, 5) and Coggi (19, 20, 21) that the ampullae of Lorenzini are homologous with the end buds of teleosts. Three supports are offered to this hypothesis. One is based on the assumption that the center for fibers supplying ampullae in selachians is the same as

Johnston, Morphology of the Head.


the center for fibers supplyinp^ end buds in teleosts. As has been pointed out in numerous places (42, 73, 65. 6^ , 68, 70, 71, 63, 49) this assumption is based merely on the misapplication of the term "lobus trigemini" by Mayser (90) to the an




End buds

Sense cells slender, rod-shaped, reaching full depth of epidermis


Pit and canal organs, Ear.

Ampullae, nerve sacs.

Vesicles of Savi.

Sense cells pear-shaped, not reaching full depth of epidermis. Supporting cells rodshaped. Merkel (91, comparative), Maurer (89, teleosts), Herrick (48, teleosts),

Johnston (68, Petromyzon). Innervated by communis components. Herrick (49) and

others. Fibers slender.

Innervated by various rami of "lateral line

nerves." Many authors ; see review in (70) and (49) .

Fibers characteristically thick.

Strong (114), Herrick (46), and others.

Central endings in lobi facialis (and vagi). Herrick (47, 48, 49),

and others. Secondary tract is the secondary vagus tract. Uncrossed (67). This tract ends in the secondary vagus nucleus beneath the cerebellum (67). Function is taste ; respond only to chemical stimuli. Herrick (49).

Distributed in mouth and branchial cavities, and over head, body and fins; in most forms without regular arrangement. Always superficial.

Central endings in acusticum, lobus lineae lateralis, cerebellum, and nuc. acust. spinalis. Many authors; see (70) and (49); Strong (ii:;).

Secondary tract is the tr. bulbo-tectalis, formed by internal arcuate fibers. Crossed. (47, 67,68).

This tract ends in the tectum opticum (47, 67, 68).

Function of equilibration and hearing. Respond to changes Probably Probably

of pressure in sur- . the same. the same.

rounding fluid.

Many authors, see (70); Parker (ioi, ,102). In canals or rows of

pits. Regular arrangement in



Appear in rows related to canals and pits. Sink in, long tubes, collected in groups.

Always depressed below the surface. (One exception, 48).

Separated from the surface.

[g6 Journal of Comparative Neurology and Psychology.

terior end of the lobus vagi (lobus facialis or lobus impar). The term "lobus trigemini" was commonly applied to the most dorsal portion of the tuberculum acusticum in selachians and ganoids. This lobe was the center for part of the lateral line fibers and has since come to be called appropriately the lohus lineae lateralis (65, 49).

In teleosts, where the facial lobe is enormously developed, it overtops the tuberculum acusticum which has no lobus lineae lateralis. Mayser was misled by the gross anatomy into applying the name "lobus trigemini" to the facial lobe. There is no relation, connection, or similarity between the lobus lineae lateralis in selachians and ganoids and the lobus facialis in teleosts to which for a time the same name was erroneously applied. The homologues of both exist side by side \x\ cyclostomes, selachians, and ganoids and their distinctness and dissimilarity can be easily demonstrated.

The second support is that the ampullary pores of the ophthalmic row stand in the same position relative to the supraorbital canal as do the first formed end buds in Amia. The writer is unable to see that the mere position of a cutaneous sense organ is a safe guide to its homology. It certainly can not have much weight in the absence of any evidence from the side of structure, innervation, or function.

Third, Allis and Coggi find evidence of homology between end buds and ampullae in the statement made by WiedERSHEiM and others that all cutaneous sense organs, including end buds, have the same fundamental structure and that the different kinds of organs pass into one another by gradual modifications. The writer has objected (71) that this statement is not borne out by the descriptions of these organs by various workers* but that there is a fundamental difference in their structure as shown in the above table. Maurer (89) has been quoted (5) as witnessing the similarity of structure. I am unable to see that the paragraph quoted (p. 300) offers such evidence. The context shows that in zvriting that paragraph Maurer had in mind only the various forms of acknowledged lateral line organs. In fact the whole paper give one the im

Johnston, Morphology of the Head. 197

pression that Maurer is not primarily interested in the homology of sense organs. He does not give much attention to end buds and does not seem to realize the possibility that they may not be homologous with lateral line organs. He overlooks end buds in Petromyzon, but clearly distinguishes them from lateral line organs in teleosts (p. 75-76). Only on p. 148, in an unexpected connection, do I find an attempt to compare the structure of end buds and lateral line organs with a view to showing a genetic connection between them. After referring to the distinction made by all authors between "birnformige Sinnes- und fadenformige Stiitzzellen" he cites Leydig's belief that both are sense cells. He then says: "Es ist also sehr leicht moglich, dass lange Zeit diese beiden Zellformen spezifische Sinnesfunktionen haben. Diese Auffassung wird durch die Thatsache unterstiitzt, dass bei Knochenfischen in grosser Verbreitung Hautsinnesorgane auftreten, welche nur aus einer Zellform bestehen und diese Zellen gleichen vollkommen den Stiitzzellen der hoher ausgebildeten Formen." Thus the genetic relation of the two kinds of organs would rest upon the homology of the sense cells of end buds with the supporting cells of lateral line organs. If this is the only way to find a genetic connection between the two kinds of organs, this fact itself would seem to offer the strongest evidence against the proposed homology.

The discussion of the morphological and physiological relationships of the visual and olfactory organs may best be reserved to a later part of this paper (Sees. 13 and 15).

6. The anterior e?id of the head.

The importance of the question of the anterior end of the head and brain can scarcely be exaggerated. It is evident that whether we suppose the anterior end of the brain to be at the so-called lobus olfactorius impar or in the region of the infundibulum, the number of segments falling between that point and, let us say, the III nerve would be greatly altered. Also, a number of important structures which would be ventral in position according to one view would be dorsal upon the other.

198 Journal of Comparative Neurology and Psychology.

Indeed it may be said that our attitude toward the facts which determine the location of the anterior end of the head largely determines our attitude toward the whole head problem.

The questions of the anterior end of the brain and of the head are inseparably connected. The anterior end of the brain may mean two things according as we consider the axis of the brain cavity or the mid-line of the medullary plate. If the neuropore is the open end of a tube which is equally long dorsally and ventrally, then the anterior end of the axis of the cavity would lie in the same dorso-ventral plane as the anterior end of the mid-line of the medullary plate, and that portion of the tube which forms the border of the neuropore would form a ring every point of which falls equally near the anterior end of the head. As such an assumption begs some of the questions at issue it is inadmissible, and we must take the safer course of considering the primitive brain as a flat plate whose anterior end is marked by the point at which its midline passes over into the general ectoderm. It is customary to say that this point is at the lower border of the neuropore. This definition of the anterior end of the brain would be satisfactory if it were shown that the lower border of the neuropore occupies the same position in all vertebrates with reference to other structures of the head and with reference to structures within the brain itself. It is probable that it does not. It is necessary, therefore, to take a wider view and determine the front end of the head as a whole.

The anterior end of the head is marked by thai point at which the brain plate meets the general ectoderm at the same time that it comes into contact with the anterior end of the entodertn in those vertebrates in which the praeoral entoderm is most completely prese^i'ed. This definition does not indicate in advance any particular point that must be accepted as the anterior end of the head. There is in every vertebrate embryo a common point at which brain, ectoderm and entoderm meet. It is a most reasonable assumption that for the given animal such a point marks the anterior end of the head. Following this mode of reasoning Koltzoff {j6) locates the anterior end of

Johnston, Morphology of the Head. 199

the brain in Petromyzon in the infundibular region some distance behind the future chiasma. In some other lower vertebrates the same reasoning would place it in front of the chiasma. The difference in the two cases is found in the length of the praeoral entoderm. Further, it appears very probable that not all vertebrates agree in the position of the neuropore and in its mode of closing. In mammals (36) the lower border of the neuropore at the time when the eye vesicles are forming lies at a point which in the adult is behind the chiasma on the topographically ventral wall of the brain. The point in question, too, meets the entoderm and is met by the hypophysis in process of formation. If this be the anterior end of the brain then the optic vesicle and the olfactory bulb are dorsal organs lying in succession in the order named caudal to the anterior end of the brain. Since the lower border of the neuropore in lower vertebrates is topographically some distance in front of the chiasma, it is evident that in mammals the chiasma must form across a closed up portion of the neuropore which is wanting in the lower forms. Although His did, not deal with this question, a re-examination of his writings with this point in mind leads me to think that his observations were in accord with the statement just made. In his paper on the general morphology of the brain (53) he defines the anterior end of the brain as follows: "Diese basilare Axe [die Mittellinie des Hirnbodens] endigt unzweifelhaft in der Basilarleiste. Versteht man dagen unter Gehirnaxe eine Linie, welche der Mitte der Rohrenlichtung folgt, so wird diese mittlere Axe in einer Ebene liegen, welche die Grund- und die Fliigelplatte des Gehirns von einander scheidet, und ihr Endpunkt trifft die vordere Endflache im Recessus opticus, bez. dicht vor dem Ort des chiasma opticum. Wollen wir zur basilaren und zur mittleren Axe noch eine dritte dorsale Langsaxe oder Langslinie annehmen, so haben wir deren Ende am oberen Rande der Lamina terminalis zu suchen, vor der Stelle, wo die Fissura chorioidea ihren Anfang nimmt." Since the Basilarleiste is situated behind the optic chiasma it would seem that His referred the ventral border of the neuropore to a point topographically ventral and

200 Journal of Comparative Neurology and Psychology.

caudal to the chiasma. His also distincly states that the front end of the entoderm meets the brain at the Basilarleiste and only in the frog does he figure this part of the entoderm (Seesel's sac) as extending forward beyond the Basilarleiste. Although in discussing a question in which phylogeny is of such great importance we should probably not consider the conditions in mammals as of primary significance when they differed from other vertebrates, yet it is obvious that we have reached no concordant results in various classes of vertebrates with reference to the criteria thus far applied.

In order to fix the anterior end of the head in primitive vertebrates it is necessary not only to locate the point of contact of brain, ectoderm and entoderm, but to locate this point in those vertebrates in ivhich the prac oral entoderm is most completely presen'ed.

The animals which best meet the requirements of our definition are the selachians. In these (119, 106, 107, 57, 96, 97, 17) the praeoral entoderm occupies a median position forming a bridge between the anterior head cavity and somite i of the two sides. It extends forward immediately beneath the brain until it meets the ectoderm at the point of junction with the brain plate. Laterally from this point are the ectodermal thickenings which constitute the nasal plates. These are connected across the middle line by a thickened ridge which is continuous with the border of the neural plate before, and for a short time after, the rolling up and closure of the neural tube. Just below this point appears the ectodermal pit which forms the hypophysis. So far as selachians are concerned, the anterior end of the head and brain can scarcely be put at any point on the brain wall which is topographically below or behind this meeting place of entoderm with brain plate and nasal plates — for all such points are behind the anterior end of the entoderm. To seek the anterior end of the brain at any point topographically in front of the lower border of the neuropore implies that some part of the margin of the neural plate may be morphologically anterior. Such a view could be accepted only if it were shown that the neuropore closed by fusion of its lips from below as well as from above.

Johnston, Morphology of the Head. 201

I do not know of any evidence that this takes place in selachians. The lower border of the neuropore lies a short distance in front of the optic stalk, a vestige of which persists as the praeoptic recess (recessus opticus) in the adult. The morphologically anterior end of the brain in selachians is therefore slightly in front of the praeoptic recess and probably behind the anterior commissure. Ventral to this point an ectodermal pit forms the hypophysis and lateral to the same point paired ectodermal patches and corresponding expansions of the brain wall adjacent to the neuropore form the nasal epithelia and olfactory bulbs respectively.

The developement in amphibia is exactly similar to that in selachians. In Amblystoma piinctatuni the continuity of the neural plate with slightly thickened ectoderm at the lower border of the neuropore is present, the entoderm sends forward a median solid prolongation in front of the mouth which meets brain and ectoderm at their point of fusion and overlies the slight depression from or in which the hypophysis is later formed. The median mass of praeoral entoderm corresponds in all essentials to the median, entoderm in selachians described above. In Necturus, Miss Platt (108) describes a median mass of mesentoderm in the same position which she homologizes with the median entoderm between the anterior head cavities in selachians. His's figure of the frog embryo which shows the praeoral entoderm in the same relations has already been referred to. The thickened ectoderm forming the anterior limb of the hypophysial depression in Amblystoma is continuous laterally with the nasal epithelium. When the hypophysial invagination takes place it forms a slender pouch which presses upon the ventral side of the praeoral entoderm slightly behind its tip. All the relations are as in selachians except that the anterior he;^d cavities are not formed and the praeoral entoderm is much shorter. Since the relations just described are most clearly seen at the time when the neuropore is closed and later, there can be no question of the closing of the neuropore from below upwards. Hence the point at which the lower border of the neuropore meets the praeoral entoderm must be taken as the anterior end

202 Journal of Comparative Xeurology and Psychology.

of the head in Amblystoma. The relations described are fuUy maintained until after the optic vesicle and stock are well formed: indeed, the in\-agination of the h\-poph3-sis does not begin until this time. It is therefore easy to determine that the anterior end of the brain lies in front of the optic groove, that is, just in front of the praeoptic recess of the adult. Immediately in front of this is formed the anterior commissure, and above this is the lobus olfactorius impar which marks the dorsal border of the neuropore.

It may be necessarj- to add that when I say just in front of the optic groove I mean to indicate an appreciable distance, a sufficient distance to have morphological value. If I were to attempt to fix ven.' accurately the point in question in Amblystoma I should put it nearer the anterior commissure than to tlie praeoptic recess. Thus it will be seen that while my result is closeh" related to that of His, there is a distinct difference which, as later pages will show, is of considerable importance for segmentation. It should be said, however, that the data on which His based his conclusion would not exclude the result at which I have arrived in the study of amphibia.

Before accepting this result as final, it is necessary to compare lower and higher forms. It seems certain that in the matter of the cephalic extent of the entoderm and mesoderm the selachians are the most primitive vertebrates. In the matter of the branchial apparatus and hypophysis, however, the cyclostomes present more primitive conditions. These condition? are ver\- well known from the work of Dohrn" (25), KuPFFER 178) and Koltzoff iyS). The hj-pophysis reaches a much greater developement than in other vertebrates and in Bdellostoma (Kupffer) communicates openly with the archenteron before the definitive mouth opening is formed. In the front wall of the hypophysial pit are formed at first a median and afterward paired lateral ectodermal thickenings which later form the nasal epithelium. The h}-poph\-sis fuses with the foremost end of the entoderm and this comes into contact with the brain at the infundibular region (Koltzoffj. Here we have relations similar to those in selachians but with important differ

Johnston, Morphology of the Head. 203

ences : the entoderm does not extend far in front of the notochord or mouth but meets the hypophysial ectoderm at a point relatively far back along the lower wall of the brain. To harmonize the facts in cyclostomes and selachians it is necessary to suppose that in the cyclostomes the hypophysis has pushed back further than in selachians while the entoderm and mesoderm have been less well developed, and the point of meeting falls beneath the infundibulum instead of at the lower border of the neuropore. The criticism might be made on this that the writer has arbitrarily chosen the selachians, in which the entoderm is well developed, in preference to the cyclostomes in which the ancient stomodaeum is better developed. That the choice is not arbitrary, however, is clear from two facts. First, not only is there actually a longer praeoral entoderm in selachians, but there are definite paired lateral structures developed (the anterior head cavities) which certainly have some segmental value whether they be considered as somites, lateral plate mesoblast, or aborted gill pouches. Second, the same region of the brain is related to the olfactory placode as in selachians and amphibia. There has actually been a greater reduction of entoderm and mesoderm in cyclostomes and the brain projects farther forward in relation to them. The hypophysis, however, in the absence of the entoderm is able to become deeper. The writer is therefore forced to the conclusion that the selachians and amphibia indicate more reliably than the cyclostomes the anterior end of the head and brain. If we fix the anterior end of the brain in cyclostomes, in harmony with the selachians, at the point of contact of the nasal pit with the brain wall it would be, according to the figures of Kupffer and Koltzoff, about midway between the praeoptic recess and the "lobus olfactorius impar, " or very near to the anterior commissure. So also the figures of Dohrn (25) and His (53).

Since finishing the manuscript it occurs to the writer that the above account might be taken to imply that the entoderm extends into the region which later in the paper is called the prostomium. Such of course is not the intention. The pro

204 Journal of Comparative Neurology and Psychology.

stomial area is extremely short in the early embryo and in fact assumes appreciable size and definite form only after the period at which the relations between the brain plate, ectoderm and entoderm above dealt with are in evidence. The term "anterior end of the head" as used here really applies to the base of the prostomium, or the most anterior point to which the two primary germ layers extend. The anterior tip of the prostomium consists of ectoderm alone and could be located only by reference to the point determined upon as the result of the above examination.

7. TJie segment containing N . III. Segment^.

Neuromere v, somite i, N. profundus, N. III.

To fully describe a given segment it is necessary to determine what representatives are present of the skin, dorsal mesoderm, lateral mesoderm, and gut. and of the somatic sensory, somatic motor, splanchnic motor and splanchanic sensory divisions of the nervous system severally related to these structures. If any of these components of a true segment are absent their absence should be accounted for, and a functional relation between the nerves recognized and the segmental organs present should be clearly shown.

In determining the neuromere to which a somatic motor nerve belongs it should be remembered that these nerves emerge from the brain as ventral roots in the constrictions between the neuromeres. Since a neuromere is considered to be the convex segment between two constrictions, it is most convenient to assign the ventral roots to either the preceding or the following neuromere. In the trunk region where in all higher vertebrates the dorsal and ventral roots unite in the spinal ganglia the ventral root should be assigned to the neuromere with which the dorsal root is connected. The relative position of the dorsal and ventral roots, the neuromeres, and the myotomes will be discussed in a later section (Sec. 19)here the result of that discussion will be anticipated and the ventral roots assigned to neuromeres next caudal to the place of origin of the roots.

Johnston, Morphology of the Head. 205

The somite which gives rise to the muscles innervated by N. Ill is universally agreed to be somite i (praemandibular) of


The third nerve arises from the ventral wall of the midbrain, in some forms far forward toward the point at which the floor of the brain bends down into the inferior lobes. The nucleus of N. Ill is well forward in the base of the mid-brain and that of N. IV follows not far caudal to it. These relations in the adult would seem to show that the third nerve belongs to the first of the two mesencephalic neuromeres, to which, indeed, some authors have referred it. On this assumption the IV nerve would naturally be assigned to the second mesencephalic neuromere. The trochlearis, however, supplies the mandibular somite (somite 2) and the sensory nerve of this segment is the trigeminus. The trigminus can not on any ground be placed farther forward than the cerebellar neuromere. If, then, Nn. IV and V be assigned to the cerebellar neuromere, N. Ill would belong to the second mesencephalic or fifth neuromere of Locy. To account for the position of the nucleus and root somewhat farther forward in the adult we have to consider that the nucleus lies at the anterior border of the large ventral commissure of the cord and brain, which here in the base of the mid-brain reaches its greatest size and importance and is known as the ansulate commissure. This great commissural system is composed fundamentally of the decussating fibers from "tract cells." It has added to it in the base of the cerebellum and mesencephalon many internal arcuate fibres from the cutaneous centers, descending tracts from the tectum opticum to the oblongata and the inferior lobes, the decussation and end nucleus of the bundles of Meynert, and other fibers. Most of these fiber tracts have developed at a period later than the formation of the infundibular depression which sets a cephalic limit to the commissure. The development of these voluminous tracts has crowded the nuclei and roots of Nn. Ill and IV forward. Not only the mere crowding due to the size of the bundles but also the fact that the fiber tracts are related to brain centers situated farther cephalad, has contributed to this result.

2o6 Journal of Comparative Neurology and Psychology.

The sensory nerve of this segment is the ophthalmicus profundus. In selachians (119, 124, 58, 59, 96, 97) the neural crest which gives rise to the profundus and trigeminus extends over the mid-brain and cerebellum. As the ganglionic anlages are being formed, those of the profundus and trigeminus are separate, the profundus anlage being connected with the dorsal wall of the mesencephalon and the trigeminus anlage with the cerebellum. At a later time the profundus ganglion has lost its connection with the mesencephalon and its root fibers pass into the brain with the root of the trigeminus. When the profundus and trigeminus trunks are formed the profundus runs forward along the inner side of somite 2 and the outer side of somite i, the trigeminus runs over the outer side of somite 2. The profundus and trigeminus are therefore one segment apart and belong respectively to neuromeres v and vi.

Essentially the same mode of development is seen in Petromyzon (79, 76), and Koltzoff states that the segmental arrangement is the same as in selachians.

The separation of the profundus from the trigeminus as a distinct segmental nerve is of course not new; it is in agreement with the opinion long held by v. Wijhe, Wiedersheim FuRBRiNGER and other comparative anatomists.

The profundus is the general cutaneous nerve of the segment to which somite i belongs. Its distribuion to the dorsal surface of the head and snout, above and in front of the eye, seems to indicate that it is only a dorsal ramus and also that in spreading over the snout it has annexed an innervation territory not primitively belonging to it. For the bearings of the latter point see the treatment of segments anterior to segment 5, below. When the head flexure is considered it is seen that the area of distribution of the profundus is dorsal; and the almost total absence of ventral structures in front of the mouth, i. e. between the innervation territory of the maxillaris and the eye and olfactory organ, explains the absence of a ventral ramus. The N. ophthalmicus profundus is therefore to be considered as the sole representative of the general cutaneous component in segment 5, to which neuromere v, somite i, and N. Ill belong.

Johnston, MorpJiology of the Head. 207

It is either a dorsal ramus alone or has incorporated in it whatever vestige of a ventral ramus exists.

The visceral elements which should be found in a typical head segment are absent from this one. The visceral apparatus is represented in the most primitive forms by very much reduced and transient structures (see sections on the mesoderm and on the anterior end of the head, above), and the splanchnic sensory and motor nerves with their appropiate central nuclei are wholly unknown in this segment.

In the central nervous system, as has been shown elsewhere (69, 70), the caudal part of the tectum opticum and the nucleus of N. Ill represent respectively the somatic sensory and the somatic motor divisions of this segment. A further treatment of the sensory center will be found below in the sections on the lateral line system and on the visual organs (Sees. 10, 13).

8. Segment 6.

Neuromere vi, N. V, N. IV, Somite 2, mandibular arch, sympathetic ganglion.

In the last section reference has been made to the structures found in this segment, and part of the evidence for their relations given. For the bringing together of Nn. IV and V in one segment we have a long line of investigations including V. WijHE (i 19), Hoffmann (58, 59), Platt (106), Neal {j6, 97), and KoLTZOFF (76). For the relations of somite 2 and the mandibular arch respectively to these two nerves abundant evidence is found in the same papers. It must be noticed that N. V contains both general cutaneous components for the skin of this segment and viscero-motor fibers for the muscles of the mandibular arch. These muscles are formed from the lateral mesoderm and the nucleus of origin of the motor V belongs to the lateral motor column in the oblongata {70). There are found, then, all the elements ot a typical segment except the splanchnic sensory nerve component and its center. The only vestige of this component found in any vertebrate is the sympathetic ganglion which will be treated below in the general section on the sympathetic. The small entodermal area belonging to this segment, if indeed there is any, is supplied by com

2o8 Journal of Comparative Neurology and Psychology.

munis components in the palatine branch of the seventh nerve, and there is accordingly no communis component in the trigeminus and no viscero-sensory center in this segment. In the exceptional case of Amia (3) a large number of communis fibers enter the trigeminus from the facialis root, to be distributed to end buds on the surface of the head. Even here the communis fibers do not belong to the trigeminus segment.

The central nervous structures related to the nerves of this segment include the cerebellum and the nuclei of the spinal V tract, the motor nucleus of N. V which lies a short distance in front of the nucleus of N. VII in the oblongata, and the nucleus of N. IV which lies in the base of the mid-brain. From the adult relations of centers and nerve roots it would be difficult to decide with what neuromere they are associated. This is especially noticeable when attention is called to the fact that the root of N. V is connected with the cerebellar segment in some forms, while in others it is distinctly further caudad. If we look to embryology to determine its primitive location we find the same difference in different classes of vertebrates. Considering the cerebellar segment as Locv's neuromere vi, the trigeminus root is found to have the following location. In both the embryo (76) and the adult (68) of Petromyzon it is connected with neuromere vi. In selachians (124, 106, 58. 96, Fig. 6, 97) it is connected with neuromere vi, in the embryo and shifts to neuromere vii. The actual shifting has been seen and clearly described by ZiMMERMANN. It appears also in Neal's (97) figures 1 5 to 2 1 , In teleosts(52) it has the same relations. In a Salmo embryo of 23 days it is connected with the cerebellar neuromere; at 3 i days it has shifted to the neuromere following. In amphibia (118) it is connected with neuromere vi in the embryo. In the lizard (56, 99, 87), chick (52) and pig (92) the trigeminus is connected with neuromere vii in the embryo. Giglio-Tos (40) describes a shifting of the ophthalmic and trigeminal roots in the human embryo from the mesencephalon to the hind-brain. From this series it must be thought that there has been a shifting backward of the trigeminus during the phylogenetic history of vertebrates. The reason for this shifting is probably to be found

Johnston, Morphology of the Head. 209

in the place of ending of the cutaneous fibers within the brain, although shifting of other structures (to be spoken of below) may have aided or at least offered favorable conditions. In Petromyzon (68) a part of these fibers go directly to the cerebellum, others bifurcate and send one branch to the cerebellum and the other backward, and the remainder turn directly backward to end in the acusticum or the nucleus funiculi. In selachians (63, 28) and in ganoids [f>'j) the destination of the fibers is practically the same. In amphibia (123) a part of the trigeminus fibers enter the cerebellum. In higher vertebrates few if any sensory V fibers go to the cerebellum; they run regularly to the nuclei funiculi. Here, then, is a shifting of the place of central ending of these fibers corresponding to the shifting of the root. The writer would interpret these facts as follows. The general cutaneous fibers of the trigeminus, as of the spinal nerves, typically bifurcated on entering the brain; in primitive vertebrates the ascending fibers ended in the cerebellum or farther forward, the descending fibers passed backward varying distances. The reduction of general cutaneous area in the region of the IX and X nerves due the disappearance of postauditory myotomes, and the developement of the lateral line system resulted in the general cutaneous centers in the cerebellum and oblongata coming gradually into the service of the special cutaneous sense organs (neuromasts and ear). Concurrent with this was the specialization of the cerebellum as a center of coordination. These processes account for the gradual shifting of the general cutaneous fibers of the trigeminus which we observe in vertebrates, until they are all collected in the spinal V tract leading to the nuclei funiculi at the caudal end of the oblongata. Along with these processes, then, and probably as a result of them, we find the root of the nerve shifting back one segment nearer the place of ending of the fibers. The shifting of the whole root of the profundus from the mesencephalon to join the trigeminus is a more conspicuous example of the same process carried out earlier. Both of these shiftings are repeated in the ontogeny in lower vertebrates, as we have seen. On the ground of the source of the V ganglion and the position of its root in

210 Journal of Comparative Neurology and Psychology.

the more primitive forms, and the place of ending of its ascending fibers in fishes and amphibia the writer has assigned N. V to the cerebellar neuroraere, vi.

In connection with the ganglia of the profundus and trigeminus are dorsal-lateral placodes giving rise to cells whose fibers form the Rr. ophthalmicus superficialis and buccalis of the facialis. These are of great importance for the interpretation of the lateral line components and their relation to the general cutaneous system. They will be referred to again (Sec. lO).

The exit of the trochlearis from the dorsal surlace of the brain and its decussating at that point is a puzzle for which no adequate solution has yet been offered. The writer has only a brief suggestion to make. The course of the root fibers dorsally through the brain wall may be due to the course of the fiber tracts among which they run. The position of the nucleus of the nerve relative to the tracts which form the ansulate commissure in typical fishes suggests strongly that the axones from the cells of the trochlearis nucleus may have followed some of these bundles as the path of least resistance. The tracts between the tectum opticum and the base of the oblongata, the tracts between the inferior lobes and the cerebellum, and others, all running more or less dorso-ventrally in the side wall of the brain and decussating ventrally at the level of the trochlearis nucleus — these bundles, which lie ectal to the nucleus of the trochlearis, may have constituted an effective barrier to the axones of the trochlearis iij their attempt to reach the ventrolateral surface of the brain. The axones may then have turned upward along the ental surface of these bundles until they reached the dorsal surface of the brain. If the fibers were thus directed in their course they would be carried to the mid-dorsal line before gaining an exit from the brain and if they then grew straight on they would pass to the opposite side. In Acipenser {6y, p. 139) the root fibers are so intimately wrapped up in a heavily medullated bundle which rises from the region of the nucleus of IV to the dorsal surface of the cerebellum that it is easy to imagine the course of the root being determined by the

Johnston, Mo7'phology of the Head. 2 1 1

cause just mentioned. It is a serious question, however, whether this influence would have been sufficient to determine the course of the root in more primitive forms where the tracts in question are less highly developed. In Petromyzon, although the tracts between the inferior lobes and the cerebellum are very small, the ascending and descending tracts between the tectum and the oblongata are relatively larger and more conspicuous than in Acipenser. There is good reason for thinking that these tracts are among the most primitive and fundamental in the brain, so that their influence would be felt in the lowest forms. It should also be noticed that on account of the late appearance of the trochlearis we may suppose that these sensory tracts may have been already laid down before the fibers of the nerve grew out.

It may be mentioned here that the ventral decussation of a part of the fibers of the third nerve may have been brought about in a similar way. The nucleus of III lies among the bundles of the ansulate commissure and in Petromyzon, where the decussation of the nerve is large and conspicuous, the position of its fibers (68, Fig. 24) suggests that their course may have been determined by the path of least resistance offered by fiber tracts already laid down.

p. The absence of 7iejve roots from one or more hind brain segme?its and the determination of the netiromere concerned.

The only reference which I have seen made to the blank neuromere (Sec. 4 above) as having special segmental significance is by Neal (96, 97) who uses it in support ot the hypothesis that one branchial segment has been supressed, the hyoid arch representing two arches and this neuromere being deprived of nerve roots in consequence. This hypothesis was first brought forward by vanWijhe and has been supported by Miss Platt (106) and Hoffmann (127). Koltzoff {jG) strongly argues that there is no ground in the development of the mesoderm and the gills for the supposition that a branchial segment has been suppressed. The writer is unable to see the advantage of the hypothesis of a lost gill. No reason or cause has

2 [2 Journal of Comparative Neurology and Psychology.

been given for the disappearance of a gill in the middle of the series. The only indication of it is the connection of somites

3 and 4 with the hyoid arch in selachians. This would indicate that the lost gill slit must have cleft the present hyoid arch, as V. WijHE thought. Since it is impossible to explain the nerve relations on this supposition, Neal (96) suggests that the lost branchial segment must have been "anterior to the present hyomandibular cleft," corresponding in position to the blank neuromerc between Nn. V and VII. Even so, the praetrematic and posttrematic rami of VII are made to embrace two gills, the present hyomandibular cleft and the lost one; unless, indeed,

it is supposed that the lost cleft divided the mandibular arch between the mandibularis V and the mandibularis VII nerves. On the ground of mesoderm, however, it is the hyoid that is supposed to have been cleft. Braus' Figs. 5 and 6, PI. XXI, indicate that the connection of somite 4 with the hyoid arch is secondary. One very important feature of this blank neuromere which has not been noticed by previous authors is that it is not the same in all vertebrates, but shifts caudad in higher forms just as the trigeminus does. This does not seem to have , any relation to a lost gill in the hyoid region. Since the hypothesis of a lost gill was brought forward to explain the nerve relations, it loses its interest if these relations are satisfactorily explained on some other basis.

For the explanation of the blank neuromeres the most suggestive facts are found in the position of the auditory vesicle and in the position of the nerve centers related to the cranial nerves. In Petromyzon (Koltzoff's figures), the auditory vesicle when first formed lies oposite neuromere vii and the facial ganglion passes down mesial to its caudal portion. In the adult (i, 68) N. VII passes through the mesial wall of the auditory capsule and out through the ventro-cephalic angle. There has apparently been some shifting back of the auditory sac during the ontogeny of Petromyzon. In selachians and all Gnathostomes the auditory sac is situated farther back, about opposite the tenth or eleventh neuromeres. During the early development in Sgiialus acantJiias (125) the auditory pit

Johnston, Morphology of the Head. 213

shifts back from neuromere ix to neuromere xi. Such an amount of shifting of the relative position of the auditory vesicle and brain can scarcely have failed to influence the segmental position of the nerve roots. Hill's figures (52) also show a slight shifting of the auditory vesicle during development in teleosts. The relative position of the auditory vesicle and the nerve roots also changes, as indicated in part above. From lying opposite a blank neuromere between Nn. V and VII in the embryo of Petromyzon, in the lower Gnathostomes it comes to lie between the roots of Nn. VII and IX which are crowded apart by the vesicle although they remain connected with successive neuromeres. In reptiles, birds and mammals the vesicle is found between Nn. VII and IX and there is a blank neuromere opposite it, in addition to the blank between Nn. V and VII. It is a noteworthy fact that in mammalian embryos the VII-VIII ganglionic complex lies wholly in front of the vesicle, while in. lower vertebrates it lies between the vesicle and brain.

If we couple with this shifting of the auditory vesicle a tendency to concentration of the centers for the V, VII and X nerves, I think we shall find sufficient cause for the observed arrangement of the nerve roots-. The visceral sensory fibers form a descending tract {^fasciciilus cornniunis, f. solitaritis in man) and end in relation to cells accompanying the tract and in the nucleus co.mmissuralis at the caudal end of the oblongata. The general cutaneous fibers run caudally in the spinal V tract. The motor roots are formed by fibers which arise from nuclei caudal to the place of exit of the roots and run forward parallel with the somatic motor fasciculus. Under these conditions it is reasonable to suppose that there may have been a shifting backward of the nerve roots toward their centers. The efficiency of this influence seems especially clear in the case of the communis components, many of which primitively decussated in the dorsal commissure of the oblongata (Sec. 17). As the formation of the choroid plexus crowded this commissure backward to the present position of the commissura infima Halleri,

2 14 Journal of Comparative Neurology and Psychology.

the root fibers would certainly be influenced to shift backward with it.

The history of events which have led to the appearance of blank neuromeres in the oblongata, then, is thought to be as follows. The auditory vesicle attained a size in cyclostomes sufificient to push apart Nn. V and VII between which it was situated. Later in the phylogeny there was a migration of the auditory vesicle backward, which was possibly connected with the reduction of myotome 4 and following. This migration, together with a process of concentration of the nerve centers, resulted in carrying all the nerves in question back one segment; Nn. profundus and V following the shifting vesicle for the advantage of approaching their centers, and Nn. VII, IX and X crowded back by the vesicle at the same time that they come nearer their center. The movement of the vesicle was more rapid than that of the roots and the vesicle came to lie between Nn. VII-VIII and IX. Finally, after this position had been assumed, the further expansion of the vesicle crowded N. IX farther back and caused a blank neuromere at this point in reptiles, birds and mammals.

The shifting from the primitive condition which we have supposed may be summarized thus:

Petromyzon: V not shifted; VII and IX shifted one neuromere, crowding upon X.

Lower gnathostomes: V shifted one, VII, IX and X shifted two neuromeres.

Reptiles, birds and mammals: V shifted one, VII shifted two, IX and X shifted three neuromeres.

We may suppose that there was first, in the lower vertebrates a crowding together of IX and X and then a shifting back of both into those neuromeres whose branchial nerves have been collected by N. X (see Sec. 12).

The blank neuromere was originally the one following the cerebellum and the elements composing the head segment to which this neuromere belonged were: Neuromere vii, somite 3, hyoid arch, somatic motor N. VI, splanchnic motor N. VII, splanchnic sensory N. VII, and a general cutaneous component

Johnston, Morphology of the Head. 215

of N.VII which is unknown. (But see 73). The position of N. VI and the absence of a general cutaneous component have not yet been accounted for. N. VI innervates the muscles derived from somite 3 and its nucleus of origin and root are found at approximately the same level with N. VII in all vertebrates except Petromyzon. Whether its usual position is to be accounted for by supposing that it has shifted back at the same time with the other nerves, or by supposing that it represents the ventral nerves of two segments requires further investigation. I am inclined to think that Neal's reasoning is right in this matter and that the latter is the more probable interpretation. However, no support for this was found by Koltzoff in the development of Petromyzon where we should have most expected it. The writer is convinced that the adult relations of N. VI in Petromyzon are not correctly described and he hopes soon to re-investigate them.

For the explanation of the absence of a general cutaneous component from this segment, also, we must await further investigation. Such a root should be present for the innervation of the skin around and above the hyomandibular cleft in cyclostomes and selachians, and in operculated forms for the innervation of the operculum. The hypothesis may be suggested that in the process of shifting above described the general cutaneous fibers which originally accompanied N.VII remained in neuromere vii and were absorbed by N. V, so that the general cutaneous fibers of three segments are concentrated at the profundus-trigeminus root. The fact that a part of the operculum in teleosts (46) is innervated by the trigeminus is consistent with this hypothesis. If the corresponding area in cyclostomes and selachians should have the same innervation, it would give strong support to the suggestion just made.

10. Segment 7. The aciistico-lateral system.

In the primitive condition this segment was probably composed of the following elements: neuromere vii, somite 3, hyoid arch, somatic motor N. VI, splanchnic motor N. VII, general cutaneous component ( ?), and splanchnic sensory N. VII, with the

2i6 Journal of Comparative Neurology and Psychology.

appropriate centers. In cyclostomes neuromere vii is blank and neuromere viii is related to the structures enumerated. In gnathostomes neuromere viii is blank and neuromere ix is related to the same structures — not that neuromere viii in cyclostomes or ix in gnathostomes lies in the same transverse plane as the other structures of segment 7, but that it is connected with those structures by the nerve roots. In Fig. 10 the relations are shown as in Petromyzon; in Table B the same facts are presented by giving to each element its hypothetical primitive position and stating the shifting which is supposed to have taken place.

In all craniata the neuromasts with their fiber components and centers are added to the above structures. These now require some attention.

The essential unity of the acustico-lateral system both peripherally and centrally and the close relation of these organs and centers to the general cutaneous system is assumed {dj, 68, 70). It is a natural inference that the neuromast system has been derived from the general cutaneous. Further, it has been shown by the work of numerous authors that the whole neuromast system grows peripherally from t/ie vicinity of the auditory pit as a starting point, that the fibers enter the brain at a common point near the auditory nerve, and that the same sensory nuclei serve as the centers for all these fibers. It therefore seems pretty clear that we may look for the origin of the neuromast system as a modification of the general cutaneous system of a given segment or small number of contiguous segments, and that the center of the neuromast system was primitively in about the region of neuromere vii, opposite which the auditory pit now makes its appearance in the lowest known vertebrates.

Although the neuromast nerves have usually been regarded as dorsal rami (when they have not been given the dignity of complete dorsal roots) a difficulty arises in the fact that the ganglion cells of the neuromast rami are derived from ecodermal thickenings, and that the placodes of nerves for the head lines are at first in connection with such undoubted dorsal rami as the ophthalmicus profundus and ophthalmicus superficialis trigemini.

Johnston, Morphology of the Head. 217

The origin of the nerves from ectodermal placodes has led to the development of the hypothesis of primary and secondary sense organs, to which the Hertwigs, Retzius, v. Kupffer and others have contributed. The hypothesis may be stated as follows. The dorsal-lateral placodes in the vertebrate embryo represent a series of primary sense organs in which the sensory cells were of a type illustrated by the cells of the olfactory organ in vertebrates or by the usual sense cells of invertebrates. Such sense cells send their own axones into the central nervous system. There has been a gradual sinking of these sense cells beneath the surface comparable to the migrating inward of sensory epithelial cells in invertebrates. The sinking in of cells from the primary sense organs is repeated in the ontogeny and the cells become the ganglion cells of the lateral line nerves. Then there appear in the placodes newly differentiated sense cells, the neuromast or pear cells characteristic of lateral line organs which constitute secondary sense organs.

It should be noticed with regard to this account of the lateral line organs and nerves, first that there is no direct evidence for this interpretation of the placode cells; and second that it does not imply any close connection between this system and the general cutaneous system. Also, it seems to the writer that a somewhat more simple explanation of the origin and relationship of the neuromast system may be found. The above hypothesis postulates the existence primarily of the sense organs of the invertebrate type, the transformation of these into ganglia, and the new formation of sense organs of the neuromast type. While the transformation of sense cells into ganglion cells is not unthinkable, yet the transformation of a cell possessing a specialized sensory peripheral process into one whose peripheral process is adapted to receive stimuli from another cell with which it enters into complex relations is a very profound change. Moreover, what advantage is gained by the replacing of primary sense cells by secondary sense cells and the addition of one link to the sensory chain is difficult to see. Any change which increases the length of, and the number of relay stations in, a reflex arc would generally be considered to be disadvantageous.

2i8 Journal of Comparative Neurology and Psychology.

Finally, the process by which the primary sense cells are supposed to be transformed into ganglion cells is compared to the process in lower invertebrates by which the epithelio-sense cells wander inward to help form the central nervous system. While this again is conceivable it should be noticed that it is quite the opposite to what we see going on in the vertebrates ; namely, the wandering of cells out from the central nervous system toward the areas of their distribution.

The writer would suggest that the facts may be equally well explained by supposing that the placodes represent material comparable to the neural crest which has separated more completely from the brain tube than has the remainder of the crest, and that the new formation of neuromast cells is the only peculiar feature of the acustico-lateral system. It is not certainly known what nerve components owe their origin to the neural crest but the writer believes that the general cutaneous component is universally regarded as having that origin and the same assumption is at least tacitly made for the general visceral component. It seems probable from our present knowledge that; (i) in Amphioxus all or the greater part of the material representing the neural crest of vertebrates is included within the brain tube ; (2) in selachians, teleosts, and amphibia at least, a part of the cells whose dendrites supply the skin and which are homologous with the general cutaneous ganglion cells remain within the spinal cord during embryonic and in some cases adult life ; and (3) the neural crest presents all degrees of separation from the neural tube during the development of the embryo. In some selachians (59) it separates from the neural tube by a process of hollow budding, while portions of some of the ganglia are formed of cells which migrate out from the lateral surface of the brain at a comparatively late period. In amphibia the relations of the neural crest vary greatly, but in Necturus (108) there is an unusually close connection between the brain and the ectoderm through the neural crest and in the case of the VII ganglion cells cells continue for a long time to migrate from the wall to the ganglion, while the V ganglion is very largely formed by cells coming from the ectoderm. The fact is that there is a

Johnston, Morphology of the Head. 219

very great variation in the relation of the neural crest to the brain and to the ectoderm. The suggestion presents itself that there has been a progressive separation of the material for the ganglia from the brain and, if the acustico-lateral ganglia be included, it appears that the more highly specialized components came off first, the less specialized last.

The neural crest is to be regarded as fundamentally a part of the central nervous system. This is evidenced by the presence of its equivalent within the nerve cord in Amphioxus, by the fact that it is more or less completely included in the neural tube in craniata when the tube first rolls up, and by the fact that a part of its cell's remain within the spinal cord in fishes (38, 39, 66). If this view of the neural crest be accepted, is the neuromast system to be considered as a new structure of peripheral origin, or is it to be brought into relation with other structures represented in the neural crest? The view that it is of peripheral origin gives no explanation of its very intimate connections with the centers of the general cutaneous system in the medulla oblongata and cerebellum. The neuromast system must be supposed to have come into relation with the general cutaneous centers by accident and then to have largely usurped them. Even the suggestion made by Koltzoff {^6, p. 502) that the general cutaneous ganglia represent an earlier category of primary sense organs which have been transformed into ganglia does not help in this matter. Yet the relation of the neuromast and general cutaneous components in their nerve centers is of so fundamental a character that the writer is convinced that it must be reckoned with as one of the most important factors in the interpretation of the acustico-lateral system.

Two structures, both of which are primitive and transient, and both of which point to the ancestral structure of the nervous system, are significant for the explanation of the origin and relationships of the acustico-lateral system: first, the neural plate stage in the ontogeny of the central nervous system, and second, the existence in the spinal cord of fishes of cells whose dendrites are distributed to the skin and whose neurites behave

220 Journal of Comparative Neurology and Psychology.

as do those of the general cutaneous ganglion cells. The first indicates that the nervous system was in primitive vertebrates a specialized area of the superficial ectoderm. The second fact indicates that the primitive form of cutaneous innervation was by means of dendrites of cells in this specialized area growing out to the rest of the ectoderm. This general cutaneous innervation corresponds to the free nerve endings in invertebrates which come from cells in the central nervous system. That this is the primitive form of innervation for tactile and pressure stimuli in vertebrates, I have no doubt. The peripherally situated tactile cells (hair cells) of the highly specialized arthropods do not appear to me to be in the line of vertebrate development. Granted that the sensory cells which supply free nerve endings to the ectoderm in worms may (or must) once have been derived from the ectoderm, we must nevertheless believe that since these cells gained their position in the central nervous system in the worms that relation has been maintained in the line of vertebrate development. Now the relation of the neuromast system in present vertebrates to the general cutaneous system compels us to seek the origin of the neuromast system as a modification of that already established cutaneous system. This seems the more necessary on account of the phylogenetically late appearance of the neuromast system.

If the line of reasoning thus far is good, we have to make only two suppositions for the explanation of the chief relationships of the acustico-lateral system. First, the area from which the system has been developed originally belonged to the neural plate area ; Second, there have appeared in this area a special type of sense cells characteristic of the neuromast organs. It is not hard to conceive that the neuromast area once formed the marginal portion of the specialized ectoderm which constituted the central nervous system. The idea that the neural crest and neuromast area together formed a marginal band along the primitive neural plate is supported by the position of the neuromast area just behind the broad cephalic plate and by the fact that opposite this area there is a long space in which the neural crest contains no cutaneous elements and in which it is partially

Johnston, Morphology of the Head. 221

or completely interrupted at two points. We may imagine, then, that when the neural plate rolled up and sank beneath the surface, the neuromast area was left outside ; and we may find the cause for its failure to be included, as the rest of the neural crest was, in the fact that the special sense cells were already developed in this area. If now we consider that the primitive cutaneous sense cells remained in the neuromast area, as their absence from the neural crest of this region indicates, we have all the essential relations of the neuromast organs accounted for. The further development consisted in the sinking in of the area to form a pit for the betcer protection of the sense cells, or for their better functioning, or both, and the migration of the ganglion cells into the space between the infolded neural tube and the auditory pit, where they found more room. The depressed area became the forerunner of the auditory vesicle and the head and trunk lines of neuromast organs. The general cutaneous sensory cells of the area acquired permanent relations with the newly formed sense cells, and their migration inward to form the ganglia of the lateral line nerves has become a permanent feature of vertebrate development.

The extent of the neuromast area has thus far been indicated only in the most general way for the double reason that we should not expect it to be sharply circumscribed at least in the beginning, and that there are reasons for thinking that it extended farther forward than is generally supposed. The whole of the acustico-lateral system is described as arising in some selachians (93, i25)andteleosts(i2 i, 122) from a single somewhat extensive thickening of ectoderm from which the two head lines grow forward and the trunk lines grow backward. In cyclostomes (79, "j^), however, there are two large and separate placodes, in front of the acustico-facialis placode, with which the profundus and trigeminal ganglia come into contact. One would think from the relations that these two placodes, which give rise to the ophthalmicus superficialis facialis and buccalis nerves, belonged originally to the profundus and trigeminal segments. This idea is strengthened by the fact that in the Gymnophiona (11) a placode appears in the same situation

222 Journal of Comparative Neurology and Psychology.

which invaginates and forms a vesicle in all essential respects like the auditory vesicle. This is then converted into a ganglion which comes into close relation with the maxillo-mandibular ganglion and gives rise to the ophthalmic nerve. A comparison of this account with Coghill's (22) account of the profundus and trigeminus nerves in Amblystoma indicates that the nerve in this case is the ophthalmicus superficialis facialis, and the placode is the homologue of that which gives rise to the R. ophth. sup. VII in cyclostomes, some selachians, and ganoids. That the ganoids are probably to be included here is indicated by the fact that the continuity of the head lines with the auditory placode has not been found in Amia (6). From these facts the writer thinks it probable that (i) the acustico-lateral system was not originally confined to a single segment ; (2) that it extended forward one or two segments in front of the site of the auditory pit ; (3) that it was largest, however, in that segment in which the auditory sac was afterward formed, and that in the other segments it appropriated a greater or less part of the neural crest material.

Two further questions may now be considered: that of the segmental relations of the system, and that of its relation to the general cutaneous nerves. We have come in late years, and rightly so, to deny segmental value to the adult lateral line system ; but here is evidence that primarily it may have extended through more than one segment. The fact that three separate placodes are formed in the lowest forms, cyclostomes, and in some other fishes and amphibia, suggests that the whole neuromast area was not fully continuous at the start but that it was early divided into three parts corresponding to the three segments which it occupied. (Compare 125, p. 578). It seems probable that the supraorbital line, formed from the placode related to the profundus ; the infraorbital line, formed from a placode related to the trigeminus ; and the hyomandibular line, formed from the auditory placode, represented three segments of the primitive neuromast system. The lateral line is to be thought of as a late outgrowth from the auditory placode. I am well aware that this is in opposition to the conclusion which

Johnston, Morphology of the Head. 223

the writer and many others have reached in recent years that the acustico-lateral system should be referred strictly to the same segment with the faciaHs. But the reasoning here employed is in perfect harmony with the spirit and method which have led us to that conclusion, and the present result is based on new facts. It should be said that the writer does not accord to the neuromast system any value to segmental study except as a landmark, and that the truth or error of the segmental arrangement suggested here does not in any way affect the validity or the general interpretation of the acustico-lateral system given above.

By the relation of this system to the general cutaneous nerves I mean the question whether it corresponds to one or more dorsal rami, or whether it has used up a larger part of the general cutaneous component. From what goes before it is clear that the neuromast system must be thought to have absorbed a variable part of the general cutaneous component in the segments to which the profundus, trigeminus, and facialis belong. It may even be that the whole of the general cutaneous component in the facialis segment has been absorbed, and that this accounts for its apparent total absence in that segment. It is of course evident that the communis component in the neural crest could not enter into the neuromast system, and so all schemes of segmentation which raise the VIII or any or all of the lateral line nerves to the dignity of complete segmental sensory roots are faulty.

//. Segments S and folloiving.

The processes which have led to the conditions in the caudal part of the head are discussed in the following section. It will be unnecessary to describe in detail each segment and the changes which take place in each. Figure 10 and Table B show the component parts of these segments in lower vertebrates more clearly than a description in the text could make it. For the facts in the higher vertebrates I can do no better than refer the reader to the extensive descriptions and tabular statements of FuRBRiNGER. These must be read, however, it seems to me,

2 24 Journal of Comparative Neurology and Psychology.

in the light of the general considerations contained in the following section.

12. TJie occipital region and the postenor limit of the head. The constitjition of the vagus. Shifting of organs.

In the occipital region of the head we find the following peculiar condition. (i) A number of myotomes (beginning with somite lO in Petromyzon, somite 9 or 8 in selachians) immediately behind the gills send ventral buds downward behind and then forward beneath the gill sacs to form hypobranchial muscles. These myotomes and their ventral derivatives are supplied by a series of ventral nerve roots from the medulla oblongata which usually join into a plexus known as the hypoglossal nerve.

(2) A variable number of myotomes following the ear (beginning with somite 4) have disappeared in gnathostomes and with them have disappeared the somatic motor nerves.

(3) Throughout this region the general cutaneous, splanchnic sensory and splanchnic motor components are wholly separate from the somatic motor component. This is the condition which persists throughout life in the whole body in Amphioxus.

(4) Usuall}^ these three components are united into a complex known as the vagus or vago-accessory nerve, which supplies the skin, mucosa, and visceral (lateral plate) musculature of the region following somite 4 and the first branchial arch, including trapezius musculature, stomach, etc.

(5) The relative position of myotomes and gills has been so much changed by growth and expansion of one part and shifting of another that it has been difficult or impossible to trace any fundamental relationship between the myomerism and branchiomerism.

(6) The above conditions which obtain in embryos are further complicated in the adults of many vertebrates by the secondary addition of vertebral elements to the skull.

Before undertaking a general account of this region it will be an advantage to give a summary of the conditions in Petromyzon, of which we have a clear and judicial account by Koltzoff.

Johnston, MorpJiology of the Head. 225

As noticed above, the medial parts of somites 4 and 5 disappear without forming muscle, the lateral parts form muscles on the dorsal side of the head. All following somites persist as myotomes. The somatic motor nerves of somites 4 and 5 are not found. The nerve of somite 6 seems to innervate somites 4 and 5 for a time, then disappears, and somites 4, 5 and 6 are innervated in the adult by the nerve of somite 7. Following this each myotome has its ventral nerve.

Following the ear there is in the embryo one dorsal nerve root or ganglionic anlage between each two somites, corresponding to Nn. IX, X, Sp. I, 2, etc., and according to Hatschek (44) each of these is represented in the adult by a dorsal ramus. Now since the vagus root contains both the sensory and motor visceral components for all the following branchial segments and since the dorsal ramus is typically general cutaneous in vertebrates, it seems clear that there is present in each segment behind the ear in Petromyzon a general cutaneous root, the series continuing without interruption into the trunk.

The glossopharyngeus is independent in Petromyzon and contains (68) general cutaneous, splanchnic sensory and splanchnic motor components which form the typical rami for the segment containing somite 4 and the first branchial arch, ramus dorsalis, r. posttrematicus, r. praetrematicus, and r. pharyngeus. The vagus is treated by Koltzoff as constituted of a full branchial nerve similar to N. IX, innervating somite 5 and br. arch 2, and of a ramus branchio-intestinalis. This later consists of a series of ganglia, one over each gill slit, connected by a longitudinal nerve trunk. P'rom each ganglion arise the three rami which innervate a branchial segment in the typical manner. Each ganglion is formed at least in part at the expense of an epibranchial placode. The nerve trunk is continued backward as the ramus intestinalis.

Koltzoff concludes from his study of Petromyzon that the vagus nerve with its ramus branchio-intestinalis is not a compound nerve formed by the fusion of two or more segmental branchial nerves, but is a collector which has taken up the ventral rami of several segrments. In this he seems to the writer to

2 26 Journal of Comparative Neurology and Psychology.

be entirely justified by the facts. The above described conditions seem to indicate clearly that the ancestors of Petromyzon had in the branchial region one complete nerve in each segment of the same constitution as the present glossopharyngeus. It appears that the visceral sensory and motor fibers of the caudal branchial segments, instead of continuing to reach the brain by way of the dorsal roots proper to their segments, have progressively changed their course so as to run to the brain through the root of the next more cephalic segment. This process may have begun as the result of the expansion of the gill sacs and growth of branchial arch muscles which served to crowd the ventral rami of the nerves at the caudal end of the gill region. When the shifting had brought the penultimate gill back to the level of the ultimate nerve root the fibers of the ultimate ventral rami shifted their course to the root of the penultimate nerve. The dorsal ramus was unaffected by these conditions and hence remained independent. As the process went on progressively the branchial and pharyngeal rami became united by the common branchio-intestinal trunk which enters the brain by way of the vagus root. In Bdellostoma (35, 80) it seems that the process has gone one segment farther and the rami of the whole branchial region behind the ear have been collected into a single root which may be supposed to occupy the place of the glossopharyngeus. The writer sees no reason for the supposition made by Koltzoff that Nn. IX and X in Petromyzon may have arisen by the division of such a root.

It seems to the writer that this interpretation of the vagus as a collector, which is so obviously applicable to the cyclostomes, offers the most simple and satisfactory solution of the vexed question of the vagus region in gnathostomes. Without going into a detailed examination of the literature, it is evident that facts directly supporting this interpretation are not wanting in gnathostomes themselves. So-called dorsal hypoglossal roots, either as ganglionic rudiments or as fully developed roots and ganglia, have been described by numerous workers (30, 31, 100, 26, 27, 59, 12). Some of these ganglia undoubtedly lie in segments whose branchial portions are supplied by vagus

Johnston, MorpJiology of the Head. 227

rami. This is probably true of the dorsal anlage ;/, v, w, x described by Braus in Spinax; of two ganglia in Acanthias des-^ cribed by Hoffmann; an of and uncertain number of the anlages said by Dohrn to be present in all his selachian embryos. In the sheep embryo the hypoglossal ganglion lies over the caudal part of the vagus ganglion and its roots enter the oblongata dorsal to and independent from the roots of the vagus. It is evident from our present point of view that these dorsal roots of occipital nerves are composed of general cutaneous elements alone. As in the cyclostomes, the visceral elements have been collected into the vagus. There is therefore no difficulty arising from the presence of dorsal roots in those segments which possess ventral rami from the vagus. In fact, the presence of a dorsal general cutaneous root is to be expected to fill out the complement of functional divisions, and only its absence requires a special explanation. Where it is absent, as it commonly is from a few segments following the vagus root, two explanations are possible. One is that the general cutaneous fibers have been collected into the vagus, which I regard as very improbable since there is no cause for it, mechanical or other. The true explanation probably is that as a result of the reduction of myotomes behind the ear and shifting forward of cervical myotomes, the cutaneous area of the original occipital region has been greatly reduced and the general cutaneous nerve roots have aborted in consequence. The general cutaneous component of the vagus supplies the skin of this region. It may be added here that, as is indicated in Fig. 10, I believe that the shifting of the vagus root described in a previous section has resulted in gnathostomes in the absorption by the vagus root of the general cutaneous root of the segment next following the proper vagus segment. This is an entirely different matter from the process of collecting the visceral nerves. It is rendered probable not only by the consideration of embryonic conditions but by two other facts: namely, the large area of skin supplied by the dorsal rami of the vagus between the vagus and first spinal nerve; and the existence of two distinct dorsal rami of the vagus (114, 46) with general cutaneous

228 Journal of Comparative Neurology and Psychology.

distribution. The nerve which has been absorbed in gnathostomes is the one in cyclostomes which Hatschek calls "Vagusanhang" and which Koltzoff calls the first spinal.

If the vagus nerve really became a collector of the visceral sensory and motor components in all the following branchial segments in primitive vertebrates, the condition of these components in gnathostomes is clear. The number of gills has been reduced and the ramus branchio-intestinalis has consequently lost the corresponding ganglia and rami. It may be suggested here in view of the evidence that a much larger number of gills once existed (Sec. 3), that the ramus intestinalis itself would be regarded as a trunk which owed its formation in the first instance to the collecting process above described. In other words the ramus intestinalis is the collector of all branchial nerves which may at one time have existed behind the present gill region. In accordance with this view the oesophagus and stomach, innervated by the r. intestinalis, would be included in the primitive gill region. Kingsley's argument (74) that the stomach is formed by a rapid expansion of a part of the alimentary canal immediately following the gills is perhaps to be carried one step further by adding: a region which had earlier been reduced by the loss of more caudally situated gills. The effect of the present suggestion would be to regard the distribution of the ramus intestinalis as a measure of the primitive gill region.

The formation of the spinal accessory nerve should be explained in this connection. The visceral motor nuclei are enlarged in the oblongata and this enlargement continues into the spinal cord where it gives rise to the roots of the accessory nerve. This nerve innervates the trapezius musculature which is formed from the lateral mesoderm, and the nerve is homologous with the visceral motor components in the vagus. Its attachment to the vagus is due to processes similar to those which brought about the union of the branchial nerves with the vagus. V It has been "collected" and probably at a time when the pectoral arch was a gill arch. The position of the roots offers no difficulty, since the true vagus roots in some cases have a similar position ; that is, more or less spread out along the

Johnston, Morphology of the Head. 229

medulla oblongata and between dorsal general cutaneous roots and ventral somatic motor roots. The accessory illustrates the method by which the collection of motor roots may take place. The shifting of motor roots obviously can not be due directly to such peripheral causes as the crowding together of rami which explained the collection of sensory nerves. The motor roots must have been collected under the influence of the sensory components and so indirectly on account of the same causes. As the motor fibers grow out from their nuclei in the brain they must follow some path of low resistance in travelling to their muscles. Since the motor fibers develop late they find such a path already provided in the near-by sensory root. The motor fibers follow this and a mixed trunk is formed. When the sensory fibers of a given root shift their course between the epibranchial ganglia and the brain to the root next cephalad, the motor fibers on issuing from the brain find no path in that segment but must turn forward to the next cephalic sensory root and follow it. As this goes on gradually from segment to segment there are formed a number of roots emerging from the cord or brain caudal to the complex and running along the side of the brain to join it. This is the description of the accessory roots. The more cephalic roots have become more compactly arranged and their fibers run backward within the brain, parallel with the somatic motor fasciculus, to their nuclei of origin. This is the description of the vagus motor roots. Both vagus and accessory motor roots are formed by the same process and constitute a graduated series.

This is a very brief summary of views on a complex region to which hundreds of pages have been devoted by observers in this special field. But even from this it must be evident that the region between the head and trunk is one of transition and the attempt to set a caudal limit to the head which shall be valid for all vertebrates is a fruitless one. Reduction and loss of myotomes with loss of somatic motor nerves, reduction of cutaneous area and nerves, growth and expansion of gills followed by their reduction and loss, shifting and reduction of visceral sensory and motor nerves, shifting of skeletal elements, fusion

230 Journal of Comparative Neurology and Psychology.

of vertebral arches with the skull and secondary separation of the same ; all these and other changes have been seen in this region by various workers and perhaps all have taken place in different degrees in one or another group of vertebrates. The region as represented is almost the most variable in development and structure that can be imagined. No one criterion can be taken by which to determine the posterior end of the head in vertebrates. The caudal end of the skull varies aiid is of significance only when interpreted in the light of the nerves and other structures. The use of the first nerve which has both dorsal and ventral roots gives widely differing results, and in no case are we sure that the two roots constitute a complete segmental nerve. Myotomes shift, gills are reduced. In truth, the vertebrate head has no posterior limit, although such a limit may be more or less arbitrarily agreed upon for any individual vertebrate species.

What is needed now in the study of this intricate region is the analysis of its nerves into their components, the discovery of the source of each component and the tracing of its history, and the determination of the segmental position of each nerve and ganglionic anlage with reference to the series of somites. A great number of data are already available for this but nearly all of them must be interpreted in the light of principles which were not appreciated by the original authors. It would not be profitable at this time to attempt to work over this material because there are too few researches in which the myotomes are certainly identified by reference to a standard series such as V. Wijhe's, and because the nerve roots and anlages are not located with reference to the somites. This material has been treated in a comparative manner by Furbringer (35) and his vast work might have cleared up this intricate region if he had understood the relations of the components in the sensory nerve roots. As the writer has pointed out elsewhere (70) he failed to appreciate this. There are indications that the relations in this region are not so complicated and the work of unraveling them need not be so arduous as Furbringer has made it. For Furbringer the dorsal roots of the cranial and spinal nerves are

Johnston, Moj-pJiology of the Head. 231

homologous ; he does not think of the possibihty of two separate kinds of dorsal roots being found in the same segment under some conditions. Therefore if nerves of spinal type are found in the vagus region they must have reached their present position by shifting through the long branchial region. As the above discussion has shown, the dorsal and ventral "hypoglossal" roots need not be considered as spinalajiige nerves. They probably are not equivalent to spinal nerves at all but are only the general cutaneous and somatic motor components of nerves of the vagus region, the visceral^ensory and motor components of which have been collected into the single large vagus root. The presence of these nerves in the vagus region, then, does not require the hypothesis that they have shifted forward from the postbranchial region but is directly opposed to such a hypothesis. So far as these nerves are concerned a shifting forward of trunk elements into the head simply has not occured. The nerves are just where we should expect to find them if no shifting or other modification has taken place. This fact is of course not inconsistent with the observed shifting forward of trunk myotomes and the fusion of vertebral elements with the skull. As stated elsewhere, the disappearance of the spino-occipital nerves is due to the reduction of their innervation territory concurrent with the shifting forward of trunk elements. The structure of the nervous system is opposed, however, to the view which supposes that the vertebrate head once (palingenetic) included all the structures innervated by the vagus and the segments in which they were situated, and that all structures now found in the vagus region and not innervated by the vagus have moved forward from the trunk. Owing to the formation of a rigid cranium to protect the brain and sense organs and the secondary extension of this cranium farther caudad by the fusion of more or fewer vertebral elements to the cranium, there has been a reduction of unnecessary myotomes and a shifting forward of trunk myotomes to gain attachment to the cranium. This may be ground for recognizing a coenogenetic cranium ; but the cranium is not the head. The shifting of any organ is relative to all the others. The expansion of the branchial ap

232 Journal of Comparative Neurology and Psychology.

paratus may be expressed as a shifting backward with relation to the vertebrae and muscles, and when the consequent shifting of the heart and blood vessels and other organs (see Hoffmann, 60) is taken into account it is an open question whether the shifting backward has not been greater than the shifting forward in the vertebrate body. Shifting should be referred to organs which are known to have some cause for movement with reference to all the others, as the heart is pushed backward by the branchial apparatus, and the trunk myotomes are enabled to push forward by the reduction of those in the head. The analysis of the nervous system and the description of the components present in each of the spino-occipital nerves is the best guide to the explanation of this region. Now that we know the comparative anatomy, much more is to be gained by the study of the actual shifting of organs in the ontogeny than by further speculation.

It is probable that all attempts to divide the vertebrate body transversely into primitive regions will have to be given up. The w'hole trend of reasoning in the present paper has been against the idea of palingenetic and coenogenetic head regions (Gegenbaur-Furbringer). Three main objections to this hypothesis may be stated: (i) in Petromyzon the development of the mesoderm clearly indicates the existence in ancestral vertebrates of a continuous series of similar muscle segments throughout the head and trunk. There is no limit to the palingenetic head to be found in the mesoderm. This seems to be the best key to the interpretation of the conditions in selachians and higher vertebrates as well.

(2) It is by no means certain that the pharynx with its gill slits was confined in primitive vertebrates to the palingenetic head. There are indications that the gills once extended far into the trunk and that the specialization of the branchial nerves had its beginning from that condition.

(3) The interpretation of the nervous system from the functional point of view strongly favors the view that the ancestral vertebrate had a very slightly specialized head region. It is certain that the fundamental divisions of the nervous system

Johnston, Morphology of the Head. 233

in the trunk region find their exact counterparts in the pahngenetic and also in the coenogenetic regions of the head, although with various modifications in different classes of vertebrates. There is good reason for thinking that the special sense organs of the head,' their nerves and brain centers, are modifications of one or other of these fundamental functional divisions and not in any sense new structures. The so-called palingenetic head does not havq distinct boundaries and is not constituted of distinct and characteristic structures.

Recently Froriep (33, 34) has stated his objections to the theory of the palingenetic head, insisting especially that it is false to suppose that there was first developed a primordial head and that this was then modified by various processes of reduction, shifting and dovetailing by which the coenogenetic head was added to the palingenetic. The triple division of the body which he adopts in common with v. Wijhe rests on a scarcely better foundation. The regions are ill-defined and the distinctions between them are traceable to processes which are still indicated in the ontogeny of vertebrates so that the divisions can not be regarded as primitive. The achordate and caducichordate regions are specialized through the reduction of the mesoderm and this is due to the development of the definitive mouth (neostoma) and the reduction of the paleostoma.

i^. The morphology of the visual organs.

It is important in approaching the subject of segmentation in the rostral part of the head to have in mind all the structures which have segmental significance. The entoderm and mesoderm of this region have been discussed above. The anterior head cavities of Platt and a median prolongation of entoderm are the only structures of segmental value afforded by these two germ layers. When we turn to the ectoderm and brain we find in addition to the two so-called nerves of special sense, the optic and olfactory, the nasal epithelium itself and a placode which comes to form the lens, two vestigeal nerves, the N. thalamicus of Platt and the N. olfactorius accessorius of Pinkus and LocY, and two epiphysial organs, not to mention several epi

234 Journal of Comparative Neurology and Psychology.

branchial placodes of doubtful character described by v.Kupffer. . Do all of these organs belong to the one segment represented by the anterior head cavities, or are most of them to be regarded as organs without segmental significance? It seems to the writer that the nervous structures represent more segments than have been preserved in the mesoderm. In other words, there are preserved vestiges of nerve structures belonging to segments whose entodermal and mesodermal organs have disappeared for the most part. The physiological relationships and the structure of these organs should give the clue to their segmental relationships. The division of the brain wall of this region into four neuromeres in embryos of selachians, teleosts and the chick gives a clue to the number of segments.

A full consideration of the visual organs is important for any scheme of segmentation. The paired lateral eyes arise from the lateral surface of the brain tube and remain attached by a hollow stalk to the second neuromere of Locy — in higher vertebrates the cerebral segment. The stalk becomes replaced by the optic tract whose primary place of ending in the brain is the optic lobes. Comparative studies (37) show that there are two epiphyses in vertebrates, one caudal to the other, both of which possess the structure of visual organs with nerve cells and fibers often present in the adult. These two organs are descfibed as hollow outgrowths of the dorsal wall of the diencephalon. The anterior one in Petromyzon, Acipenser, and reptiles sends nerve fibers into the ganglion habenulae ; the posterior one in Petromyzon, selachians, teleosts, Amia, Anura and reptiles sends its fibers back to the posterior commissure. It is probable that they end in the optic lobes. Such being the general description of the visual organs, have they any relation to other sensory structures and what is their morphological and segmental significance?

The lateral eyes are physiologically related to the skin and the general cutaneous sensory system. Both retina and skin receive stimuli from the external world and both are sensitive to light. It has been shown experimentally by Parker (102) that the skin of the frog is sensitive to light and that in a degree

Johnston, Morphology of the Head. 235

sufficient to enable the animal to orient itself with reference to the source of light. The skin of many other animals also has been shown to be sensitive to light. It may be added that the retina is sensitive to pressure, sending an impulse to the brain which is there interpreted as light.

There are certain resemblances between the eyes and the general cutaneous ganglia in the mode of their development. In Petromyzon {y6) the eye appears first as a solid outgrowth of the wall of the brain and the cells have the appearance of migrating cells. In this there is a similarity to the mode of formation of the trigeminus ganglion (j6, figs. 4 and 7). In selachians, on the other hand, both the optic vesicle and the ganglia of the cranial nerves appear as hollow buds or outpocketings of the brain wall. This has been mentioned by Hoffmann (59) and a comparison drawn between the eye vesicle and the ganglia in this regard. His figures suggest very strongly that something more than an accidental resemblance obtains. In amphibia (29) the eyes appear first as pigment spots toward the margins of the cephalic plate where a certain rearrangement of cells takes place long before the vesicles begin to bulge out.

From their position in amphibia one would think that the eyes were dorso lateral structures belonging to the same neuromere as that to which the stalk is attached after the vesicle is formed. This appears more clearly in the selachians figured by LocY (125). In all vertebrates the vesicle includes a large part of the lateral wall of the fore-brain and owing to the constriction taking place from above downward the stalk comes to have a ventral position. The mode of formation of the optic stalk strongly suggests that the vesicle was not originally a ventral structure, as it is usually regarded, but rather a dorsal one whose attachment to the brain has moved downward or been constricted downward, as happens in the ontogeny. In Lepidosiren (72), indeed, this process is recapitulated in much fuller form than in most vertebrates. The solid optic process appears at first extending laterally from the dorsal wall of the brain. A cavity appears in it and becomes continuous with the cavity of the tore-brain. "And at the same time, by differential growth, the

236 Journal of Comparative Neurology and Psychology.

attachment of the optic stalk becomes carried downward to the ventral side of the brain." The figures clearly illustrate this as a progression of the optic stalk from the dorsal to the ventral side of the brain. When this is taken together with what is said below about the optic tract I think it will be seen to have considerable significance. At all events the idea that the eye is a dorsal structure involving the dorso-lateral wall of the brain, rather than a ventral one, is of the first importance, if true. I believe that the ontogeny favors this view, although the opposite has been universally held. In correlation with this I wish to point to an important fact in the gross anatomy of the forebrain. In all lower vertebrates there is between the prosencephalon and diencephalon a very slender region made up chiefly of fiber tracts. This slender portion is due to a very deep indentation of the dorsal-lateral wall of the brain above the optic chiasma, and the choroid plexus, is widest at this point. This slender region becomes less noticeable in higher vertebrates, where the fore-brain grows larger and more rapidly, but it is clearly seen in embryos and is marked in the adult by the wide choroid plexus. The facts suggest that the nervous material of the dorso lateral wall of the brain tube has been withdrawn in this region and its place is filled merely by a thin membrane. I believe that this is due to the formation of the optic vesicle and that the nerve centers which are recognized as constituting the retina have been derived from the brain at this point, namely in the caudal portion of the fore-brain dorsal to the permanent attachment of the optic stalk.

If we turn to the central relations we shall find a number of very significant facts which explain this relation of the optic vesicle to the brain wall. The optic tracts form a decussation in the ventral wall of the brain which in lower vertebrates is complete and simple. These tracts then enter the optic lobes which were primitively their sole place of ending. These lobes are also the place of ending of the internal arcuate fibers (tractus bulbo-tectalis) from the general and special cutaneous centers in the medulla oblongata. These tracts from the retina and from the cutaneous centers constitute primarily the only afferent

Johnston, Morphology of the Head. 237

tracts to the optic lobes. The inference is that the optic tracts are to be compared with the internal arcuate fibers. In support of this inference further facts are obtainable in the fish brain. The first is the intimate relation between the two sets of fibers at their place of ending in the optic lobes. The writer has thought that he could distinguish in Acipenser the sets of cells which served as the center for the optic fibers from those which served as the center for internal arcuate fibers. This distinction is at best a difficult one, can not be extended to the whole of the tectum opticum, and is much more doubtful in Petromyzon (68) and selachians (63). It is to be noticed, however, that the structure which corresponds to the colliculus of higher forms is partly differentiated from the tectum opticum in Acipenser and teleosts and is related to the secondary cutaneous fibers only and not to those of the optic tract. From these facts it must be thought that the tectum opticum was in primitive fishes the common center for the optic tract and the secondary cutaneous tracts and that a differentiation into two centers is in progress in the higher fishes.

A second fact is found in the examination of the two chief ventral decussations of the brain, the ansulate commissure and the decussations of which the optic chiasma forms a part. Although both of these (especially the latter) require further investigation, this much may be said : both contain secondary and tertiary somatic sensory tracts destined to the same or similar centers. The most important and primitive tracts in these decussations are:

Chiasma and postoptic decussation. Ansulate commissure.

Secondary: optic tract Tractus bulbo-tectalis (int. arc).

Tertiary : tractus tecto-lobaris Tractus tecto-bulbaris et lobaris

? : tractus lobo-bulbaris Tractus lobo-bulbaris

The correspendence between the second and third elements suggests that the two decussations belong to one commissural system which has been interrupted by the downgrowth of the inferior lobes. Reasons were given in a previous paper (69) for thinking that the inferior lobes are constituted of cells belonging to the category of commissural and tract cells which are found

238 Journal of Comparative Neurology and Psychology.

in the cord and oblongata. Whether the homology is correct or not, it should be noted that the inferior lobes are formed as an expansion of the mid-ventral part of the brain wall, their cavity being the greatly expanded ventral furrow of the third ventricle ventral of the thalamic nucleus of the somatic motor fasciculus. Fig. i shows this relation. An expansion ventrolaterally of this part of the brain wall would very naturally cause a separation of the ventral commissure belonging here into two parts, one running in front and the other behind the inferior lobes. The homology of the tractus lobo-bulbaris is suggested in Fig. 7 by the .symbols used, but the question of this homology does not affect the present argument, since the tractus lobobulbaris is present in both decussations.

The presence of centrifugal fibers from the tectum to the retina should receive explanation in a hypothesis regarding the origin and morphology of the eye. From the study of the cutaneous centers in lower vertebrates the writer has been led to suggest (69) that these centers primitively possessed a co-ordinating apparatus consisting of fibers which arose in the dorsal horn, decussated in the ventral commissure and ended in the dorsal horn of the opposite side at a higher or lower level. It was supposed that these fibers have gradually passed farther forward until they came to form the tractus spino- et bulbotectalis which ends in the tectum opticum. The cause of the shifting forward and concentration of this system of fibers would be found in the development in the cord of the tract cells which serve the same function more directly, and in the growth forward of the ascending branch of the cutaneous root fibers in the dorsal tracts to the medulla oblongata. Upon this supposition it will be seen that the centrifugal fibers in the optic nerve would represent the fibers passing from the cutaneous center of one side by way ot the ventral commissure to the cutaneous center of the other side in a more cephalic segment ; in this case, from the tectum by way of the chiasma to the retina of the other side. The fact that the centripetal fibers of the optic tract pass backward to reach the tectum as a co-ordinating center is explained simply by the fact that the cutaneous center of the only neuro

Johnston, Morphology of the Head. 239

mere rostral to the optic vesicle has never been highly developed and the tectum has early become specialized as a co-ordinatingcenter. Whether this explanation of the centrifugal fibers shall prove to be supported by further investigations of the cutaneous centers, the main fact remains that the tracts which connect the retina with the brain are closely similar to those which connect the cutaneous centers with the tectum, and that the rise of a large general cutaneous nerve (the profundus) in connection with the tectum in the ontogeny indicates that the tectum was originally a primary general cutaneous center.

Following this line of thought it is possible to see in the retina two chief layers which may correspond to the ganglia and central nucleus respectively, and also to draw some comparisons between the cells of the retina and those of the cutaneous sensory centers. The outer nuclear layer with the rods and cones may be thought to represent a general cutaneous ganglion. The outer molecular layer is the place of meeting of the central axones of these ganglionic cells with the dentrites of central cells, and as such suggests comparison with the substantia gelatinosa. The bipolar rod and cone cells of the inner molecular layer correspond to the small cells characteristic of cutaneous centers which in the cerebellum and acusticum of fishes become specialized as granules. The horizontal cells may perhaps be compared with the simple II type cells in the cerebellum and other cutaneous centers. The large cells of the ganglion cell layer correspond to the large cells of cutaneous nuclei which give rise to internal arcuate fibers. The comparison of the afferent and efferent fibers has been made above. It may be said further that there is even a correspondence between the retina and the cutaneous centers in the relative position of the nerve elements. Beginning at the original outer surface of the ectoderm we find in each case the elements in the following order : the primary receptive elements or ganglion cells, the substantia gelatinosa, the small distributing cells of the sensory center, and the large cells whose axones go to other brain centers. These fibers travel at first next to the original inner surface of the ectoderm and as they pass to the opposite side through the ventral com

240 Journal of Comparative Neurology and Psychology.

missure and longitudinally to enter the tectum they remain "outside" of the central grey, that is near the original inner surface of the ectoderm. When we take into consideration the morphology of the optic stalk and the process of the brain ventricle which it contains, the course of the optic tract fibers from the retina to the brain is comparable in every respect to that of the internal arcuate fibers from the cutaneous centers to the tectum opticum. The migration of the optic stalk from the dorsal to the ventral wall of the brain is now readily understood. The retina must be carried out some distance from the brain wall in order to reach the surface of the head. Since the centers are carried out in the retina the stalk consists only of internal arcuate fibers running through the ventral commissure. These shift to the ventral wall as a matter of economy since in their original position they must follow two sides of a triangle. Furthermore, it will be noticed that if the retina is derived from the dorsolateral wall of the brain its inner wall which becomes thin is originally continuous with the choroid plexus and the dorsal border of the outer wall of the retina may be regarded as the dorsal border of the brain wall. In this respect the retina is analogous to the acusticum which forms the dorsal part of the wall of the oblongata, and the choroid plexus of the fore-brain is analogous with that of the hind-brain. We could say homologous instead of analogous except for the presence of the equivalent of the peripheral ganglion in the retina.

Finally, reference may be made again to the fact that a frog is able to orient itself with reference to the source of light which falls on the skin alone. It is even shown (102) that the animal orients itself in the same way by means of the skin as by means of the eyes, namely turns its head toward the light and jumps toward it. This shows a close relation between the central paths involved in motor reflexes in response to impulses received from the skin and those received from the eyes. The writer believes that the same paths were originally used for coordinated reflexes by way of the tectum and the fact that these relations are retained in an animal so highly developed as the frog is a strong argument for the fundamental unity of the

Johnston, MorpJwlogy of the Head. 241

visual and cutaneous apparatus in primitive vertebrates. The specialization of the secondar}' and tertiary centers, visual and cutaneous, we see beginning in the fishes and going on in the higher classes of vertebrates.

We may sum up these several lines of evidence in the following hypothesis. The retitta belongs morpJiologically as ivell as physiologically to the cutaneous sensory system. It cojisists of the general ciitaneous ganglion and corresponding centers belojiging to nenrouiere ii, zuhich have passed out of the brain ivall in order to remain exposed to the light. The optic tract is Jiomologoiis ivith the internal arcuate fibers of the cutaneous centers.

Reference may be made here of a possible comparison of a part of the retina with the acustico-lateral system. It will be seen that in the above there is supposed to be included in the neural tube when it rolls up all that area which at the level of the acustico-lateral anlage is supposed to be left out. The thought quickly presents itself that perhaps the rod and cone cells are homologous with the neuromast cells in the acusticolateral sense organs, the supposition being that both are derived from cells of the original neural plate which have the power to develop into special sense cells. The presence of sense cells in the saccus vasculosus of lower vertebrates [f)'], 9, 8) gives color to this supposition although they are situated at the median line instead of the lateral border of the neural plate. The writer sees one morphological objection to this modification of the above hypothesis; namely, that the rod and cone cells are bipolar ganglion cells and not merely sense cells. Another objection is that we have no evidence on the physiological side of close relation between the neuromast cells which are specialized for pressure sense and the rods and cones which are sensitive to light waves. Since we have experimental evidence of a close physiological relation between the general cutaneous ganglion cells and the rods and cones, the writer is inclined to accept the hypothesis in the more simple form stated obove.

The writer has read with great interest Boveri's effort (10) to connect the eye of vertebrates with the simple invertebrate eye through the pigment spot of Amphioxus. The hypothesis

242 Journal of Comparative Neurology and Psychology.

which he sets forth does not deal with the same field as that which I have dealt. The two hypotheses may prove to be consistent. At one point, however, they are absolutely opposed. BovERi bases his reasoning on the assumption that the eye, both in Amphioxus and in vertebrates, is a ventral structure. I have supposed, for reasons stated, that the vertebrate eye is a dorsal structure. If the eye of Amphioxus is necessarily to be regarded as a ventral structure, I doubt whether there is any relation between the eye of Amphioxus and that of vertebrates.

Morphology of the lens.

At this time when the lens is being made the subject of experimental investigation it would not be profitable to discuss whether it belongs to the series of epibranchial placodes, or to the dorso-lataral series, or is a structure of an entirely different character. Kupffer's hypothesis that it is a modified epibranchial placode fits very well with the absence of the gill slit and the visceral rami to which such a placode would belong. On the other hand, the invagination of the lens suggests relation to the lateral line organs and the auditory vesicle. This is strengthened by the fact noted above that in Gymnophiona the placode of the supraorbital lateral line nerve forms first as a vesicle like the auditory vesicle.

Morphology of the pineal organs.

In spite of the great number of investigations bearing upon these organs they still remain only imperfectly understood, owing probably to their very rudimentary character. In addition to their general structure whicli fits them to be light percipient organs, two facts point to a relation between these and the lateral eyes. These facts are the presence of rods or cones in Petromyzon (116) and the presence in the lateral walls of the diencephalic and mesencephalic neuromeres of pouches similar to the optic vesicles which may develop into the epiphyses (84, 125). If the epiphyses are formed from lateral pouches of the brain wall, as seems probable, it is not known whether a pair of such pouches fuse to form a median organ, or one of the

Johnston, Morphology of the Head. 243

pair disappears while the other moves nearer the middle, or whether one or the other process may occur in different forms. More important for the question of segmentation is the fact stated above that one epiphysis is situated behind the other. Since the anterior one has its stalk and center in the diencephalon, it may be assigned without hesitation to the diencephalic (iii neuromere and head segment 3. If the posterior one is one segment farther back it must fall into the anterior of the two mesencephalic neuromeres (iv) and head segment 4. This is consistent with the fact that its fibers pass (apparently) into the optic lobes. It must be noticed, however, that since the stalk of this second epiphysis stands between the superior and posterior commissures its assignment to neuromere iv implies that the anterior limit of the mesencephalon is not the posterior commissure, which has been taken as the most conv^enient mark. The limit must lie between the two epiphyses, some distance in front of the posterior commissure.

On the supposition that the epiphyses are homologous with the lateral eyes, and the evidence is all in favor of this, the writer would consider them to be metamorphosed cutaneous ganglia belonging to the third and fourth neuromeres and head segments. In the case of the epiphyses, their wholly dorsal position, the direct entrance of the fibers into dorsal centers without a chiasma, and the presence of the cutaneous center (tectum) in the segment of the second epiphysis all indicate that the cutaneous centers have remained in the brain. It seems probable that the centers for the first epiphysis have degenerated with the reduction of the organ. This view is suggested by the fact that in Petromyzon, where the organ is functional, the left ganglion habenulae divides into an anterior part which comes into connection with the epiphysial vesicle, and a posterior part which receives the tractus olfacto-habenularis through the superior commissure. The finding of these facts in Petromyzon changed the view which I had formed from the study of Acipenser, namely that the ganglion habenulae had changed its function after the reduction of the epiphysis. The structure and relations of the ganglia habenulae can be better

244 Journal of Comparative Neurology and Psychology.

explained on the assumption that they belong to the central grey.

/^. T/ie nefves of Piatt, Locy and Pinkus.

Miss Platt found in Acanthias (io6, 107) a process from the neural crest at the groove between the mesencephalon and diencephalon which later disappeared. She believed this to be the vestige of a nerve which was formerly developed, and gave it the name N. thalamicus. Its history has been followed more completely by Hoffmann (58, 59). The strand of neural crest cells descends over the optic stalk and meets the profundus strand at a point where later the ciliary ganglion is formed, and then loses its connection with the brain. Since the profundus ganglion is distinct from the ciliary and is formed from a different part of the neural crest, it seems altogether probable that the ciliary ganglion permanently represents the N. thalamicus. KuPFFER (79) finds the same structure forming the ciliary nerve in Ammocoetes. Hoffmann describes a small cutaneous branch from this ganglion but it is not clear that it is not composed of profundus fibers. In the case of the ganglion ciliare we seem to have the formation of a sympathetic ganglion from the ganglion of a nerve which has otherwise aborted. Since the sympathetic is derived from the visceral sensory and motor systems, we may consider the N. thalamicus as the vestige of a splanchnic sensory nerve root. In position it corresponds to the second epiphysis and lies one segment in front of the profundus. It would therefore belong to the first mesencephalic neuromere. The N. thalamicus may be considered as the splanchnic sensory nerve of neuromere iv and head segment 4.

Pinkus (104) described a nerve in Protopterus which arose from the brain in the vicinity of the praeoptic recess and, running forward, was distributed to a part of the olfactory epithelium, mingling with the olfactory nerve fibers. Locy (85, 86) has described the development and anatomical relations of a nerve which he has found in 19 genera of selachians. In most of the genera described the nerve enters the ventral surface of the fore-brain, in several it enters the dorsal surface near the anterior

Johnston, Morphology of the Head. 245

end. The nerve has the same distribution as the nerve of PiNKUS and is considered by Locy to be homologous with it. Locy's nerve possesses a distinct gangHon and ends centrally in a mesial eminence of the infolded pallium (126).

Until this nerve is better understood it is not worth while to attempt to explain the difference in its point of attachment to the brain in different species. It is important to notice, however, that its adult position indicates that it arises either lateral or dorsal to the neuropore in selachians. Its position is therefore morphologically dorsal and both the position of its root and its distribution to the extreme anterior end of the head indicate that it belongs to the first segment. Since it is distributed to the skin it may be supposed to be general cutaneous in character. Until further investigation, then, the hypothesis is suggested that the nerve described by Pinkus and LocY represents the primitive general cutaneous nerve of the first segment.

If it is true that this nerve ends in the pallium in selachians, the fact suggests the very interesting possibility that before the olfactory cortex was developed a general cutaneous center was located in its place. More than this can not be said until the central relations of the nerve are more fully studied.

75. Morphology of the olfactory organ. Functional and morphological relations to the gustatory system.

The resemblance of the olfactory sense cells to the typical sense cells of segmented worms may be taken as a starting point. Since the ectoderm of these worms is provided with free nerve endings (113, no, 82) comparable to the general cutaneous endings in vertebrates, the sense cells probably do not have a tactile function. It is more probable that they serve the function of taste or smell. Or perhaps it would be better to say that they have an indifferent chemical sense. These cells are especially numerous in the prostomium and roof of the mouth in the earth-worm. In Amphioxus (24) such cells are present as scattered cells in the lining of the buccal cavity, the velum and tentacles, the oral cirri, and the skin of a large part of the body; and in the form of cell groups or organs they are found in the

246 Journal of Comparative Neurology and Psychology.

cirri and velar tentacles. The region in which the olfactory epithelium is formed in vertebrates corresponds closely in position to Hatschek's pit in Amphioxus. This is based upon the description of v. Wijhe (120) and Legros (83). This pit lies in the roof of the invagination which corresponds to the hypophysis of vertebrates. It follows that Hatschek's pit of Amphioxus and the olfactory organ of vertebrates lie in the roof of the ancestral mouth and in this position correspond to the prostomial area of invertebrates in which sense organs are especially numerous. The hypothesis is suggested, therefore, that the olfactory organ represents a collection of sense organs of the type found in segmented worms which at first was related to the vertebrate paleostoroa as the special collection of sense organs is related to the annelid mouth. These organs probably had an undifferentiate chemical sensibility and served the worm and the primitive vertebrate in seeking food. It is not necessary to this hypothesis that Hatschek's pit should prove to be a sense organ in Amphioxus. The reference to it is made only for the sake of indicating as clearly as possible the point at which the sense cells are supposed to have collected.

The question of the segmental value of the olfactory nerve was discussed in a previous paper with the result of denying it primary segmental value. This was based on the description of its nerve centers and the specific conclusion reached was that the olfactory nerve and its centers cannot serve to define the segments in which they lie because they have no resemblance to the typical nerves. I see no way to minimize that argument but it may be said nevertheless that we have in the first segment of the head in addition to a general cutaneous nerve a nerve whose functions are visceral sensory. And now I may go further and suggest how the very differences between the olfactory apparatus and the other cranial nerves may be taken to indicate the morphological status of the region of the head in which it is. If the comparison of the olfactory epithelium with the special collection of sense cells in the prostomium of annelids is accepted, it implies that the olfactory segment is equivalent to the invertebrate prostomium. So far, then, the writer is in

Johnston, Morphology of the Head. 247

clined to think that the olfactory organ and nerve are peculiar in the vertebrate organization. In the figures, however, these structures have been indicated by the same symbols as the visceral sensory system, on the basis of the visceral function of the olfactory apparatus.

On the basis of their function the writer has for several years felt impelled to seek some morphological connection or relation between the olfactory apparatus and the visceral sensory, especially the gustatory. After what has just been said this search would seem contradictory, if not fruitless. However, we have not yet sufficient knowledge of these systems to warrant our jumping to any final conclusion. Even if the first segment of the vertebrate body is a prostomium, its general cutaneous innervation is homologous with that of tiie rest of the body. May not its visceral sense be related to that of the typical segments also? The writer has in his notes detailed statements of several hypothesis, but it will be more profitable in the present state of our knowledge to indicate briefly one or two lines along which facts may be sought, than to spend time in speculations.

(i). When the gustatory paths and centers within the brain are known at least as well as we know the olfactory we may find some indications of their relationship.

(2). If we were to apply to the epibranchial placodes the theory of Kupffer which the writer has declined to accept for the dorso-lateral placodes, some such line of reasoning as this might be developed. The cells of the chemical sense which are scattered over the body in annelids may have collected over the gill slits in primitive vertebrates. These collections would form the primary sense organs of Kupffer. These sense cells sink below the surface as ganglion cells. Secondary sense organs develop in the placodes and these become the taste buds. If we suppose that the primarj^ sense cells sent their fibers to thebrain along with the general visceral rami which pass adjacent to the gill slits, we have a close relation shown between the gustatory and olfactory systems. It should be noticed, however, that in this statement as it stands no homology is shown be

248 Journal of Comparative Neurology and Psychology.

tween the centers of the two systems, since it is by accident that the taste fibers enters the visceral centers.

(3). If the dorso-lateral placodes represent a part of the neural plate area of ancestral vertebrates, may not the epibranchial placodes be an offshoot from that also ? This suggestion would have a point if it were shown that the end buds had their origin from the epibranchial placodes as the neuromasts have their origin from the dorso-lateral placodes. It would be necessary to explain further the wide separation between the epibranchial placodes and the neural plate, and the permanent relation of the nerves of these placodes, not to the margin of the neural plate but to the column which lies mesial (ventral) to the cutaneous area of the plate.

(4). The following suggestion seems to the writer most in accord with such facts as we have at present. The end buds are recent organs, formed comparatively late in the phylogen}^ They appeared first in the vicinity of the mouth and gill slits where they would detect indications of food in the respiratory water current. In this position relative to the vertebrate neostoma they would supplement the impressions received from the olfactory organ. The animals in which end buds first appeared may be thought of as relatively inactive, so that the respiratory current formed their chief source of information as to the presence of food. They were distinguished from such animals as Amphioxus which feed by means of cilia by the fact that they captured larger kinds of food by muscular movements. This would be accomplished at first by means of the visceral muscles of the jaws and branchial skeleton alone, the food perhaps being sucked in by a sudden increase of the respiratory current. The innervation of the end buds was determined by their function. It is evident that they could not be innervated by the general cutaneous component. For, if cutaneous fibers could receive and transmit impulses due to chemical stimuli those impulses would be interpreted in the brain as tactile impulses ; that is, they would give rise to reflex movements which characteristically follow upon tactile impulses. In the present case this would be useless or harmful to the animal. On the

Johnston, Morphology of the Head. 249

other hand impulses transmitted by the general visceral sensory fibers would be followed by reflex movements of the visceral musculature. Since the visceral fibers passed close to the supposed site of the first end buds, their innervation by these fibers is easy to understand. As the end buds became more numerous and important they spread into the mouth and over the head and body and the visceral nerves followed them. While it is not possible to explain in detail the arrangement of these nerves, the disposition of such nerves as the ramus recurrens facialis in teleosts is in support of this view. If any relation is to be found between the gustatory and olfactory systems it must rest upon the central relations of the two systems. If the gustatory centers and fiber tracts are so related to the olfactory as to show that the primary olfactory centers belong to the same longitudinal zone of the brain as the visceral centers, the close physiological relations between the olfactory and gustatory organs have their explanation in their morphological relations. The writer expects that this will prove to be the case and in another section offers some slight suggestions looking in this direction.

16. The first fo7ir head segments.

These may be passed over briefly since all of the structures concerned have been treated in the above general discussion.

a. Segment 4.

Neuromere iv, anterior head cavity, palatine arch (?), epiphysis 2, N. thalamicus (ciliary ganglion and nerve), lens placode (?).

The structures peculiar to this segment are all rudimentary. Only the brain is well developed and that chiefly because of its relation to the cutaneous and optic tracts. The dorsal mesoderm is represented by the anterior head cavity in selachians, the lateral mesoderm only by the palatine arches of Kupffer in Petromyzon, if they actually exist. The splanchnic sensory nerve is represented by the rudimentary N. thalamicus from which the ciliary ganglion and nerve are derived; the somatic sensory nerve by the second epiphysis. If the lens is developed from an epibranchial placode it should be placed here in con

250 Journal of Coniparative Neurology and Psychology.

nection with the ciUary nerve (79) as the epibranchial placode of the fourth segment.

b. Segment 3.

Neuromere jii, most anterior portion of the median entoderm (?), epiphysis i.

Neuromere iii represents the thalamus but according to Hill's results (52) does not include the hypothalamus which is developed in teleosts as a ventro-caudal expansion of the lower wall of neuromere ii. No mesoderm and no motor nerves are present but within the brain is the thalamic nucleus of origin of the somatic motor fasciculus. Its location here possibly indicates the former presence of a motor nerve in this segment; no other explanation occurs to the writer. The representative of the general cutaneous nerve is found in the anterior epiphysis. The innervation of the skin is taken over by the profundus. As suggested above, the central nuclei belonging to the cutaneous division have probably become reduced to an insignificant remnant on account of the degeneration of the epiphysis. The ganglia habenulae in their structure resemble the central grey and they may be regarded as one of the earliest special centers to be differentiated from it. The suggestion has been made earlier (69) that the central grey belongs to the category of the commissural and tract neurones which are found in the oblongata and cord. The reason for the large development of these central elements in the diencephalon is to be found in the reduction of the somatic and visceral sensory and visceral motor centers. The posterior commissure, whose nucleus of origin lies in this neuromere, is spoken of below (Sec. 17).

c. Segment 2. Neuromere ii, lateral eyes, hypophysis.

The interpretation of the eyes as the representative of the cutaneous division has been given above. In the central nervous system important centers are developed. Neuromere ii includes the striatum area and also that of the pallium. The writer has previously referred to the striatum as probably representing- the tract cells of the cord and oblongata. The nature

Johnston, Morphology of the Head. 251

of the specialized nucleus known as the epistriatum is a difficult question. It serves in fishes as the end nucleus of two very different classes of fibers. The one comes from the inferior lobes and corpora mammillaria by way of the anterior commissure; the other comes from the cells of the olfactory bulb through a somewhat isolated portion of the anterior commissure. It is probable that this is the more primitive tract and that it is gradually supplanted by tracts from the h)^pothalamus and perhaps other regions. In any case the epistriatum seems to belong chiefly to the visceral sensory apparatus.

The origin and relationships of the cerebral pallium have been discussed by the writer elsewhere (70). The massive pallium which is related to the olfactory apparatus is regarded as a neomorph in gnathostomes and as such it is not of especial importance for our present subject. It has been suggested above, however, that this olfactory pallium may have been preceded by a structure which was the general cutaneous nucleus of the first segment. The pallium, then, is to be regarded as a new structure developed in the space from which the lateral eyes were derived.

If Hill's interpretation of the neuromeric relations is cor_ rect, we are brought to the interesting result that the diencephalon is formed from two neuromeres, the hypothalamus belonging to the same neuromere with the striatum and pallium and lying morphologically ventral to them. The writer has shown that in fishes (Acipenser, 6]) throughout the whole length of the ventral wall of this second neuromere the structures adjacent to the middle line (nucleus thaeniae, inferior lobes in part, and corpus mammillare) are histologically identical. In Petromyzon (68) the inferior lobes and corpus mammillare are fundamentally alike, there being but slight differentation. It seems, then, that a single nucleus has been interrupted by the optic chiasma and accompanying decussations. The posterior portion of this nucleus was associated with the saccus vasculosus and also came to receive tertiary tracts from both the olfactory and the somatic sensory centers. It has consequently developed into a large coordinating center. The anterior portion of the nucleus (nuc.

252 Journal of Comparative Neurology and Psychology.

thaeniae) received secondary olfactory tracts and remained at a lower stage of development.

The absence of epibranchial placodes in segments 2 and 3 is noteworthy. If the hypophysis is in segment 2 no placode is to be expected. The absence of one in segment 3 may argue against the supposition that a gill was ever present in that segment. The lense placode and the N. thalamicus would then appear to be the vestiges of the nervous structures connected with the most anterior ancestral gill slit.

Whether the hypophysis is to be assigned to this segment or to segment i involves the question whether we are to recognize a prostomium in vertebrates. The hypophysis invaginates about opposite the groove between neuromeres i and ii and extends back beneath neuromere ii. This canal is perhaps equivalent to the buccal cavity of invertebrates. The treatment of the olfactory organ in this paper and the interpretation of the hypophysis as the invertebrate mouth are both consistent with the recognition of the first segmeut as a prostomium.

d. Segment i. Prostomium.

Neuromere i, nerve of Locy, olfactory epithelum and nerve.

The writer is unable to accept Kupffer's idea of a median olfactory organ. It does not appear from his descriptions that the median thickening between the olfactory placodes is ever sensory. Its connection with the brain for some time is to be explained as a mere persistence of the continuity of the neural tube with the ectoderm at the lower edge, of the neuropore. The olfactory sense cells exist from the first in bilateral groups and are innervated by paired nerves. As may be inferred from references to Amphioxus made above, the writer thinks that the interpretation of the olfactory pit as a true olfactory organ requires further evidence.

The olfactory bulbs are formed by great growth of the lateral walls of the first neuromere. In lower fishes {6j, i 6) and cyclostomes (68) these bulbs still show a structure comparable to that of the primitive brain wall. The highly speciali zed mitral cells are not primitive features of the olfactory apparatus,

Johnston, Morphology of the Head. 253

but are neomorphs in craniates which are just beginning their specialization in cyclostomes.

The general cutaneous nerve of Locy and the olfactory nerve are to be taken together as roughly the equivalent of the nerve of the prostomium in segmented worms.

77. The dorsal commissures of the bi'ain.

In Figs. 1—7 are shown the longitudinal zones of the brain belonging to the four functional divisions of the nervous system. A further review of these zones does not fall within the scope of the present paper. Some remarks upon the dorsal commissures and their relations to the zones, however, will perhaps be of value. It should be said at the outset that the fiber crossings in the brain of lower vertebrates, both dorsal and ventral, are in the greatest part if not wholly mere decussations. True commissures are almost unknown. When the word "commissure" is used in the following paragraphs it is used because it has come to be part of the accepted name of a given fiber-crossing.

The dorsal decussation of the spinal cord in higher vertebrates contain splanchnic sensory (sympathetic) fibers, collaterals from cutaneous fibers, and perhaps secondary elements. At the junction of the spinal cord and brain in all vertebrates this decussation is greatly enlarged just behind the choroid plexus of the IV ventricle. This enlarged portion, known as the commissura infima Halleri, is composed chiefly of splanchnic sensory fibers from the roots of the VII, IX and X nerves. These fibers for the most part end in the median nucleus of the commissure of Cajal (46, 6'], 68). A part of them without crossing (and a part of the crossed fibers also?) pass on back into the spinal cord (68). Other fibers in the commissura infima come from the cells of the nucleus funiculi (90, 46). From these facts it is evident that both splanchnic sensory and somatic sensory elements cross in the dorsal decussation of the cord. The somatic sensory elements are chiefly or wholly secondary fibers, or only collaterals.

These two components must be rigidly distinguished if we are to understand the dorsal decussations of the brain. The

2 54 Journal of Comparative Neurology and Psychology.

dorsal decussation of the medulla oblongata is not obliterated on account of the non-nervous roof, but its elements are crowded forward or backward. Behind tiie choroid plexus the commissura infima contains the splanchnic sensory elements proper to the segments of the VII, IX and X nerves. It is probable that the course of the root fibers of these nerves within the brain has been influenced by the crowding backward of their decussation and median nucleus by the choroid plexus. It is further probable that those fibers which take this caudal course are the more primitive components of these nerves, namely the general visceral fibers as distinguished from end bud fibers. It is known that the end bud fibers in teleosts (48) end chiefly in the lobus facialis. The point of especial interest in the present connection is that the concentration of the visceral decussation for the VII, IX and X nerves behind the choroid plexus precludes the expectation that the visceral elements of the first order will be found in the dorsal decussations farther forward. There are no visceral nerves anterior to N. VII.

The somatic sensory elements have behaved differently with reference to the IV ventricle. Instead of concentrating behind it they have concentrated in front of it. In those vertebrates in which the cerebellum is most primitive (Petromyzon, Protopterus, Urodeles) a commissure constitutes a prominent part" of it. This decussation consists of axones of granule cells situated in the cerebellum destined to the somatic sensory nuclei of the medulla oblongata. This is therefore to be considered as the homologue of the somatic sensory portion of the dorsal decussation of the spinal cord. It is an important decussation in all lower vertebrates.

A second prominent cerebellar decussation is found in fishes (67). This is situated farther ventrally and cephalad and instead of connecting the dorsal portions or lateral lobes of the cerebellum, connects two nuclei which in fishes lie in the lateral walls distinctly ventral to the somatic sensory centers, the secondary vagus nuclei. The fibers of the secondary vagus tract coming from the vagus lobe end in part in the secondary vagus nucleus of the same side and in part cross

Johnston, Morphology of the Head. 255

to the opposite side. The remainder of the decussation is formed by axones of cells of these nuclei. The destination of these axones is unknown, so that we are uncertain whether true commissural elements are present. It is evident, however, that the decussation belongs to the splanchnic sensory division of the nervous system.

In mammals the tract which corresponds to the secondary vagus tract is the direct cerebellar tract arising from the cells of Clarke's column in the cord. This tract ends in a nucleus which lies in the roof of the IV ventricle in the vermis. It is, however, probably homologous with the secondary vagus nucleus of lower vertebrates.

The dorsal decussation of the optic lobes requires but brief mention. In accordance with the interpretation of the lobes given in this and previous papers it must be regarded as a somatic sensory decussation comparable wifh the somatic portion of the decussation in the spinal cord.

The posterior commissure is a decussation of fibers arising from a nucleus in the dorsal part of the diencephalon and mesencephalon (75. 68). After crossing the fibers run toward the base of the medulla oblongata. The destination of the fibers is unknown, and nothing is known of the fiber tracts which may come to end in the nucleus of origin. The fact that the decussation is somewhat intermingled with the dorsal decussation of the optic lobes suggests that the two may be related. Some of the cells of origin of the posterior commissure in Petromyzon are so situated as to be indistinguishable by their position from those of the optic lobes. Without knowing the connections and functions of the posterior commissure, nothing further can be said.

In the roof of the diencephalon two decussations are present, the well known superior or habenular commissure and a decussation closely related to the base of the epiphysis known only in a few forms (67, 62). The superior commissure contains decussating fibers from the olfactory nuclei of the fore-brain (tractus olfacto-habenularis) and also, according to Catois (16) and Cameron (15) true commissural fibers. A certain similarity

256 Journal of Comparative Neurology and Psychology.

is to be noticed between this and the decussation between the secondary vagus nuclei in the cerebellum. Decussating fibers of the second and third order are present in each case. A further point of comparison is found in the fact that the ganglia habenulae appear to be a part of the central grey some distance removed from the dorsal wall of the brain, which here is a choroid plexus. If we accept the supposition that the somatic sensory nucleus of the diencephalon has degenerated, the morphological position of the ganglia habenulae corresponds rather closely to that of the secondary vagus nucleus beneath the somatic sensory centers of the cerebellum. The choroid plexus of the diencephalon is regarded as morphologically equivalent to the optic lobes or dorsal part of the cerebellum and the ganglia habenulae are morphologically lateral. These hypotheses have a significance in connection with the question of the relation of the olfactory and gustatory organs and centers discussed above.

Not enough is known about the epiphysial decussation to admit of any discussion.

The anterior commissure consists of two distinct parts: a smaller, anterior portion, sometimes isolated, composed of fibers from cells of the olfactory bulb destined to the epistriatum ; and a larger portion composed of fibers from the inferior lobes also destined to the epistriatum. A third element, consisting of fibers from the lateral cortex to the opposite epistriatum, need not be considered since it is not primitive in vertebrates. The tract from the inferior lobes is a tertiary tract which may be either olfactory or somatic sensory in character. We have seen reasons for thinking that the hypothalamus has been modified by the entrance into it of the somatic sensory tracts. We must therefore regard the secondary (and tertiary?) olfactory tracts as constituting the primitive portion of the anterior commissure. The writer is unable to see any resemblance between this and the somatic sensory decussation, while there is a certain similarity between the anterior commissure and the splanchnic sensory decussations of the ganglia habenulae and the secondary vagus nuclei. If the anterior commissure is to be compared

Johnston, MorpJiology of the Head. 257

with the other brain commissures it is doubtless to be considered as a part of the dorsal splanchnic decussations.

We may summarize the dorsal commissures by saying that the dorsal decussation of the cord has been differentiated in the brain into separate somatic and splanchnic sensory decussations. The commissura infima Halleri, the decussation between the secondary vagus nuclei, the superior commissure, and the anterior commissure represent the splanchnic portion. The dorsal commissure of the cerebellum, the dorsal decussation and posterior commissure of the mesencephalon constitute the somatic portion. Whether the pallial commissure will be brought into the latter category on account of the relations of Locv's nerve remains to be seen.

Some reference was made to the ventral commissures in discussing the morphology of the eye. The writer has nothing further to add.

18. The sympathetic sysieitt.

The relation of the sympathetic to the four functional divisions of the nervous system is important in connection with the subjects dealt with in the present paper. Although the sympathetic in lower vertebrates, where we must look for its genetic relations, is very imperfectly understood, yet sufficient is known to indicate that the sympathetic is morphologically related to the visceral divisions. This is what the function of the sympathetic and the central relations of its fibers in higher forms has led us to expect. In the trunk regions in selachians (60) the sympathetic ganglia are budded off from the distal portion of the spinal ganglia after the union of the ventral with the dorsal root. It does not appear that anything is contributed to the sympathetic by the ventral root in fishes. In the adult (64, 98) sensory fibers coming from the viscera pass through the sympathetic and spinal ganglia to end in the region of Clarke's column in the cord, and motor fibers arising in the lateral motor nucleus enter the sympathetic by way of the dorsal root (and in higher forms by way of the ventral root also). If we interpret

258 Journal of Comparative Neurology and Psychology.

the formation of the sympathetic gangHa in the Hght of the adult structure, we must suppose that the spinal ganglion cells which migrate to the sympathetic ganglia are those of the general visceral component.

It is a striking fact which has not passed unnoticed that the composition of the sympathetic closely resembles that of a branchial nerve minus the cutaneous components of the latter. In cyclostomes and selachians no sympathetic gangHa have been found in the head, except the ganglion ciliare. If this be the true condition then in these forms the cranial ganglia must be considered to contain the equivalent of the sympathetic. That they can not as a whole be considered homologous with the sympathetic ganglia of the trunk is obvious from the presence of taste fibers and general cutaneous fibers. In teleosts (46) sympathetic ganglia are present in connection with the typical ganglia of the profundus, V, VII, IX and X nerves, in addition to the ciliary ganglion which is the sympathetic ganglion of the segment anterior to the profundus. The ganglia are connected by a longitudinal fiber strand as in the trunk. The high development of the sympathetic in the head of teleosts suggests that it may yet be found in selachians.

/p. Relation of dorsal and ventral nerve roots to the myotomes.

In Amphioxus the ventral root emerges from the spinal cord opposite a myotome and enters the muscle directly. The dorsal root emerges from the cord between two myotomes, passes along the myoseptum to the surface and is distributed to the ectoderm overlying the myotome next anterior to its root. I can add from my own observation that although the ventral ramus (the visceral fibers of which do not concern us here) usually passes vertically ventrad and so crosses the surface of three myotomes, it not infrequently happens, especially in the caudal half of the body, that the ventral ramus runs caudoventrally at the same angle as the myotomes and a single nerve is confined to the area overlying a single myotome.

In Ammocoetes by putting together the work of Hatschek (44) and Koi.TZOFF we can account for every one of the dorsal

Johnston, Morphology of the Head. 259

and ventral roots of the cranial and first spinal nerves and can state their origin, relation to the myotomes and distribution. Reference to Fig, 10, with which Hatschek's Fig. 11 and Koltzoff's diagrams may be compared, will make the relations clear. The profundus passes over the caudal and ectal surface of somite i, which is innervated by N. Ill, and is distributed to the ectoderm outside and in front of that somite. The trigeminus passes over the caudal and ectal surface of somite 2, innervated by N. IV, and is distributed to the ectoderm overlying that somite and its mandibular arch. N. VII passes caudal to somite 3, innervated by N. VI, and supplies the hyoid arch. It has no cutaneous component. N. IX lies between somites 4 and 5 and has a dorsal ramus passing out along this septum (Hatschek's Oec. /). The vagus ganglion appears between somites 5 and 6 and its dorsal ramus passes out along this septum. There is no ventral root to either somite 4 or 5 on account of the early reduction of the mesial part of each of these somites. Behind somite 6 (Koltzoff) both dorsal and ventral roots are present. The ventral root of somite 6 probably innervates myotomes 4, 5 and 6 in the larva. The dorsal root sends a ramus to the skin of the occipital region {^Occ. j, Hatschek). We should expect the dorsal rami of N. IX, X and Sp. i to be distributed to the ectoderm overlying somites 4, 5 and 6 respectively. It is probable that this is their actual position in the embryo but owing to the migration forward of myotomes 4, 5, 6 and 7 during development myotome 5 is brought beneath the ectoderm belonging to somite 4, myotome 6 beneath the ectoderm of somite 5, etc. This accounts for the apparent distribution of Hatschek's Occ. i, 2 and j to the area over myotomes 5, 6 and 7. These dorsal nerves all persist in the adult but the ventral nerve of somite 6 disappears and myotomes 4, 5, 6 and 7 are all innervated by the ventral nerve of myotome 7. From the work of these two authors, then, it would appear that the relations in cyclostomes are the same as in Amphioxus, the dorsal nerves innervating the ectoderm overlying the myotomes next anterior to the dorsal roots. Two further remarks are to be made. Neal's suggestion that

26o Journal of Comparative Neurology and Psychology.

N. VI includes the ventral nerve of somite 4 (and that of somite 5 ?) does not affect the the present question. It is an important question, however, whether Hatschek's Occ. i, 2 and 7 are all general cutaneous nerves. If they are lateral line nerves without general cutaneous fibers they are worthless for the present purpose. The writer has described a branch from the N. lateralis to this part of the skin. The cranial nerves of Petromyzon require to be reinvestigated, but so far as our present knowledge goes the above account of the relation of the dorsal roots to the myotomes appears to the writer to be perfectly sound.

The same relation of the nerves to the somites is found in Spinax (12) where N. IX Hes behind somite 4, N. X behind somite 5, and a rudimentary ganglion is formed behind each of the following somites. Six of these rudimentary dorsal nerves eventually disappear owing to the reduction and shifting of myotomes and consequent reduction of cutaneous area.

Hatschek in 1892 came to the same conclusion as I have stated with regard to the relation of dorsal roots to myotomes in Amphioxus and cyclostomes. In 1893 he published a brief correction (45) in which he said that in all higher vertebrates the dorsal root unites with the next following ventral root. This can be made out most readily, he says, in amphibian larvae. I have examined such sections of amphibian embryos as I have by me and although I have none old enough to show the distribution of the nerve rami, I find two facts opposed to Hatschek's statement. The dorsal roots are always a short distance caudal to the ventral roots, and the ventral roots run across the cephalic face of the ganglia. I do not see how this could result if each dorsal root joined the next following ventral root. I hope soon to investigate the matter farther in Amblystoma.

The second statement of Hatschek seems to have been generally accepted by later workers. I am inclined to think that his first statement was correct for the fishes at least. 20. Comparison of head and Mink.

Aside from the development of special organs in the head, the chief differences between the head and trunk lie in the skele

Johnston, Morphology of the Head. 261

ton and somatic muscles and in the relation of the dorsal and ventral nerve roots to one another and to the myotomes. The formation of a rigid cranium has had very important consequnces in the disappearance of postauditory myotomes and the development of complicated conditions in the occipital region. To this also is due the fact that certain myotomes have been released from their attachment to a segmented skeleton and have come to move the eyeball. It may be said that the same cause has indirectly determin-ed the relation of the nerve roots to the myotomes. In cyclostomes, where the buccal apparatus has demanded the preservation of the dorso-lateral portion of the postauditory myotomes, the nerve roots are covered by these very much as in the trunk. In gnathostomes, however, when postauditory myotomes completely disappear the nerves are given free access to the ectoderm. When trunk myotomes shift forward it is their ventro-mesial portions which become attached to the cranium and so the nerves retain their superficial position. The union of the dorsal and ventral roots in the trunk is a late development and the head has retained the more primitive condition. In the constitution of the dorsal and ventral roots also the head has retained the primitive condition, while the trunk nerves have been somewhat modified at least in higher vertebrates. The dorsal roots originally contained the somatic sensory, splanchnic sensory and splanchnic motor components, as in Araphioxus. In the trunk a part of the splanchnic motor component has come to run in the ventral root and perhaps other modifications have taken place.


1. The primitive vertebrate was a segmented animal with probably very slight cephalization.

2. Each segment consisted of derivatives of the ectoderm, dorsal mesoderm, lateral mesoderm and entoderm, and of somatic sensory, somatic motor, splanchnic motor, and splanchnic sensory divisions of the nervous system primarily related to the skin, the myotomes, the visceral muscles, and the visceral surfaces respectively. The definite relations existing between the

262 Journal of Comparative Neurology and Psychology.

functional divisions of the nervous system and the organs severally innervated by them make the nervous system the best guide to the segmentation of the head.

3. The anterior end of the head is indicated in existing vertebrates b)' the point at which the neural plate meets the general ectoderm and the entoderm, in those forms in which the rostral entoderm is most completely preserved. Lateral to this the olfactory epithelium is formed. Ventral to it the hypophysis represents the vertebrate paleostoma or invertebrate mouth. That part of the head and brain which extends over the hypophysial opening may be compared to the invertebrate prostomium and as such contains only cutaneous and olfactory nerves.

4. It is probable that one or two pairs of gills formerly existed between the paleostoma and the present mouth. This is indicated by the presence of epibranchial placodes and palatine and trabecular arches in Petromyzon, by the mode of formation of the ciliary ganglion, and by the labial cartilages of selachians. Similarly one or two formerly functional myotomes have been lost from the anterior end of the head. The anterior head cavity of selachians is the rudiment of one of these. Another more anterior, is possibly indicated by the somatic motor nucleus far forward in the thalamus. The loss of these several organs has been followed by the loss of the typical nerves or their modification into special sense organs. The nerve of Locy and the ciliary ganglion are the only nerves in the first four segments which retain their primitive relations.

5. Following this region of greatest modification, it is believed that the segments can be reduced to common terms and the segmental position of all the organs determined. See Table B and Figure 10.

a. The gill slits were originally intersegmental.

b. Each typical dorsal root contained a general cutaneous component distributed to the skin ; a communis component distributed by way of the posttrematic, pretrematic, and pharyngeal rami to the mucosa; and a visceral motor component to the visceral musculature by way of the posttrematic ramus. The

Head segments







7 8




12 13

















xli xiii








a. h. c.















mandib. arch

hyoid arch

br. I

br. 2

br. 3

br. 4

br. 5

br. 6

a larger number of branchial arches in lower cyclostomes and primitive vertebrates.

Somatic sens, division general culaneoiis l)eriplicral central

N. of I.ocy



in N. V. N. ophth. shifts [joins prof. to vil N. V ?]

(tectum), cerebellum, acu

comp't inN. IX shifts to Ix

sticiim, ge


in N. X. 1 sp. Pet. shifts to joins X in N. X, Gnath. Gnath.

1. cutaneous nucleus.

2 sp. Pet. rudiment in Gnath.

sp. V tract

3 sp. Pel. rudiment in Gnath.

, nuc. funi

4 sp. Pet. rudiment in Gnath.


5 sp. Pet. rudiment in Gnath.;

jjermancnl root In .\canthias?

6 sp. Pet. permanent root in .\canthias

7 sp. Pet. permanent root in Gnath.

8 sp. Pet.

2 sp. ? Gnalh.

special cutaneous peripheral central

acustico-lateral system of organs and nerves.

cerebellum, acusticum, spinal VIII tract to the nucleus acustici spinalis.

special sense

later, eyes

■ St.


2nd. epiph.


Somatic motor division peripheral central

nuc. s. m. f.

s. m. f.

N. Ill nuc. Ill

N. IV nuc. IV

N. VI nuc. VI

\s KUrb.J 1 [ / ] s. m. f. 1 s. m. f.



s. m. f.

s. m. f.

w s. m. f.


s. m. f.


s. m. f.

s. m. f.

I sj). Fiirb. ventral mot. col.

2 sp.

ventral mot. col.

Si'l.ANCHNIC sens. DIV.

general visceral



N. thai. N. cil.

comp't in N.VII shifts to viil

comp't inN. IX shifts to ix

lob. fac.

comp't inN. X; shifts toxin Gnath.

lob. vagi

comp't col'ct'd into N. X

lob. vagi

comp't col'ct'd into N. X

lob. vagi

comp't col'ct'd into N. X

lob. vagi

comp't col'ct'd into N. X


comp't in r. br.-inl. X

vagi and nucleus't br. int. X jommlssuralis

comp't br.-int. X of Cajal.

omp't br.-int.X

special visceral

The gustatory components and their centers have not been fully distinguished from the general visceral in any vertebrate.

special sense

N. olf.

1 1

Splanchnic motor div. peripheral


mot. N.V



lat. mot. nuc. V

mot. N. VII shifts to viii

lat. mot. nuc. V

mot. N IX shifts to ix

lat. mot. nuc. VII

mot. N. mot. N. mot. N. mot. N. mot. N. X shifts col'ct'd col'ct'd col'ct'd col'ct'd toxGn. intoN.X into N.X into N.X into N.X N. access.

lat. mot. , . , , nuc. IX and X ; continued as the lateral motor column of spinal cord

N. access.

N. access.

N. access.

Journal of Comparative Neurology and Psychology. Vol. XV To face p.'.'i62.

Johnston, Morphology of the Head. 263

nerve root passed behind (caudal to) the myotome of the segment to which it belongs.

c. Various shiftings of organs have taken place, due to (i) the expansion of the branchial apparatus and (2) the formation of a rigid cranium with the consequent loss of dorsal musculature. The shifting which may be observed during the ontogeny is a valuable source of information which should be further studied.

d. Largely on account of the peculiar relations with the ear, the nerve roots connected with the medulla oblongata have shifted backward during the phylogeny, and have left two neuromeres without nerve roots.

e. As a result of the expansion of the branchial apparatus the visceral sensory and motor components of the nerves caudal to the glossopharyngeus have been collected into a common root, the vagus. The somatic sensory and motor components of more or fewer of these nerves are present as rudimentary or complete dorsal and ventral roots which occupy the position of the original nerves. The number of these nerves retained depends upon the reduction of myotomes and cutaneous area in the postauditory region. In cyclostomes all the general cutaneous roots are present and independent; two of the somatic motor roots have disappeared. In gnathostomes, owing to the shifting backward of the vagus root at its attachment to the brain, the general cutaneous root following it has been absorbed and appears in the adult as the second r. auricularis vagi.

f. There has been no shifting forward of spinal nerves into the oblongata.

6. The typical sense organs of vertebrates fall into two categories: the neuromast organs constituting the acustico-lateral system, and the end buds or taste buds. These two sets of organs are absolutely distinct in structure, function, innervation, and central nerve relations.

7. It has previously been shown that the acustico-lateral system was derived from the general cutaneous. It is thought that the neuromast organs appeared first in a marginal portion

264 Journal of Comparative Neurology and Psychology.

of the primitive neural plate. The material used in the formation of these organs was a portion of that part of the neural crest which gave rise to the general cutaneous component. The neuromast area was largest in the region of the VII nerve and neuromere vii, but extended forward probably one or two segments. It can not be exactly compared with the dorsal rami. It is thought that this area, together with the cutaneous ganglion cells, remained outside the neural tube and that later the ganglion cells migrated into the space between the ectoderm and neural tube. This is repeated in the ontogeny in the phenomena of the dorso-lateral placodes and the formation of ganglia from them.

8. The visual organs are thought to have been derived from the neural crest area also, in segments 2, 3 and 4. The retina of the lateral eye is supposed to contain the equivalent of a general cutaneous ganglion and its primary brain center. The optic tract is homologous with the internal arcuate iibers which arise from the cutaneous centers, decussate in the base of the brain and run to the tectum opticum. The epiphyses are thought of as modified general cutaneous ganglia whose centers have remained in the brain.

g. The olfactory organ is regarded as a special collection of sense cells of the invertebrate type, sensitive to chemical stimuli, which are gathered above the hypophysial opening as the organs of the same type in the invertebrate are gathered in the roof of the mouth and on the prostomium.

10. On account of the similarity of their function it is expected that some morphological relation will be found between the olfactory and gustatory organs. The source of the end buds, the origin and history of the nerve components which innervate them, and the course of the gustatory paths in the brain constitute one of the most important problems in vertebrate morphology.

11. It is important that the origin of each of the sensory components in the cranial nerves should be fully worked out in at least a few forms. The cells which enter the ganglia from the neural crest and from the placodes should be traced con

Johnston, Morphology of the Head. 265

tinuously until their fate is determined. The cells derived from the neural crest should be distinguished into general cutaneous, general visceral, or others as the result might be.

Naples, October 2j, igo4.


For numerous other papers see the literature lists in Nos. 35 and 70.


1. Untersuchungen iiber das Gehirn der Petromyzonten. Zeit. f. zuiss.

ZooL, Bd. 39, 191-294. 1883.

2. Ueber den Ursprung und Austritt der Hirnnerven von Petromyzon.

Zeit. f. wiss. ZooL, Bd. 40, 286-308. 1884. AIMS.

3. The Cranial Muscles and Cranial and First Spinal Nerves of Amia

calva. Jour. Morph., Vol. 12, 487-808. 1897.

4. The Lateral Sensory Canals, the Eye-Muscles, and the Peripheral

Distribution of certain of the Cranial Nerves of Mustelus laevis. Quart, Jour. Micr. Set., Vol. 45 N. S., 87-236. 1901.

5. On Certain Features of the Lateral Canals and Cranial Bones of

Polyodon folium. Zool. Jahrb., Abth. f. Anat. u. Ontog., Bd. 17, No. 4. 1903. Beckwith.

6. The Early History of the Lateral Line and Auditory Anlages of

Amia. Abstract. Scietice, N. S., Vol. 15, 575. 1902. Bigelow.

7. The Sense of Hearing in the Goldfish, Carassius auratus L. Amer.

Nat. Vol. 38, 275-284. 1904. Bochenek.

8. Nowe szcze goly do budowy przysadki mozgowej plazow. (Neue

Beitrage zum Bau der Hypophysis cerebri bei Amphibien). Bull. Internat. Akad. Krakozv, 1902, 397-403. Boeke.

9. Die Bedeutung des Infundibulums in der Entwickelung der Knochen fische. Anat. Anz., Bd. 20, 17-20. 1901. Boveri.

10. Ueber die phylogenetische Bedeutung der Sehorgan des Amphioxus.

Zool. Jahrb., Suppl. 7 (Festsch. Weismann), 409-428. 1904. Brauer.

11. Beitrage zur Kenntniss der Entwickelung und Anatomic der Gym nophionen. Zw/. y^.^^^., Suppl. 7, 381-408. 1904.

266 Journal of Comparative Neurology and Psychology.


12. Beitrage zur Entwickelung der MuscuJatur und des peripherischen

Nervensysteras der Selachier. Morph. Jahrb., Bd. 27, 415-496; 501-629. 1899. Burckhardt.

13. Die Einheit Sinnesorgansystcms bei den Wirbeltheren. Ber. it. d.

Verh. d, J. Internal. Zool.-Cong. Berlin, 1901. 621-628. Cameron.

14. On the origin of the Pineal Body as an Amesial Structure, deduced

from the Study of its Development in Amphibia. Anal. Anz., Bd. 23) 394-395- 1903. Same title. Proc. Roy. Soc. Edinb., Sees. 190203. Vol. 24, 572-581. 1904.

15. On the Presence and Significance of the Superior Commissure through out the Vertebrata. Jour. Anal, and Physiol., Vol. 38, 20-21. 1904. Catois.

16. Recherches sur I'histologie et I'anatomie microscopique de I'en cephale chez les poissons. Pp. 172. Lille. 1901. Chiarugi.

17. Di un. organo epitheliale situato al dinanzi della ipofisi e di altri

punti relativi alio sviluppo della regione ipofisaria in embrioni di Torpedo ocellata. Monit. Zool. Ital., Vol. 9, 37-56. 1898. Coggi.

18. Sulla sviluppo della Ampolle di Lorenz. Atti d. Reale Accad. dei

Lhicei. 1 89 1, Ser. 4, Vol. 7.

19. Nouvelles recherches sur le developpment des Ampoules de Lor enzini. Arch. Ital. de Biol., T. },%, 321-333. 1902.

20. Nuove richerche suUa sviluppo della ampolle di Lorenzini. Nota i.

Atti Accad. Lincei, Rendic. CI. Sc.Jis., mat.. Anno 299, Ser. 5, Vol. II, Fasc. 7, Sem. t, 289-297. Nota 2, ibid., Fasc. 8, 338-340. 1902.

21. Sviluppo degli organi di senso laterale, della ampolle di Lorenzini e

loro nervi rispettivi in Torpedo. Archivio Zool., Vol. I, 59-107. 1902. Coghill.

22. The Cranial Nerves of Amblystoma tigrinum. Jour. Comp. Neur.,

Vol. 12, 205-289. 1902. Gushing.

23. The Taste Buds and their Independence of the N. Trigeminus. De ductions from Thirteen Cases of Gasserian Ganglion Extirpation. Johns Hopkins Hosp. Bull., Vol. 14, March-April, 1903. Dogiel.

24. Das periphere Nervensystem des Amphioxus (Branchiostoma lanceo latum). Anat. Heftc, Bd. 21, 147-213. 1902. Dohrn.

25. Die Entstehung der Hypophysis bei Petromyzon planeri. Mitth.

Zool. Sta. Neapel, Bd. 4, 172-189. 1883.

26. Bemerkungen iiber den neuesten Versuch einer Losung des Wirbel.

thierkopf-Problems. Anat. Anz., Bd. 5, 53-64. 1890.

Johnston, Morphology of the Head. 267

27. Studien zur Urgeschichte des Wirbelthirkorpers. IV. Neue Grund"

lagen zur Beurtheilung der Metamerie des Kopfes. Mitth. Zool. Sta. Neapcl, Bd. 9, 330-434. 1890. Edinger.

28. Das Cerebellum von Scyllium canicula. Arch./, iiiik. Anat., Bd.

58, 661-678, I(jOI.


29. The Development of the Optic Vesicles in Amphibia. Jour. Morph.,

Vol. 8, 189-I93. 1890. Froriep.

30. Ueber ein Ganglion des Ilypoglossus und Wirbelanlagen in der Oc cipital region. Arch. f. Anat. n. Physiol. , Anat. Abth., 1882,


31. Ueber Anlagen von Sinnesorganen am Facialis, Glossopharyngeus

und Vagus, iiber die genetische Stellung des Vagus zum Hypoglossus und iiber die Herkunft der Zungenmusculatur. Arch. f. Anat. u. Physiol., Anat. Abth., 1885, I-55.

32. Ueber die Ganglienleiste des Kopfes und des Rumpfes und ihre

Kreuzung in der Occipitalregion. Beitrag zur Entwickelungsgeschichte des Selachierkopfes. Arch. f. Anat. it. Physiol., Anat. Abth., 1901, 371-394.

33. Zur Entwickelungsgeschichte des Wirbelthierkopfes. Verhdl. d.

Anat. Ges. a. d. 16. Vers. Halle, 34-46. 1902.

34. Einige Bemerkungen zur Kopffrage. Anat. Anz., Bd. 21, 515-553.

1902. FUrbringer.

35- Ueber die spino-occipitalen Nerven der Selachier und Holocephalen

und ihre vergleichenden Morphologie. Gegenl>aiirs Festschrift, Bd.

3. 349-788. Leipzig, 1897. Gage, Susanna Phelps.

36. Serial Order of Segments in the Forebrain of Three and Four-week

Human Embryos ; Comparisons with Lower Forms. Abstract. Science, N. S., Vol. 17,486. 1903. Gaupp.

37. Zirbel, Parietalorgan und Paraphysis. Merkel u. BonneVs Ergeb.,

Bd. 7, 208-285. 1898. van Gehuchten.

38. Les Cellules de Rohon dans la moelle epiniere et la moelle allongee

de la truite. Bull, de P Acad. Roy. de Belgique, 1895, 495-519.

39. Contribution a I'etude des cellules dorsales (Hinterzellen) dans la

moelle epiniere des vertebres inferieurs. Bull. P Acad. Roy. de Belgique, 3. Ser., S. 34, 24-38. 1897. Giglio-Tos.

40. Sul' origine embryonale del nervo trigemino del' uomo. Anat. Anz.,

Bd. 21, 85-105. 1902.

268 Journal of Comparative Neurology and Psychology.


41. On the Structure of the Excretory Organs of Amphioxus. Part i.

Quart. Jour. Micr. Sci., Vol. 45, 1902. Goronowitsch.

42. Der Trigemino-Facialis Complex von Lota vulgaris. Gegenbaurs

Festschrift, Bd. 3, 1-44. 1897. Gregory.

43. Beitrage zur Entwickelungschichte der Knochenfischen. Anat.

Ht'fte, 1902. 151-230. Hatschek 44. Die Metamerie des Amphioxus und Ammocoetes. Verhdl. d. Anat.

Ges. 6. Vers. Wien, 136-162. 1892.

45. Zur Metamerie der Wirbelthiere. Anat. Anz., Bd. 8. 1893. Herrick, C. J.

46. The Cranial and First Spinal Nerves of Menidia ; A Contribution

upon the Nerve Components of Bony Fishes. Jour. Comp. A^eur., Vol. 9, 158-455. 1899.

47. A Contribution upon the Cranial Nerves of the Cod Fish. Jour.

Comp. Netir., Vol. lo, 265-322. 1900.

48. The Cranial Nerves and Cutaneous Sense Organs of North American

Siluroid Fishes. Jour. Comp. Neur., Vol. 11, 117-249. 1901.

49. On the Phylogeny and Morphological Position of the Terminal

Buds of Fishes. Jour. Comp. IVeur., Vol. 13, 121-13S. 1903. Hesse.

50. Untersuchungen iiber die Organe der Lichtempfindung bei niederen

Thieren. IV. Die Sehorgan des Amphioxus. Zeit. f. wiss. ZooL, Bd. ei. 1808.



51. Primary Segments of the Vertebrate Head. Anat. Anz., Bd. 16,

353-369- 1899 52. Developmental History of Primary .Segments of the \'ertebrate

Head. Zool. Jahrb., Abth. f. Anat. u. Ontog., Bd. 13, 393-446. 1899.

53- Zur allgemeine Morphologie des Gehirns. ArcJi. f. Anat.u. Physiol., Anat. Abth., 1892, 346-383.

54. Ueber das frontale Ende des Gehirnrohres. Arch.f. Anat. u. Entw.,

1893, 157-171Hoffmann.

55. Reptilien. In Bronn's Klassen und Ordnungen des Thierreichs. 1888.

56. Ueber die Metamerie des Nachhirns und Hinterhirns und iiber Bezie hungen zu den segmentalen Kopfnerven bei Reptilien-Embryonen. Zool. Anz., Bd. 12, 337-339- 1889.

57. zur Entwickelung der Selachii. Morplt. JaJirb., Bd. 24,

209-286. 1896.

58. Fortsetzung. Morph. Jahrh., Bd. 25, 250-304. 1897.

59. Fortsetzung. Morph. Jahrb., Bd. 27, 324-414. 1899.

Johnston, Moifhology of the Head. 269

60. Zur Entwickelungsgeschichte des Sympathicus. i. Die Entwick elung des Sympathicus bei den Selachiern (Acanthias vulgaris). Verhdl. K. Acad. Wetensch. Amsterdam, 1900. 80 p.

61. 2. Die Entwickelungsgeschichte des Sympathicus bei den Uro delen. Ibid., 1902. loi p. Holt.

62. Observations on the Development of the Teleostean Brain with

especial Reference to that of Clupea harengus. Zool. Jahrb., Abth. f. Anat. u. Ontog., Bd. 4, 478-500. 1890. Houser.

63. The Neurones and Supporting Elements of the Brain of a Selachian. Jour. Coinp. New., Vol. 11, 65-175. 1901.

Huber. *

64. Lectures on the Sympathetic System. Jour. Cotnp. Neur., Vol. 7,

73-145. 1897. Johnston.

65. Hind Brain and Cranial Nerves of Acipenser. Anat. Anz., Bd. 14,

580-602. 1898.

66. The Giant Ganglion Cells of Catostomus and Coregonus. Jour.

Comp. Neur., Vol. 10,375-381. 1900.

67. The Brain of Acipenser. A Contribution to the Morphology of the

Vertebrate Brain. Zool. Jahrb., Abth. f. Anat. u. Ontog., Bd. 15, 59-260. 1901.

68. The Brain of Petromyzon. Jour. Comp. Neur., Vol. 12, 1-86. 1902.

69. An Attempt to define the Primitive Functional Divisions of the Central Nervous System. Jour. Comp. Neur., Vol. 12, 87-106. 1902.

70. Das Gehirn und die Cranialnerven der Anamnier. Merkel u.

Bonnefs Ergebnisse, Bd. II, 973-1I12. 1902.

71. The Homology of the Selachian Ampullae. A Note on Allis' Recent

Paper on Mustelus laevis. Atiat. Anz., Bd. 21. 1902. Kerr.

72. The Development of Lepidosiren paradoxa. Part 3. Development

of the Skin and its Derivatives. Quart. Jour. Mtcr. Sci.,Vo\. ^6,

417-459. 1902. Kingsbury.

73- The Structure and Homology of the Oblongata in Fishes. Jour.

Comp. Neur., Vol. 7, 1-36. 1897. Kingsley.

74. The Vertebrate Stomach. Abstract. Science, N. S., Vol. 17, p.

486. 1903. Kolliker

75. Gewebelehre. 6. Auflage, Bd. 2. 1896. Koltzoff.

76. Entwickelungsgeschichte des Kopfes von Petromyzon planeri.

Bull, de la Sac. Imper. d. Natural, de Moscou, Ann6e 1901, No. 3-4, 259-589- 1902.

270 Journal of Comparative Neurology and Psychology.

von Kupffer.

77. Studien zur vergleichenden Entwickelungsgeschichte des Kopfes der

Cranioten. Heft i. Die Entwickelung des Kopfes von Acipenser sturio an Medianschnitten untersucht. Miinchen u. Leipzig, 1896.

78. Heft 2. Die Entwickelung des Kopfes von Ammocoetes planeri.


79. Heft 3. Die Entwickelung der Kopfnerven von Ammocoetes

planeri. 1895.

80. Heft 4. Zur Kopfentwickelung von Bdellostoma. 1900. Lamb.

81. The Development of the Eye Muscles in Acanthias. Atner. Jour. Anat. Vol. i, 185-202. 1902. Langdon.

82. The Sense Organs of Lumbricus agricola Hoffm. Jour. Morph., Vol. II, 193-232. 1S95. Legros.

83- Developpement de la cavite buccale de I'Amphioxus lanceolatus. Contribution a I'etude de la morphologie de la t^te. Arch. Anat. Micros., T. i, 497-542. T. 2, 1-43. 1898. Locy.

84. The Derivation of the Pineal Eye. .-:/««/. ^ws., Bd. 9, 169-180. 1893.

85. New Facts regarding the Development of the Olfactory Nerve.

Anat. Anz., Bd. 16. 1899.

86. A New Cranial Nerve in Selachians. Mark Anniversary Volume,

Art. in, p. 39-55. 1904. McClure. "

87. The Segmentation of the Primitive Vertebrate Brain. Jour. Morph.,

Vol. 4, 35-56. 1890. Marenghi.

88. Alcune Particolarita di struttura e di innervazione della cute delP

Ammocoetes branchialis. Zcit.f. 7viss. Zool., Bd. 75, 421-429. 1903. Maurer.

89. Die Epidermis uiid ihre Abkoramlinge. Leipzig. 1895. ^P- 352Mayser.

90. Vergleichend-anatomische Studien iiber das Gehirn der Knochen fische mit besonderer Beriicksichtigung der Cyprinoiden. Zeit. f. wiss. Zool., Bd. 36, 259-364. 1 88 1. Merkel.

91. Uber die Endigungen der sensiblen Nerven in der Haut der Wirbel thiere. Pp. 214. Rostock. 1 880. Minot.

92. A Laboratory Textbook of Embryology. Pp. 380, Philadelphia.

1903. Mitrophanow.

93. Etude embryog^nique sur les Sdlachiens. Arch, de Zool. exp. et gen., 3. Ser., T. i. 1893

JoHiNSTON, Morphology of the Head. 271


94. The Chiasma of the Toad (Bufo lentiginosus) and of some other

Vertebrates. Zcit.f. JMorph. u. Anthrop., Bd. 3, 183-207. 1901.

95. Beitrag zur Kenntniss des Chiasmas und der Commissuren am Boden

des dritten Ventrikels. Inaug.-Diss. Leipzig. 1902. Neal.

98. A Summary of Studies on the Nervous System of Squalus acanthias. Anat. Anz., Bd. 12, 377-391. 1896.

97. The Segmentation of the Nervous System in Squalus acanthias. A

Contribution to the Morphology of the Vertebrate Head. Bu//. J/tis. Co7)ip. Zool. Harvard Coll., 31, 147-294. 1898. Onuf and Collins.

98. Experimental Researches on the Central Localization of the Sym pathetic with a Critical Review of its Anatomy and Physiology. Arch. Neur. and Psychopath., Vol. 3. 19CX). Orp.

99. A Contribution to the Embryology of the Lizard. Jour. Morph.,

Vol. I, 3II-372- 18S7. Ostroumoff.

100. Ueber die Froriep'schen Ganglien bei Selachiern. ZooL Anz., Bd. 12, 313-364- 1889.


101. Hearing and Allied Senses in Fishes. Contributions from the Biol. Lab. of the U. S. F. C, Woods ILole, Mass., 1903. P. 45-64.

102. The Skin and the Eyes as Receptive Organs in the Reaction of Frogs to Light. Amer. Jour. Physiol., Vol. 10, 28-36. 1903.

103. The Function of the Lateral Line Organs in Fishes. Abstract. Amer. Nat., Vol. 38, p. 496. 1904.

Pinkus 104. Ueber ein noch nicht beschrieben Hirunerven des Protopterus annectens. Anat. Anz., Bd. 9, 562-566. 1894.

105. Die Hirnnerven des Protopterus annectens. Morph. Arbeit., Bd. 4, 275-346. 1S95.


106. A Contribution to the Morphology of the Vertebrate Head based on a Study of Acanthias vulgaris. Jour. Morph., Vol. 5, 79-II2. 1891.

107. Further Contribution to the Morphology of the Vertebrate Head. Anat. Anz., Bd. 6, 251-265. 1891.

108. Ontogenetische DitTerenzirung des Ektoderms in Necturus. Arch. /. mik. Anat., Bd. 43. 1894.


109. Some Points in the Development of a Myxinoid (Bdellostoma stouti, L.). Verhdl. .Anat. Ges. 10. Vers. Berlin. 1896.


110. Die Smirnow'schen freien Nervenendigungen im Epithel des Regenwurms. Anat. Anz., Bd. lO, 570-578. 1894.

2/2 Journal of Comparative Neurology and Psychology.


111. Lo sviluppo ed il valore della cosi detta tasca di Seesel. Arch.

Ital. di Anat. e di Embriol., 2, 344-367. 1903. Sewertzoff.

112. Beitrag zur Entwickelungsgeschichte des Wirbelthierschadels. Anat. Anz., Bd. 13, 409-425. 1897.


113. Ueber freie Nervenendigungen im Epithel des Regenwurms. Anat. Anz. Bd. 9. 1894.


114. The Cranial Nerves of Amphibia. A Contribution to the Vertebrate Nervous System, /our. Morph., Vol. 10, 101-230 1895.

115. The Cranial Nerves of Squalus acanthias. Abstract. Science N. S., Vol. 17, No. 424. 1903.


116. Ueber den feineren Bau der Parietalorgane von Petromyzon marinus. Sitzungsber. d. k. bohm. Gessel. d. Wiss. Math.-nat. CI., 1898, 4 pp.


117. Das Labyrinth der Fische, ein Organ zur Empfindung der Wasserbewegungen. Bihang till K. Svenska J^et.-Akad. Handl., Bd. 28, Afd. 4, p. 21-25. 1903.


118. Primitive Segmentation of the Vertebrate Brain. Quart. Jour. Micr. Sci. Vol. t,t„ 457-477. 1892.

van Wijhe.

119. Uber die Mesodermsegmente und die Entwickelung der Nerven des Selachierkopfes. Amsterdam. 1882.

120. Beitrage zur Anatomic der Kopfregion des Amphioxus lanceolatus. Pctrus Camper, Deel i, Afl. 2. 1900.


121. The Embryology of the Sea Bass (Serranus atrarius). Bull. U. S. Fish Co»i., Vol. 9, 209-277. 1891.

Wilson and Mattocks.

122. The Lateral Sensory Anlage in the Salmon. Anat. Anz. Bd. 13, 658-660. 1897.


123. Kleinhirn des F"rosches. Arch.f. Anat. u. Physiol., Physiol. Abth. Suppl., 1S87.


124. Ueber die Metamerie des Wirbelthierkopfes. Verhdl. d. Anat. Ges. _5. Vers. Miinchen, 1891, 107-113.


125. A Contribution to the Structure and Development of the Vertebrate Head. Jour. Morph., Vol. li, 497-594- 1895.

Johnston, Morphology of the Head. 273


126. Review of Locy No. 86. Jottr. Comp. Neitr. and Psych., Vol. 14, p. 281. 1904.


127. Zur Entwickelungsgeschichte des Selachier Kopfes. Anat. Anz., Bd. 9, 638-653. 1S94.


Abbreviations .

I, 2, J, etc. — mesodermic somites of van Wijhe.

i, ii, iii, etc.— neuromeres of Locy.

A - praeoral entoderm connected with the origin of somite i and the anterior head cavity.

a-l. — area of the primitive neural plate which gave rise to the acustico-lateral system. Hypothetical.

au. V. — auditory vesicle placode.

bi, b2, etc. — branchial clefts.

Bi, B2, etc. — branchial apches.

ch. — notochord.

cil. ^.—ciliary ganglion.

e. I — anterior epiphysis.

e. 2 — posterior epiphysis.

h. — hyomandibular cleft.

H — hyoid ^rch.

hab. — ganglion habenulae.

h. I. — portion of common acustico-lateral placode in selachians which goes to form the head lines.

int. arc. — internal arcuate fibers from cutaneous centers.

L — nerve of Locy (Fig. 10) and its center (Fig. 7).

/. /. — portion of acustico-lateral placode which forms the lateral line.

m. — mouth.

M — mandibular arch.

mes. — mesoderm.

N. I, II, III, etc. — cranial nerves.

N. olf. — nervus olfactorius. — nervus ophthalmicus profundus.

N. thai. — nervus thalamicus of Platt.

op. V. — optic vesicle.

s. — striatum.

s. m. — somatic motor column.

sp. m. — splanchnic motor column.

sp. J'. —splanchnic sensory column.

274 Journal of Comparative- Neurology and Psychology.

s. s. — somatic sensory column. s. V. — secondary vagus tract. sy. g. — sympathetic ganglion.

T — hypothetical center for the nervus thalamicus (visceral). t. b-t. — tractus bulbo-tectalis. /. l-b. — tractus lobo-bulbaris. /. t-b. — tractus tecto-bulbaris. t. t.l. — tractus tecto-lobaris. /. o. — tractus opticus. t. o-h. — tractus olfacto-habenularis. /. .r.///. —tractus sacco-thalamicus. tr. n. — trunk nerve.

An arrow point in Fig. 8 indicates the hypothetical margin of the primitive neural plate; in Fig. lo indicates the anterior end of the brain.


Figures i-y illustrate the longitudinal zones of the brain and the relation of the brain commissures to them. Fig. 7 is a median sagittal section of the brain of a fish on which are projected the zones of the right half of the brain. The location of the various decussations is also indicated by symbols which are explained on the plate. The places of exit of the nerve roots are indicated by the usual Roman numerals. The lateral line roots arise near to or above VIIL The olfactory centers have been shown in the same shading as the visceral sensory centers in the medulla oblongata. The secondary vagus nucleus and the hypothetical nucleus of the nervus thalamicus are also shown in the same way. The ventro-mesial zone, marked by small circles in this and Fig. i, is made up of cells which closely surround the mid-ventral furrow of the ventricle, "hese cells form important nuclei in the inferior lobes, the nucleus thaeniae, corpus inter pedunculare and nucleus of Meynert's bundle (and possibly the lower olive in fishes). This area corresponds to or includes the Bodenplatte distinguished by His in the embryonic brain. Fig's. 1-6 represent cross sections at the levels indicated by the reference lines. More structures are shown in some of the fi-gures than would appear in an actual section.

Fig. I. — Section through the superior commissure, ganglia habenulae, thalamic nucleus of the somatic motor fasciculus, and the inferior lobes in front of the saccus vasculosus. The position of several fiber tracts is shown.

Fig. 2. — Through the tectum opticum, dorsal decussation, ansulate commissure, and the nucleus of N. III.

Fig. J. — Through the cerebellum and the root of N. V. Both somatic and visceral decussations in the cerebellum are shown.

Fig. 4. — Through the root of N. IX. The difference in the course of the secondary tracts from the somatic sensory and the visceral sensory centers is shown.

Fig. 5. — Through the commissura infima Ham.eri.

Fig. 6. — Through the root of a spinal nerve and showing the dorsal and ventral decussations of the spinal cord.

Journal of Comparative Neurology and Psychology. Vol. XV.

Somatic sensory division Splanchnic sensory division Splanchnic motor division Somatic motor division

X X X X Splanchnic sensory decussation, first order Splanchnic sensory decussations,

second and third orders Somatic sensory decussation, second order; posterior commissure IBB Decuss. tr. tecto-lobaris n n n Som. sens, dec, internal arcuate fibers; optic tract Decuss. tr. lobo-bulbaris; ventral

Journal of Comparative Neurology and Psychology. Vol. XV. Plate II.


Journal of Comparative Neurology and Psychology. Vol. XV. Plate 111

F'iQ. 9.

Journal of Comparative Neurology and Psychology. Vol. XV.


Somatic sensory div. Splanchnic sensory div. Splanchnic motor div. Somatic motor di>

Johnston, Morphology of the Head. 275


Fig. 8, A, B, C. — Three diagrams to illustrate the hypothesis of the origin of the acustico-lateral system from the margin of the neural plate. Each functional division of the nervous system is represented by one cell and fiber. Several general cutaneous ganglion cells which are destined to form the acustico-lateral ganglion are shown. The mesoderm and somatic motor nerve are omitted from C, because by the time the acustico-lateral system has reached the grade of development represented, somite 4 no longer forms a myotome.


Fig. g, A-G. — .Seven outline diagrams to illustrate the hypothesis of the origin of the eye by modification of a general cutaneous ganglion and its corresponding centers. The retinal area is indicated by a row of large dots. The relation of the inner wall of the optic vesicle to the choroid plexus, spoken of in the text, is shown in F and G. A comparison of Fig. 9 F with Fig. 4 will show the similarity of position between the retina and acusticum. It will also show that the optic tract fiber from the retina takes a course corresponding to that of the secondary fiber (internal arcuate) from the cutaneous nucleus.


Fig. 10. — A diagram of segmentation in a generalized vertebrate head. In its general features the diagram follows Petromyzon more closely than any other form. So, the number of the gill clefts, their position relative to the somites, the position of the auditory vesicle, and the formation of hypobranchial muscles from myotome 10 and following, are taken from Petromyzon. The neuromeres, the nerve of LocY, the nervus thalamicus, the relation of the head lines to the auditory vesicle, and the praeoral entoderm are taken from selachians. The sensory nerve roots are represented as retaining their attachment to the dorsal surface of the neural tube where they were formed from the neural crest. The segmental position of these roots is about that which they have in the embryo of Petromyzon. except the root of N. X, which has been shifted back a little farther than it is in Petromyzon. The general cutaneous nerve shown in dotted outline over somite 6 is the Vagusanhang of Hatschek in Ammocoetes and the nerve which unites with the vagus root in the embryo of selachians. The position of the viscero. motor nuclei somewhat caudal to the several roots is indicated. The viscero-motor nucleus of the vagus and accessorius is shown as a single large nucleus extending through two segments. It might more properly have been continued caudally until it came into connection with the viscero-motor nucleus of the trunk nerve. The accessorius nerve is not shown. The somatic nucleus and root are shown for all the somites except somite 4, where they are shown in dotted outline. The nerves for both somites 4 and 5 are absent in Petromyzon. Probably both are present in Bdellostoma (a'v and /5^v of Furbringer). It is possible that one or both of these has joined with N. VI in gnathostomes.


Loeb, Jacques. Studies in General Physiology. Chicago, The University of Chicago Press, 1905. Part I. xiii +423, Part II. xi -|- 425-782.

These volumes contain a series of thirty-eight of the author's papers, all of which are reprinted and most of which originally appeared in German. Within the field of comparative neurology and psychology we note the following papers of the series : "The Heliotropisra of Animals and its Identity with the Heliotropism of Plants ; Further Investigations on the Heliotropism of Animals and its Identity with the Heliotropism of Plants ; On Instinct and Will in Animals ; Geotropism in Animals ; The Artificial Transformation of Positively Heliotropic Animals into Negatively Heliotropic and vice versa ; Contributions to the Brain Physiology of Worms ; Has the Central Nervous System Any Influence upon the Metamorphosis of Larvae ? ; On the Theory of Geotropism."

Since the vai'ious papers of these volumes received review notice when first published it will suffice, while announcing the fact that they are now available in English, to call attention to the author's scientific life-purpose and the relation which these papers bear to it. From the preface we quote, "a single leading idea permeates all the papers of this collection, namely, that it is possible to get the life-phenomena under our control, and that such a control and nothing else is the aim of biology. " According to their method of approach to this goal of research the papers fall into three groups : first, those which are concerned with the control of movement ; second, those which deal with the control of regeneration and the determination of organ-formation ; third, which concern the control of the development of the egg.

It is needless to say that Professor Loeb has made great strides toward the accomplishment of his scientific purpose. His work is of great importance; indeed, even when it has followed wrong paths, it has been of value for its stimulating and research-impelling influence. R. M. V.

Literary Notices. 277

Herrick, C. Judson. The Organ and Sense of Taste in Fishes. U. S. Fish Commission Bulletin for 1 902, pp. 237-272, 1904.

The sense of taste in fishes is subserved by a portion of the communis system of neurones. The sensory organs of this gustatory system are either 'taste buds' situated in the mouth and on the lips, or else terminal buds or end buds {Endknospcn, Becherorgane) situated on the outer surface of the head and body and on the fins. These organs have a general resemblance to one another, and are histologically distinguishable from the neuromasts {Nervenhilgel) of the lateral-line system. The outer surface of the body is further provided with sense organs of touch having their own distinct innervation. It has been asserted and denied that the end buds or Becherorgane, found in the outer skin of the body and fins, serve the sense of taste rather than of touch ; but this paper seems to establish conclusively that they are organs of taste.

The proof is as follows. The cat-fish {Ameiurus nebulosus) normally makes very little use of its eyes in getting food, but wanders about waving its barblets gently back and forth or trailing them on the bottom, and darting instantly at any game that they happen to touch. The fish darts in just the same way at a piece of meat (on the end of a wire) that is brought in contact with the barblets or any part of the head or body ; and the reaction is quite independent of the animal's seeing or not seeing the food. Now if cotton is substituted for the meat, the animal will respond in the same way for five or six times, but will then no longer respond to cotton although it will at once respond again to meat. In other words, after a very brief training Ameiurus responds to taste but not to touch. That the response is really to taste and not to different qualities of touch is proved by the fact that while plain cotton remains unnoticed, cotton soaked in meat juice is snapped up instantly. "These experiments seem to show that in the reaction to meat, both from the barblet and from the skin of the body, the senses of taste and touch both participate. This is in accord with the known innervation of the skin and barblets, for all parts of the body surface receive general cutaneous (tactile) nerves, and all parts are plentifully provided with terminal buds (taste buds) which are innervated by communis (gustatory) nerves." The described reaction does not take place unless the stimulus is actually in contact with the skin, and yet the fieh becomes somewhat restless in near proximity to the stimulus, specially if the latter is a piece of meat that has become stale. In order to explain this Professor Herrick suggests a 'sense of smell', meaning by this a stimulation of sensory

2/8 Journal of Comparative Neurology and Psychology.

organs in or about the nostrils of the fish, although he himself elsewhere admits in a similar case the possibility of a slight and widely diffused stimulation of end buds on the barblets and body. Further experimentation showed that there are two well defined types of reflex, the gustatory and the tactile.

Similar experiments on three types of gadoids, the pollock [Pollachius 2'irens) hake ( Urophycis tenuis) and tomcod {Microgadus tomcod), gave entirely consonant results. In the tomcod the olfactory organs of several individuals were extirpated, but the gustatory reflexes seemed absolutely unmodified by the operation. The sea-robin {Prionotus carolinus) finds its food largely by the sense of vision, but also uses not a little its pectoral fin. Now this fin has neither end buds nor communis nerve fibers, and, quite as was to be expected, it shows absolutely no gustatory sensitiveness ; the last observation is confirmed by Morrill. The paper concludes with further interesting observations on the feeding habits of these and other fishes and with interesting general comparative-physiological and comparative-psychological conclusions. E. B. h.

Carlson, A. J- Further Evidence of the Nervous Origin of the Heart-Beat in Limulus. Amer. /our. Physiol., 1905, 12,471-489.

This is another important contribution, by Dr. Carlson, to the study of the relation of the nervous system to heart-beat. The author makes clear that nerve elements are demonstrable anatomically as well as physiologically in the hearts of many invertebrates which are generally supposed to lack them.

Among the striking facts in this paper we note : (i) The ganglion cells of the posterior or venous end of the heart of Limulus are more numerous and exhibit greater automatism than those of the anterior end. (2) The nerve cord on the dorsal side of the heart is a reflex centre, but there is no evidence of a local inhibitory reflex mechanism. (3) Hearts with nerve cord intact vary less in their beating in response to diff"erent intensities of stimulation than do those without neural connections. (4) Mechanical stimuli of various kinds produce or augment the rhythmic contractions of the heart with nerve cord, but after extirpation of the ganglion no contractions result from such stimuli. R. M. Y.

Kendall. W. C. Habits of Some of the Commercial Cat-fishes. Bull. U. S. Fish Commission for 1 902, 399-409, 1904.

This paper brings together the most important published and otherwise available facts on this subject. Habitat, food, feeding and breeding habits are the chief topics considered. i. a. field.

The Journal of Comparative Neurology and Psychology

Volume XV 1905 Number 4


From the Hari'ard Psychological Laboratory .

With Seven Figures.


I. Historical Setting of Subject 279

/. Hearing in frogs and fishes . . . . . . 279

2. The ear of the frog ........ 282

J. Problems ......... 285

II. Reactions of Frogs in Nature to Sounds. . . . 286

III. The Influence of Sounds on Reactions to Other Stimuli 288

/. Infiitence of sounds 071 respiration and visual reactions . 288

2. Infiuence of sounds on tactual reactio7is .... 288

IV. Hearing of Frogs in Air and in Water .... 296 V. The Range of Hearing 298

VI. Relation of the Ear to Reaction to Sound . . . 301 VII. Summary and Conclusions 303

I. Historical Setting of Subject.

/. Heanng in frogs and fishes. The question of the existence of a sense of hearing in frogs is not, as in fishes, an historical problem. For while the question, Do they hear? has been asked repeatedly concerning fishes, and answered as often with 'No' as with 'Yes', students of sense-physiology and animal behavior have assumed, without investigation, that frogs are stimulated by so and. This unquestioning belief in the existence of hearing in frogs is doubtless due to the presence of a prominent ear drum and the ability to produce sounds. Frogs differ from fishes, with respect to the structure of the ear, in that they possess an ear-drum together with a bone which con

28o Journal of Comparative Neurology and Psychology.

nects it with the ear-sac ; and they differ from mammals, to mention only the two chiefly significant features, in that they lack an organ of Corti and a pinna. Anatomically the fish and the mammal represent respectively the simplest and most complex conditions in the vertebrate auditory series. It is not hard to see why investigators should have been interested first and chiefly in the problems of the existence and state of development of the sense of hearing in those vertebrates which have labyrinth organs only (fishes), and in those which have also an organ of Corti (mammals, for example). In the case of the former, lacking as they do all external evidence of an auditory organ as well as the organ of Corti, which is the recognized special organ of hearing in the mammals, it it is not at all surprising that considerable doubt should have existed at various times concerning the existence of a sense which may properly be called hearing. Since frogs stand between fishes and mammals in the vertebrate auditory series, they would naturally be expected to hear in case fishes do. Attention has long centered where the existence of hearing seemed most doubtful, and it will be profitable for us, before taking up the special evidence of hearing in the frog to note the result of the most recent study of hearing in fishes.

The history of the investigation of the functions of the ear in fishes is a curiously interesting and instructive chapter in the progress of sense-physiology, for it exhibits the development of scientific method, as well as the gradual increase in the scope and exactness of knowledge. A brief historical sketch of the subject is given by Parker^ in a paper on hearing and allied senses in fishes. In this excellent contribution to the subject the author, after defining hearing "as that sensory activity resulting from a stimulation of the ear by material vibrations" (p. 46), states that Fundulus and several other fishes, which when normal respond definitely to vibrations of a frequency of 128 per second transmitted directly to the water of an aquarium,

^Hearing and Allied Senses in Fishes. Bu//. U. S. Fish Com. 1902. pp. 45-64.

Yerkes, The Sense of Hearing in Frogs. 281

fail to respond after the eighth nerves have been cut. The conclusions to which Parker was led by his vivisectional experiments are clearly and briefly stated in his own words, thus :

"The skin, lateral-line organs and ears represent, figuratively speaking, three generations of sense organs. The oldest is the skin stimulated by varying pressures, such as are produced by irregular currents, and capable of initiating equilibration responses. From the skin have been derived the lateral-line organs stimulated by water vibrations of low rate, and also significant for equilibration. Finally, from the lateral-line organs have come the ears stimulated by water vibrations of a high rate and important for equilibration. The ear, unlike the skin and lateral-line organs, is differentiated for its two functions, the sacculus for hearing, the utriculus for equilibration."^

In the literature on the senses and reactions of frogs I have found nothing which contributes importantly to our knowledge of the sense of hearing. Most of the investigations which deal with the ear are concerned with the equilibrational and orientational fimctions of the labyrinth organs and have nothing whatever to say about hearing. In the natural histories the existence of a well-developed sense of hearing is usually assumed, and numerous instances of what are supposed to be reactions to sound are cited. It is to be noted, however, that none of the observations in these popular works furnishes satisfactory proof of the exclusion of the influence of visual stimuli. Among the few references to frog audition of which I have knowledge the only one which seems worthy of special notice is that of Gaupp in his "Anatomie des Frosches. " Since his few paragraphs sum up the state of our knowledge on the subject, while at the same time furnishing an illustration of the assumption of hearing on the basis of analogy, I present the substance of them in free and slightly abbreviated translation.

"The labyrinth organ has an acoustic and a non-acoustic (static) function^ For these two functions, according to the leading if not generally accepted view, entirely different portions of the organ are in question, and since the non-acoustic is attributed to the three Cristae acusticae ampidlai-iun and the three Alaculae {M. recessiis utrictdi, M. sacculi, M. lagenae) there remain for the acoustic function only the Papilla hasilaris and the Macula neglccta. It is not certain, however, that the non-acoustic organs do not participate in the acoustic function.

"With regard to the acoustic sense of the frog nothing exact is known..

'Abstract of paper read before Section F. of the Amer. Assoc, for the Advancement of Science in Philadelphia, 1904. Science, 21, 265, 1905.

282 Journal of Comparative Neurology and Psychology.

That it exists, and that in good development, is certain. The existence of the drum and columella, and the fact that frogs have a voice are unmistakable proofs of hearing. The participation of the Papilla basilaris in acoustic functions is rendered certain by comparative anatomical studies : the Papilla basilaris is the nerve end-organ from vi^hich, in the mammalia, the undoubtedly acoustic organ of CoRTi arises. From analogy of structure we may also infer an acoustic function in the Macula neglecta : on this as on the Papilla basilaris, there is a simple tectorial membrane, and further the Pars neglecta, like the Pars basilaris, has a strong thick wall which only in a limited region, namely, where it approaches a part of the perilymphatic space, is markedly thinner. (For fish Breuer (1891) has already stated that if they really hear — which is not proved — the Macula neglecta^ alone can come into consideration in connection with the function for there is no Papilla basilaris in fishes and the six other nerve end-organs apparently serve the non-acoustic function.)

. "In what way the sound waves come to the Macula neglecta has already been discussed : it is most probable that the movements of the perilymph, which are produced in the Spatium sacculare cave perilymphatici are transmitted to the endolymph of the Succulus and that the waves of the endolymph continue from the last directly to the Pars basilaris and Pars neglecta. The thin walled portions of these two regions would have value as elastic vibratory structures, but they could not serve, as might appear, for the transmission of waves from the perilymph to the endolymph.

"The fact that the vibration of the drum must be transmitted by the plectrum and the Spatium sacculare cave perilymphatici first to the Sacculus deserves consideration from the point of view of the assumption of a non-acoustic function of the Sacculus. It must remain a question whether the movements produced in the endolymph of the Sacculus are without effect upon the Macula sacculi, perhaps by reason of the very large otoliths, or whether they do not come to perception also at this place [Macula sacculi) . This same question arises also for the Lagena, in which the wave of the endolymph of the Sacculus must be continued, as in the Pars neglecta and the Pars basilaris. It is less probable that the movements may be continued as far as the Macula recessus utriculi and the three Crista/ ampullarum. The question of Hensen is certainly justified in the case of frogs (1893) "Hier (bei den Froschen) ist aber schon der acustiche Zuleitungsapparat nicht erheblich viel schlecter als bei den Vogeln entwickelt ; sollte das nur zu Gunsten des kaum im Warden begriffen Schnelkenanhangs geschehen sein?'"

2. The ear of the frog. The anatomical references in Gaupp's discussion may be made more intelligible by a brief description of the ear. It consists of a large external membrane, the tympanum, usually not more than a centimeter in diameter in the green frog (see Fig. 3), the columella which connects the tympanum with the labyrinth (see Fig. 2), and the

'Anatomic des Frosches. VI. Lehre von Integument und von den Sinnessorganen. S. 751-52. 1904.

Yerkes, The Sense of Heanng in Frogs.


labyrinth with its several sense organs. The labyrinth is constituted by three semicircular canals each of which has a sense-organ, the Crista acustica, in its ampulla, the utriculus, the sacculus and

Amp. lat. Sacc.

Ca. B. c. ant.

Ca. 8. c.


Duct, endol.





ant •

Fig. I. Membranous labyrinth of Rana esculenta. After Gaupp's modification of Retzius' drawing ; A, seen from the outside ; B, seen from the inside.

the lagena, in which together there are five sense organs. Figure i, a and b are reproduced from Gaupp's "Anatomic des

284 Journal of Comparative Neurology and Psychology.

Frosches." Fig. i a is a lateral view of the labyrinth organs; Fig. I B a median view. These figures together represent clearly the important sensory portions of the ear of the frog. Since there is no organ of Corti, whatever ability to hear the animals may possess must be due to the functioning of some portion of the membranous labyrinth.

Figure 3 is presented for the purpose of showing the striking difference in the size of the tympanum of male and female green frogs. The individual on the right is a male, that on the left a female. Although these two frogs were of the same length and weight the maximum diameter of the tympanum in




Fig. 2. Transverse section of the head of the frog to show the relations of the various parts of the ear (diagrammatic), tymp. memb., tympanic membrane; col., columella; stp., stapes ; memb. lab., membranous labyrinth ; Nv. VIII., auditory nerve; med. obi., medulla oblongata ; ch.plx., choroid plexus; tymp. cav., tympanic cavity \fen. ov., fenestra ovalis ; buc. cav., buccal cavity ; eus.i.. Eustachian tube ; an. tymp., annulus tympanicus ; b. hy., body of hyoid ; o. st., episternum ; mnd., mandible ; //f., pterygoid ; qu.ju., quadrato-jugal. From Parker and Haswell.

the male is one-third greater than that in the female. Measurements of several other individuals yielded the following results. For ten males whose average length was 7.31 cm. the average maximum diameter of the tympanum was 8.15 mm. ; for ten females 6.52 cm. in length it was 5.99 mm. Or if we compare individuals of like size: for a male 6.2 cm. the diameter was 6.3 mm. in contrast with 5.0 mm. for a female of the same

Yerkes, The Sense of Hearing in Frogs.


size; in case of two individuals 6.5 cm. the measurement of the male was 7.5 mm., that of the female 6.0 mm.; the tympanum of a male 8.8 cm. long measured 11. 5 mm., that of a female 9.0 cm. long, 9 mm. In almost all cases the difference in the size is so marked as to be noticeable to a casual observer.

J^ig. J. Photograph of green frogs (Rana ciatnitans), male on the right, female on the left. To show the difference in size of tympana.

J. Problems. The purpose of the investigation of which this paper gives an account was the study of the influence of sounds on the reactions of the frog. The definite questions for which answers were sought are four: i. Do frogs react to sound? 2. Do they hear? 3. If so, what sounds are heard? 4. Under what conditions do reactions to sounds occur ? No attempt was made to localize the function of audition within the labyrinth organs, for the problems of research concern behavior rather than the functions of special organs.

In studies of this sort it is to be remembered that normal behavior can not be studied by vivisectional methods. Oper

286 Journal of Comparative Neurology and Psychology.

ations on sense-organs, nerve cutting and like methods may aid us in determining what organ or portion of an organ is necessary for a particular function, but they can give us no trustworthy information concerning the relation existing between the senses and the normal behavior of the animal. Therefore, in the experiments now to be described, with the exception of those which were made to ascertain whether the ear is necessary for reaction to sounds, normal frogs were studied either in their native haunts or in the laboratory. All the detailed work was done with the green frog, Rana clainitans, but tests were made also with the leopard frog, R. pipiens, and the bull frog, R. catesbiana.

II. Reactions of Frogs in Nature to Sounds.

My attention was first drawn to the subject of frog audition by failure to obtain motor reactions to sounds in an investigation of the time relations of the neural process of the green frog. Although a large number of sounds of different qualities, pitches and intensities were employed, no visible motor reactions were observed. This led me to seek the significance of what appeared to be either a surprising lack of sensitiveness to changes in the environment which would naturally be expected to stimulate the animal, or an interesting and important case of the inhibition of reaction to auditory stimuli. The question to be answered is, Are frogs deaf, or do they under certain conditions completely inhibit their usual reactions to sound ?

Since they bear upon the question of deafness I quote the following observations on the influence of sounds in Nature from the auditory-reaction section of my earlier reaction-time paper. ^

In order to learn how far fear and artificial conditions were causes of the inhibition of responses to sounds in the laboratory, and how far the phenomenon was indicative of the animal's inability to perceive sounds, I observed frogs in their native haunts.

By approaching a pond quietly, it is easy to get within a few yards of frogs sitting on the banks. In most cases they will not jump until they have evidence

^The Instincts, Habits and Reactions of the Frog. Harvard Psychological Studies, 1, 629-630 {Psychological Review Monogy-aph, 4), 1903.

Yerkes, The Sense of Heaj'ing in Frogs. 287

of being noticed. Repeatedly I have noted that it is never possible to get near to any frogs in the same region after one has jumped in. In this we have additional proof that they hear the splash-sound. To make sure that sight was not responsible for this on-guard condition in which one finds the frogs after one of their number has jumped into the water, I made observations on animals that were hidden from one another. The results were the same. I therefore conclude that the splash of a frog jumping into the water is not only perceived by other frogs in the vicinity, but that it is a peculiarly significant sound for them, since it is indicative of danger, and serves to put them ' on watch.'

A great variety of sounds, ranging in pitch from a low tone in imitation of the bull frog's croak to a shrill whistle, and in loudness from the fall of a pebble to the report of a pistol, were tried for the purpose of testing their eflfects upon the animals in their natural environment. To no sound have I ever seen a motor response given. One can approach to within a few feet of a green frog or bull frog and make all sorts of noises without causing it to give any signs of uneasi' ness. Just as soon, however, as a quick movement is made by the observer the animal jumps. I have repeatedly crept up very close to frogs, keeping myself screened from them by bushes or trees, and made various sounds, but have never succeeded in scaring an animal into a motor response so long as I was invisible. Apparently they depend almost entirely upon vision for the avoidance of dangers. Sounds like the splash of a plunging frog or the croak or pain-scream of another member of the species serve as warnings, but the animals do not jump into the water until they see some sign of an unusual or dangerous object. On one occasion I was able to walk to a spot where a large bull frog was sitting by the edge of the water, after the frogs about it had plunged in. This individual, although it seemed on the alert, let me approach close to it. I then saw that the eye turned toward me was injured. The animal sat still, despite the noise I made, simply because it was unable to see me ; as soon as I brought myself within the field of vision of the functional eye the frog was off like a flash.

Many observers have told me that frogs could hear the human voice and that slight sounds made by a passer-by would cause them to stop croaking. In no case, however, have such observers been able to assert that the animals were unaffected by visual stimuli at the same time. I have myself many times noticed the croaking stop as I approached a pond, but could never be certain that none of the frogs had seen me. It is a noteworthy fact that when one frog in a pond begins to croak the others soon join it. Likewise, when one member of such a chorus is frightened and stops the others become silent. This indicates that the cessation of croaking is a sign of danger and is imitated just as is the croaking. There is in this fact conclusive evidence that the animals hear one another, and the probability is very great that they hear a wide range of sounds to which they give no motor reactions, since they do not depend upon sound for escaping their enemies.

The phenomenon of inhibition of movement in response to sounds which we have good reason to think the frogs hear, and to which such an animal as a turtle or bird would react by trying to escape, is thus shown to be common for frogs in nature as well as in the laboratory. This inhibition is in itself not surprising, since many animals habitually escape certain of their enemies by remaining motionless, but it is an interesting phenomenon for the physiologist. We

288 Journal of Comparative Neurology and Psychology.

have to inquire, for instance, what effects sounds which stimulate the auditory organs and cause the animal to become alert, watchful, yet make it remain rigidly motionless, have on the primary organic rhythms of the organism, such as the heart-beat, respiration, and peristalsis. It is also directly in the line of our investigation to inquire how they affect reflex movements, or the reaction time for any other stimulus — what happens to the reaction time for an electrical stimulus, for example, if a loud noise precede or accompany the electrical stimulus.

III. The Influence of Sounds on Reactions to Other


/. Influence of sound on respiration and visual reactions. — Observations described in my earlier paper prove that respiration is modified by sounds, and it was also noted that the attempts of a frog to seize a moving object are reinforced by the sound of a tuning fork.^

2. Influence of sounds on tactual reactions. — A more detailed study has been made of the influence of sounds on tactual reactions and of the significance of the temporal relations of the stimuli.

A reflex movement of the leg was chosen as an indication of the action of the stimuli and the influence of sounds was observed under the following conditions. A frog was placed on a saddle-like holder and kept in position by linen bands over the back and a wire screen cap over the head, as shown in Fig. 4. Under these conditions the hind legs usually hang free and limp and any movement which may be made by them in response to a stimulus can be read in millimeters from a scale, attached to the holder. This method of measuring the value of stimuli in terms of leg reflex has been used by several other investigators, most recently by Merzbacher. In this connection it is of interest to note that neither Merzbacher nor Ewald were able to detect movements of the leg of the frog in response to sounds. Even the croaking of another frog near by had no apparent effect.^

'1. c. p. 634. »

'Merzbacher, L. Ueber die Beziehungender Sinnesorganezur den Reflexbewegungen des Frosches. Pfiilger's Arch., 81, 254, 1900.

Yerkes, The Sense of Hearing in Frogs. 289

I found it desirable, as did Merzbacher, to observe the ■movements of a shadow of the leg on the scale and thus read the amount of movement, rather than to watch the leg itself and attempt to project it upon the scale.

As is indicated in the figure the auditory and tactual stimuli were given automatically by means of a swinging pendulum (/*) which was held in position by the magnet a until released by the experimenter. Early in its swing the pendulum turned the key in thus completing a circuit which caused the auditory stimulus to be given; later in the swing the key 71 was turned, and the tactual stimulus thus given through the magnetic release of the lever /. The interval between the auditory and the tactual stimuli could be varied from.i" to .9" by change in the position of the key n. For giving the two stimuli simultaneously a double hand key was employed.^

The auditory stimulus was either the sound of a quick hammer blow or the ringing of an electric bell for a certain interval. Figure 4 shows the bell. It was placed 80 cm. from the frog, and in order that the influence of vibrations might be avoided was suspended from the pendulum frame. When the hammer was used it was placed 60 cm. from the frog, on the pendulum table. The frog and the apparatus for tactual stimulation occupied a separate table which was not disturbed by the jars of the pendulum table. The tactual stimulus was given by a rubber cone, T, Fig. 4, 2 mm. in diameter at its apex. This rubber point, after the electric release of the lever to which it was attached, struck the frog at the middle point of a line drawn between the posterior margins of the tympana. The intensity of the stimulus could be varied by weighting the lever (see w in the figure).

Under the conditions of experimentation described above, a tactual stimulus regularly causes a reflex movement of the

^A full account of this method and the results of a study of the phenomena of auditory-tactual reinforcement and inhibition may be found in Pfiilger' s Archiv, Bd. 107, S. 207, IQ05, under the title "Bahnung und Hemmung der Reactionen auf tactile Reize durch akustiche Reize beim Frosche." In this connection I mention only such aspects of the investigation as bear on the subject of audition.

290 Journal of Comparative Neurology and Psychology.

Fig. 4

Yerkes, The Sense of Hearing m Frogs. 291

suspended leg, which varies in amount with the state of the animal and the strength of the stimulus. A sound, on the contrary, never causes the slightest movement, and that no matter how loud or high it be. This makes it possible to study the influence of sounds on tactual reactions, under different temporal relations of the stimuli, for even when the sound precedes the touch there is no reaction until after tactual stimulation.

The results now to be presented were given by the green frog, but test experiments indicate that both the leopard and the bull- frog are influenced in similar manner by sounds.

The experimental procedure was as follows. After an animal had been placed in the proper position and had ceased to struggle to escape, reactions to stimuli were taken in pairs regularly at half minute intervals, first a reaction to the tactual stimulus alone, then a reaction to the same intensity of touch when accompanied or preceded by an auditory stimulus.

The series consisted of 50 pairs of reactions taken without pause. So far as the frog is concerned there seems to be nothing undesirable in long series, for there is no indication of fatigue, and so long as the animal is kept moist and in a comfortable position it does not often struggle to escape. The advantage, for the purposes of this investigation, of taking the reactions in pairs is obvious. It permits us to compare directly the reactions of each pair, and to note at once whether the auditory stimulus has reinforced or inhibited the tactual reaction.

During a series the intensity of the tactual stimulus was changed as conditions demanded, but for any one pair of reactions it was always the same. It not unfrequently happened that an intensity which at first caused only slight movement of the leg, later in the series uniformly brought about a maximum contraction, or the reverse might be true, and since the maxi Ftg. 4. Auditory-tactual apparatus. (Drawn by Dr. Wm. E. Hocking.) P, pendulum; p, contact point of /"; b, attachment for electro-magnet, a; m, key for electric bell circuit ; B, electric bell ; n, key for magnet circuit of touch apparatus ; K, hand-key for release of pendulum and temporary closing of electric bell circuit ; k^, k^, P, keys in circuits ; e,/, g, magnetic release for touch apparatus ; /, pivoted lever bearing rubber cone, T, and weights, w.

292 Journal of Comparative Neurology and Psychology.

mum amount of movement left no opportunity for judging of the influence of the auditory stimulus it was always necessary in such cases so to alter the intensity of the tactual stimulus that a medium reaction would result.

The animals seldom struggled during experiments, but if too firmly bound they became irresponsive to the stimuli.^ It was therefore necessary to place them carefully in position, and then, after they had ceased to struggle, to draw the linen bands over them just tightly enough to prevent change in position. For the purpose of excluding the influence of visual stimuli a wire screen cap covered with black cloth was put over the head. This served to help keep the frog in the proper position as well as to exclude visual stimulation.

TABLE I. The Influence of Sound on the Tactual Reactions of the Green Frog.

Green frog A.

Tactual Stimulus


Auditory and Tactual Stimuli

Position of shadow


Position of shadow


No. of Exp.

on scale


on scale


before after





the stimulus

the stimulus

Pair I





90 M*


" 2





90 M


" 3





90 M


" 4





90 M


" 5







" 6





90 M


" 7







" 8







" 9





" ID







Green frog B

Pair 1





90 M


" 2





90 M


" 3





90 ]\I


" 4





90 M


" 5





90 M


" 6







" 7





90 M


" 8





90 M


" 9





90 M


" 10





90 M


■M indicates maximum movement of leg.

'A case of inhibition.

Yerkes, The Sense of Hearing in Frogs. 293

The two series of reactions of Table I, chosen at random from several hundred, will serve to indicate the method of recording the results as well as the nature of the results themselves. The first ten pairs of reactions fairly represent the variableness of the reactions; the second show still more clearly the influence of the auditory stimulus.

The results obtained with four frogs, two males (Nos. i and 3) and two females (Nos. 2 and 4) will suffice to indicate the influence of sounds on tactual reactions. For each of the four frogs fifty pairs of reactions were taken in series, and that for each of seven different temporal relations of the two stimuli. The individual averages therefore are based upon fifty reactions, and the total number of reactions for each individual is seven hundred. As is clear from column one of Table II, the temporal relations of the stimuli ranged from simultaneity to .9" (i. e., the auditory preceded the tactual by .9").

The influence of the sound, which for these experiments was a sudden hammer blow, is discovered by direct comparison of the tactual reaction of each pair with its corresponding auditory-tactual reaction. When the tactual reaction is the greater, we infer that the sound has partially inhibited reaction ; when it is the smaller, that it has reinforced reaction ; when the two are equal, that it has been without influence. The influence of sound may be expressed either in terms of the number of reactions reinforced, inhibited and equal, or in terms of the amount of reinforcement or inhibition. Both methods have been employed. Table II presents the percentage value of the auditory-tactual reactions in comparison with the tactual, and also the number of reactions over half which were reinforced or inhibited, while Figures 5 and 6 graphically represent the amount of influence in terms of the tactual reaction. The auditory-tactual reaction is always expressed as so many per cent greater (+ i. e., reinforcement) or less ( — i. e., inhibition) than the tactual. In the table + always indicates reinforcement, — inhibition, and in the curves the portions above the zero line indicate reinforcement, those below it inhibition. The number of reactions over half, that is over twenty five, since there were

294 Journal of Comparative Neurology and Psychology.

fifty pairs of reaction for each interval and each animal, which were reinforced or inhibited furnishes an excellent quantity for comparison with the averages. As such comparison reveals close agreement between amount of influence of the auditory stimulus and the number of reactions either reinforced or inhibited it is clear that the averages are trustworthy, even though the variability of the reactions is enormous.


Influence of Sound on the Tactual Reactions of the Green Frog for

Different Temporal Relations of the Stimuli.

Length of interval

Dift". in per cent, between tactual and auditory-tactual reactions

Number of reactions reinforced or inhibited

Diff. in per cent, between the tactual and auditory tactual reactions

Number of reactions reinforced or inhibited

Male No. i.

Male No. 3.

■AS .90

-f 62.0 fo + 30.3 + 33-Z + 1-7

— 22.3

— 10. + 0.7

+ 17-0 + I7-0 + I3-0 + 0.5


— 6.0 + 0.5

+ 85.9' + 31-3 + 6.4

— 4-8

— 15.0

— 2.1

+ 18.0 + I7-0 + 12.5 -\- 2.0

— 4-5

— 6.5

— 2.5

Female No. 2

Female No. 4.





+ 60.9 Cfo

+ 19.8 + 30.4

- 6.3

— 8.1 -15-8

— 0.9

— 14-5

— 6.5 -(- 12.0

— 3-0

— 3-5

— 9-5 + 3-0

+ 40-7 I0 + 3I-I + 49-2

— 132

— 19.8

— 7-8

— 51

+ 1..0 + 10.5 + 13-5

— 3-5

— II.

— 5-5

— 4.0

The results tabulated in Table II show that the sound when occurring simultaneously with the tactual stimulus greatly increases the amount of the reflex, while on the contrary it decreases the amount of reaction if it precedes the tactual stimulus by .5 to i.o". Number i, for example, exhibited reinforcement equal to 62 % of the amount of the tactual reaction when the stimuli were given simultaneously, but when the sound occurred .65" before the touch the resulting reaction was 10 % less than the tactual.

In Figures 5 and 6 are curves, representing the amount of reinforcement and inhibition for the four frogs, constructed

Yerkes, The Sense of Hearing in Frogs.


according to the following method. Let the zero point on the ordinates represent the value of the tactual reactions; then the value of the auditory-tactual may be indicated on the same ordinate as so many per cent, greater (above the zero point) or less (below the zero point) than the tactual. In Fig. 5, for instance, the sound when simultaneous with the touch caused a

Fig. 5 Fig. 6

Fig. J-. Reinforcement-inhibition curve for momentary auditory stimulation, contracted on the basis of amount of reaction. Curve for male No.

1 ; Curve for Male No. 3.

Fig. 6. Reinforcement-inhibition curve for momentary auditory stimulation,

constructed on the basis of amount of reaction. Curve for Female No.

2 Curve for Female No. 4.^

'The curve for No. 4, Fig. 6, as originally published in Pft.i)ger's Archiv, Bd. 107, S. 219, is incorrect in that it starts with 50.7 (?^ instead of vi^ith 40.7 cjf,.

296 Journal of Comparative Neurology and Psychology.

reaction 62 % greater than the tactual, in the case of one frog, and 103. 1 % greater in the case of the other. On the other hand, when the sound preceded the touch by .45" the resulting auditory-tactual reaction was 22.3 % and 4.8 %, less, respectively, than the tactual.

The figures on the left margin of the curves above the zero point indicate reinforcement in per cent, of the tactual reaction, those below the zero point, inhibition. Below the base line the time intervals are given in tenths of a second. Each curve is plotted, from the data of Table II, on the basis of seven hundred reactions. It is to be noted that the curves of Figure 5, for the two males, show considerably more reinforcement than those for the females in Figure 6. Furthermore, the curves for the females cross the transition line between reinforcement and inhibition sooner and return to it more slowly than do those for the males. In other words, inhibition begins with a shorter interval between the stimuli and continues longer.

These experiments prove conclusively that sounds, although they do not call forth the reflex movement under consideration, modify in important ways the action of other stimuli. It is therefore certain that lack of auditory reaction is due to some form of inhibition and not to insensitiveness.

IV. Hearing of Frogs in Air and in Water.

It has* been held by many investigators of the sense of hearing that sounds in the air cannot be heard by animals under water for the simple reason that the air waves cause only very slight disturbances in the water. In view of this statement it is of interest to test the ability of the frog to hear when the tympanum is exposed to air and when it is under water.

Experiments were made with the apparatus represented in Fig. 7. The reflex reaction method of testing the influence of sounds was again employed, and pairs of reactions were recorded for frogs whose ears were either exposed to air waves or submerged in the water of the aquarium. The level of the water, with reference to the ear, was controlled by changing the volume in the aquarium, and the leg of the frog was kept from

Yerkes, The Sense of Hearing in Frogs.


floating upward by the attachment of a one gram weight. The two stimuli were given by means of the hand key, K. In all cases the auditory stimulus was the sound of an electric bell, B^ suspended above the water in such a way that its vibrations could not be transmitted either to the aquarium or the frog holder. Precautions were taken to exclude also the influence of visual stimuli.

During the month of December, 1904, ten green frogs were tested with this apparatus and it was found, contrary to the expectations aroused by previous results, that the females

Fig. 7. Auditory apparatus for testing hearing in air and in water. A^ aquarium ; B, electric bell ; T, tactual stimulus apparatus ; K, hand-key for giring stimuli ; tc, weight to hold leg.

reacted far more uniformly and vigorously to sounds than the males. Indeed the irregularity of response and irresponsiveness of the males were so marked that it was seldom possible to get more than two or three pairs of reactions in series. For this reason it has been impossible to present averages for the males. The females reacted with a fair degree of regularity, but with notning like the vigor and uniformity which characterized the reactions of animals tested in April and May. Evi

298 Journal of Comparative Neurology and Psychology,

dently the seasonal condition of the animal is an important matter to consider in studies of audition.

The experiments included tests under three conditions: I. When the tympanum was exposed fully to air, although the body was submerged up to the level of the eardrum ; 2. When the tympanum was half under water, the head and nares being in air ; 3. When the frog was submerged to a depth of 4 cm.

The results prove beyond doubt that sounds made in the air stimulate frogs when their tympana are under water. Furthermore, there is evidence that sounds stimulate the green frog even when it is totally submerged to a depth of four centimeters. I cannot better describe these results than by giving the sample series of Table III.

Although the amount of reaction under the conditions of these experiments is small, the difference between the tactual and the auditory-tactual is sufficiently great to establish the frog's ability to hear under water.

The averages for ten reactions of another female under four sets of external conditions follow :

Condition Amount of reaction to Amount of reaction to the

tactual stimulus alone. auditory-tactual alone.

Tympanum in air 2.0 mm. 16.4 mm.

•' half under water .9 " 13.5 "

" 1 cm. under water 1. 3 " 3.5 "

" 4 cm. under water .5 " I. a "

With several other individuals reactions were obtained which just as clearly indicated audition under water. We may therefore conclude that the green frog can hear both in air and under water.

There is some evidence that the reaction to sound is greats est when the tympanum is half submerged.

V. The Range of Hearing.

By one or another of the methods of experimentation already described the value of the following sounds as auditory stimuli for the frog has been demonstrated during the course of this investigation. The croaking of green, leopard and bullfrogs ; splashing of water ; pistol explosions ; tuning forks rang

Yerkes, The Sense of Heating in Frogs. 299


Hearing of Green Frog in Air and Under Water.

Female. Dec. 5, 1904. Weight for tactual stimulus 5 grams.

No. of pair (Continuous series)

Tactual stimulus alone

Auditory and tactual stimuli

Position of shadow on scale




Position on shadow on scale

Amount of

before 1 after

before | after

the stimulus

the stimuli

Series I. Tympanum exposed to air.

I 2

3 4

28 27 26 25

30 30 29


2 mm. 3 3 2

27 26 26 25

50 30

29 28

23 mm. 4 3 3

Series II. Tympanum entirely under water.

5 6

7 8

25 25 25 19

25 27 28 19



25 25 25 19

32 28 28 21

7 3



Series III. Tympanum half under water.

9 10 II 12


19 19 19


19 20 21


I 2

19 19

19 19

22 22

25 21

3 3 6


Series IV. Frog submerged to depth of 4 cm.

13 14


19 19 18 18

20 20 18 19



19 19 18 18

22 21 20 20


2 2 2

Series V. Tympanum exposed to air as in Series I.





18 18 18 18


19 20 20


I 2

18 21 18 21 18 22 18 25

3 3

4 7

ingfrom 100 to 1000 vibrations per second; electric bell with metal gong and also with wooden gong ; sudden hammer blowj Galton whistle ; Appunn whistle, and a variety of sounds produced by the human vocal organs.

The sound of the electric bell produces the most marked modification of reaction, probably because it consists, hke the induced electric shock, of a rapid succession of stimulating

300 Journal of Comparative Neurology and Psychology.

changes. The green frog is stimulated by sounds as low as 50 vibrations per second. No experimental tests were made with lower sounds. For the purpose of determining the upper limit of audibleness tests were made with Galton and Appunn whistles. The tactual reflex was again employed as a means of testing the influence of sound. The tests, like those of the previous section of the paper, were made during the winter and responses to both tactual and auditory stimuli were obtained much more regularly with females than with males. A sufficient number of reactions have been recorded for males, however, to prove that they are influenced by sounds as high as 10,000 vibrations per second, as are the females also. The following series of averages for a female is indicative of the nature of the results of these tests :

Green frog Galton whistle Amount of tactual reaction Amount of auditory-tacfemale vibration rate Average of twenty reactions tual reaction in each case.

Series I 5, 000 2.80 mm. 4.65 mm_

•' II 8,000 2.60 " 5.45 "

" III 10,000 2.10 '• 4.20 "

" IV Puff of air alone 2.45 " 2. 30 "

Series IV serves as a control on the influence of the sound of escaping air which accompanies the tone of the whistle. Its result indicates that the tone is the effective stimulus in the preceding series.

Similar results obtained by the use of the Appunn whistles prove that up to 10,000 vibrations per second the frog is stimulated by sound. With neither the Galton nor the Appunn whistles was evidence of reaction to sounds of higher vibration rate obtained. We may therefore conclude, until further investigation by more satisfactory methods is available, that the green frog is influenced by sounds ranging in rate of vibration from 50 to 10,000 per second. It should be noted, however, that neither the upper nor the lower limit of audition has been accurately determined by these experiments.

Yerkes, The Sense of Hearing m Frogs. 301

VI. The Relation of the Ear to Reaction to Sound.

Abundant evidence has now been presented in support of the statement that the frog is stimulated by sound, but the use of the term audition in connection with these reactions has not been justified. If the investigation were dropped at this point the criticism would doubtless be made that the modifications of tactual reactions produced by sounds may be due to stimulation of certain cutaneous sense organs instead of the organs of the ear, and that therefore nothing has been proved concerning hearing in the frog. In anticipation of this objection to the conclusions which have been drawn from the results of experimentation check observations were made on frogs whose ears had been operated upon in various ways.^

The operations whose effects serve as evidence of the relation of the ear to the responses to sounds which we are now considering were three : (i) Cutting the tympana ; (2) Cutting the columellae as well as the tympana ; (3) Cutting the eighth nerves from the dorsal side."

Briefly stated, the results of the three operations mentioned are as follows.

1. After cutting of the tympana there is no apparent change in the nature of the influence of sounds. The frogs are at times less responsive to stimuli, and as a rule they do not show as marked reactions to either touch or sound as do the normal animals.

2. Sounds continue to modify tactual reactions after both columellae and tympana are cut. In one case it was noticed that the influence of the sound of a wooden gong was much increased by this operation, whereas there was no marked change in the influence of the sound of the metal gong. There

^Through the kindness of Mr. J. H. Mason, graduate student in the department of zoology of Harvard University, I was enabled to make the following tests on frogs which he had operated upon for the purpose of studying the static functions of the labyrinth organs.

^Postmortem examination in case of attempted cutting of the nerves showed in most cases that the ear or brain had been injured. In only two instances was the operation sufficiently clean to fulfill the requirements of the experiments.

302 Journal of Comparative Neurology and Psychology.

is an observation which suggests the importance of the relation of vibration rate to the nature and condition of the transmitting structures. Many individuals failed to react to any moderately intense stimuli within a period of several hours after this operation. The sample series of reactions recorded in Table IV is sufficient comment upon the conclusiveness of the experiments so far as the value of these transmitting organs for socalled auditory reaction is in question.


The Influence of Sounds on Frogs without, Coi.umellae or Tympana.

Green Fro^

r. Operated

April 77,

igo4. Reaetions taken iininediately aftt

- operation .

Tactual stimulus alone

Auditory and tactual stimuli.

Position of shadow

Position of shadow



on scale

No. of pair

on scale

of movement


before after




the stimulus

the stimuli







4 mm.






90 M







90 M













90 IVI




90 M










90 M























Average for serie

3 1 4.0 mm. 1

36.8 mm.

3. After the cutting of the eighth nerves reactions to sounds were not obtained. Tests were made with green frogs and bull frogs, but only with the latter could reactions to tactual stimuli sufficiently great for the purposes of the investigation be obtained after the operation. Of fifteen frogs used only four reacted regularly to tactual stimuli after being operated, and these gave no sign of reaction to sounds. Typical of the results for those individuals which continued to be responsive to tactual stimuli after the operation is the following series, Table V. The eighth nerves of the bull frog used in the experiments of this series were cut May 17; two days later a series of twenty five pairs of reactions was obtained. Experiments were continued at intervals until June 14. The animal was apparently in good condition, the skin wound had healed by the latter date, but auditory reactions were wholly lacking.

Yerkes, The Sense of Hearing in Frogs.


The Influence of Sounds on Frog with Eighth Nerves Cut.


Bull Frog Operated May 17, igo4.

Date of

Amount of reaction to tactual stimulus

Amount of reaction to auditory and tactual stimuli

Number of reactions

greater (-(-)

less { \

Each result is the average for 25 reactions

equal (=)

May 19 May 20 May 21 May 24 June 14

4.16 mm. 19.24 15.04 13.08


4.04 mm. 18.20 14.28 13.28


9+. II —, 5 =• 9+, 14—, 2=.

10 -f-, ID , 5 =.

10 -j-. 6 — ,9 =. 7+, 12 -, 6=..


ges 12.02

II. 71

9+,io.6— , 5.4=.

Cutting of the eight Ji nerves j-enders the frog irresponsive to sonitds which markedly influejice the tactual reactions of the 7iornial animal. We may tJierefore conclude that the reactions zuith zvhich ive have dealt in this investigation are due to stimulation of certain sense organs of the ear, and that the use of the word hearing in comiection zvith them is appropriate.

VII. Summary and Conclusions.

1. Observation of frogs in their natural habitat shows that they are stimulated by sounds. The sense of hearing apparently serves rather as a warning sense which modifies reactions to other simultaneous or succeeding stimuli than as a control for definite auditory motor reactions.

2. Experimental tests prove that sounds modify the frog's reactions to visual and tactual stimuli. When the sound accompanies the visual or tactual stimulus it serves to reinforce the visual or tactual reaction, but when given alone it never causes a motor reaction.

3. The sound of an electric bell occurring simultaneously with a tactual stimulus markedly increases (reinforces) the leg reflex of green-, leopard- and bull-frogs. If the sound precedes the touch by i" it is without effect on the reaction ; if the interval is not longer than .35" it usually causes reinforcement, whereas for an interval of from .4" to .g" there is partial inhibition of reaction. According to its temporal relation to

304 Journal of Comparative Neurology and Psychology.

another stimulus, an auditory stimulus may either reinforce or inhibit the reaction appropriate to that stimulus. What may be called reinforcement-inhibition curves for auditory stimuli are presented in this paper.

4. The green frog responds to sounds made in the air whether the tympana be in the air or in water. There is some evidence that the influence of auditory stimuli is most marked when the drum is half submerged in water. The influence of sounds upon tactual reactions is evident when the frog is submerged in water to a depth of 4 cm.

5. Sounds varying in pitch from those of 50 to 10,000 vibrations per second effect the frog. The most striking results were obtained by the use of an electric bell with a metal gong. With this sound in connection with a weak tactual stimulus a maximum reaction of the leg may often be obtained even when either stimulus alone causes no perceivable reaction.

6. Sounds modify the reactions of the frog after tympana and columellae are removed. Cutting of the eighth cranial nerves causes disappearance of the influence of sound. It is clear then that the reactions to sounds are really auditory reactions and that the sense of hearing in the frog is fairly well developed, although there is little evidence of such a sense in the motor reactions of the animal.

7. Those portions of this investigation which were carried out in the spring months show marked influence of sounds for both males and females, whereas experiments made during the winter indicate a much diminished sensitiveness to auditory stimuli in both sexes, but especially in the male.


Contributions from the Zoological Laboratory of the University of Michigan, No.ioo.

With Six Figures in the Text.


I. Introduction. II. General Habits.

III. Reactions to Light.

I — General features of the phototactic response.

2 — The negative reaction.

3 — Head and swaying movements in negative phototaxis.

4 — The eftect of contact on phototaxis.

5 — The effect of temperature on phototaxis.

6 — Phototaxis leading to fatal results.

7 — Inhibition of phototaxis by other activities.

8 — The effect of hemisecting the brain.

9 — The effect of covering the anterior half of the eyes. 10 — The effect of covering the posterior half of the eyes. II — The effect of destroying or covering one eye.

12 — Reactions of specimens with only a small part of the lateral surface of one eye exposed. 13 — Phototaxis as modified by experience and habit. 14 — Formation of habits of turning.

IV. General Considerations on the Phototactic Response.

I. Introdtction.

In endeavoring to ascertain the way in which animals of various kinds orient themselves to the rays of light I have experimented with quite a large number of species in the hope of finding forms in which the exact mode of response would reveal itself. Animals vary greatly as regards both the definiteness of their reactions to light and the ease with which their movements can be followed. Among creatures of small size such as. the Copepoda, Cladocera and Ostracoda, it is almost impossible

3o6 Journal of Comparative Neurology and Psychology.

to observe the precise movements concerned in orientation, and in many larger forms the rapidity, irregularity, or indefiniteness of their light reactions renders the same difficulty almost equally great. In studying the reactions of animals to light we are naturally confronted with the question as to how far the movements involved are the result of choice, or something analogous thereto, and how far they may be explained as the result of reflex responses to photic stimuli. If they mainly fall into the latter category we are led to inquire just what these reflexes are and how they produce the particular kind of behavior observed.

It is a quite commonly accepted hypothesis that the phototactic reactions of organisms are effected by the action of light directly or indirectly upon the tension of muscles concerned in locomotion. In nearly all insects and in a large proportion of other arthropods this tension, if it exists, must be brought about through the central nervous system, since the opacity of the integument prevents any appreciable direct effect of light upon the musculature. In most arthropods phototactic impulses are set up by means of light entering the eyes, and not as in many lower forms through the stimulation of the integumental nerves ; this is shown by the fact that when the eyes are blackened over or destroyed responses to light no longer occur. In most animals it is not possible to observe any effect of light upon muscular tension, although there is considerable indirect evidence that such an effect is produced. As Radl^ has remarked, it is difficult to explain the fact that an insect with one eye blackened over moves about in a circle except on the assumption that light affects unequally the tension of the muscles on the two sides of the body. Such circus movements are comparable to those which take place in a vertebrate animal upon the destruction of the semicircular canals in one side of the head. After this operation there is produced a marked difference in the muscular tonus of the two sides of the body and, as a consequence, the animal, instead of going in a normal manner

'Untersuchungen iiber den Phototropismus der Tiere, 1903.

Holmes, The Reactions of Ranatra to Light. 307

veers continually toward the weaker side. A small difference in the muscular tension of the two sides of an insect body which would be sufficient to cause the creature to orient itself to the rays of light might not be patent to direct observation, especially if the movements are rapid or irregular, as they frequently are. There are several forms, however, in which the effect of light upon the muscular tone is quite clearly manifested, but none more so than in the common water scorpion, Ranatra f us ca. In many ways this species is admirably adapted for the study of phototaxis ; it is of large size, its long slender legs move in a slow and deliberate manner so that one can observe just how each action is performed ; it may be readily kept for a long time in the laboratory, shows no signs of fear when being experimented with, and reacts to light with a remarkable degree of precision. For an investigation of the modus operandi of the phototactic response Ranatra is probably not equalled by any other known form.

It is especially advantageous to study phototaxis in some such organism if we wish to ascertain how far the reflex theory of orientation will carry us. If orientation is the result of comparatively direct reflexes we are better able to determine their precise mode of action. If a more involved type of reaction occurs there is a better opportunity afforded for proving its existence, and, perhaps, ascertaining something of its nature. It does not follow that because we can construct a theory to account for orientation by means of direct reflexes that the process necessarily takes place in so simple a manner. Between the stimulus and the reaction there may be processes of a complicated nature whose existence is not ordinarily betrayed by any outward and visible sign. No one would consider a dog's following the scent of a rabbit a matter of simple chemotaxis. While it is not a process requiring conscious ratiocination, it is doubtless one involving psychic operation of considerable complexity. The possibility should be borne in mind that many of the tropisms of insects may be less simple and direct reactions than is commonly supposed. If a bee finds its way to its hive over miles of woods and fields guided by its memories of

3o8 Journal of Comparative Neurology and Psychology.

the various objects that come into its field of vision, it is certainly something more than a mere reflex machine. In organisms which are capable of a higher type of response we should at least be on our guard in attempting to explain their tropisms as due entirely to direct reflexes involuntarily performed in response to outer stimuli. The conduct of higher animals is guided in large measure by their likes and dislikes, however we may interpret this kind of behavior in physiological terms. Between such behavior and those tropisms which are the result of comparatively simple reflexes there are, no doubt, numerous intermediate kinds of conduct. It is not unreasonable to suppose that tropisms which in low forms are brought about by direct reflexes may in higher animals complicate into reactions of the pleasure-pain type while still preserving outwardly the appearance of a more mechanical mode of response. At the same time an element of direct reflex action may be retained, although closely associated with and capable of being modified by more complicated neural processes. A consideration of the experiments described in this paper will lead us, I think, to some such view.

II. General Habits.

Ranatras are generally found in ponds or streams among masses of vegetation where they lie quiet the greater part of the time. Although capable under certain conditions of manifesting considerable activity, these insects are usually sluggish in their movements. Their choice of habitat is probably determined, in great part at least, by their positive thigmotaxis, since they tend to insinuate themselves between objects which afford considerable contact stimuli. Their habit of coming together to form groups is a manifestation of the same tendency. When several individuals are placed in an aquarium they mass together when at rest to form a cluster in which they are often so closely aggregated and so tangled together that those which are near the center of the group experience much difficulty in disengaging themselves. In this way they may lie for n an almost motionless state.

Holmes, The Reactions of Ranatra to Light. 309

The general form and dull coloration of Ranatra tend to make it inconspicuous in its natural habitat, especially as it does not reveal its presence by its movements. When lying in the water the long breathing tube through which air is admitted to the body rests at the surface. The two parts of which it is composed occasionally approach and recede from each other, moving the air between them to and fro, an operation which doubtless assists in respiration. Air is prevented from escaping when the valves are separated, by the rows of hairs which line the margins of the concave inner faces of these structures.

Ranatra is carnivorous in habit, seizing its prey with its anterior raptorial limbs and holding it until it has sucked out its juices. It is quite destructive of fish eggs and frequently attacks and sucks the blood from young fishes. It is also reported to prey upon young tadpoles. De la Torre Bueno^ describes the method Ranatra employs in capturing prey as follows: "When a fly attracts its attention Ranatra very slowly, almost imperceptibly, moves its fore legs, with the knife-like tarsus away from the tibia, towards its prey. When the tibiae are almost, or quite, touching the victim the movement is so sudden and quick that one is aware of it only by seeing the prey seized. Sometimes its hold is not satisfactory, and then it will let go with one tarsus, get a firmer grip with that, and then do the same with the other. Once it has the fly securely held, Ranatra slowly approaches it to its extended beak, with which it seems to touch and feel until it finds a suitable spot, and proceeds to a leisurely meal." I have usually fed Ranatras during the winter on Notonectas, or back-swimmers, as these insects were easily obtained during this time of year. The Ranatras did not pursue the back-swimmers, but as soon as their attention was attracted to the prey they lay quietly in readiness for them with their anterior limbs prepared to quickly seize the small insects should they swim sufficiently near. If a Notonecta strikes against a Ranatra the latter makes a quick

'Notes on the Stridulation and Habits of Ranatra fiisca Pal. B., Canadian Entomologist, Vol. 35, p. 235, 1903.

310 Journal of Comparative Neurology and Psychology.

grab for it, and, if successful in seizing it, proceeds to suck out its blood in the manner described by de la Torro Bueno.

In locomotion, either by swimming or walking, the anterior limbs do not usually play a part ; they are held straight in front of the body and are employed only occasionally to aid in changing the direction of locomotion or to clamber over some obstruction. Out of the water Ranatra walks rather awkwardly. Its long slender second and third pairs of legs are articulated close together near the center of the body and the insect is frequently tilted over so that one extremity or the other strikes against the surface over which it walks. While Ranatra is capable of flight, it rarely if ever flies to lights at night as many other aquatic hemiptera do ; I have never seen any specimen around electric lights where other insects are found in abundance.

Ranatras pass the winter in the adult state. I have collected numerous specimens in a small stream north of Ann Arbor late in November, but on visiting the same locality during a thaw in January following, although a diligent search was made in their favorite habitat among aquatic plants and by digging in the mud in the sides and bottom of the stream, I did not obtain a single specimen, although Zaithas and water boatmen were found to be quite common. Possibly the Ranatras burrowed more deeply than I could dig with the apparatus employed, although the general form of the animal renders this supposition improbable. Egg laying occurs in the spring. The eggs are long and narrow and furnished at one end with a pair of filamentous processes which, according to Korschelt, have a respiratory function. In ovoposition the female inserts the eggs into the stems of aquatic plants, or even into wood, the filaments projecting from the exposed ends.

Ranatras make a feeble sound by rubbing the bases of the anterior legs against the lateral processes of the prothorax. When a Ranatra is picked up in the fingers one can feel a slight tremor when the animal stridulates, although the sound is so faint that it cannot be heard farther than a few inches from the ear. What use, if any, is made of this sound is uncertain.

Holmes, The Reactions of Ranatra to LigJit. 3 1 1

The method of sound production has been described by Torre BuENO. The statement of this writer that the stridulation of Ranatra was not previously described is not correct, for the subject was briefly treated of by Locy^ in 1884.

The instinct 6f feigning death, which is remarkably well developed in Ranatra, will be described in a subsequent paper.

III. Reactions to Light.

General Features of the Phototactic Response. When Ranatras are kept in a glass dish of water near a window they are usually to be found facing the light, often swimming towards it and repeatedly colliding with the side of the dish and clawing against the invisible barrier which blocks their course. These movements may be kept up, with intervals of rest, all day. When an artificial light is used the Ranatras may be caused to swim in any desired direction by placing the light in the proper position. In experiments with this species I have usually employed an ordinary 16 candle-power incandescent lamp attached to a flexible cord of wire which permitted it to be readily moved about at will. The work was carried on in a darkened room so that the specimens experimented with were exposed to light only from this source.

When Ranatras are taken out of water and laid on a table they generally feign death, and, while in that condition, they at first give no reaction to light. One may move the light about near them or hold it almost against their eyes without eliciting the least sign of a response. This apparent insensibility, gradually wears away, and after some minutes the movements of the light are followed by scarcely perceptible motions of the head. By passing the light back and forth laterally over the body the head may be caused to rotate laterally each time the position of the light is changed. These are the first movements that can be made to appear, and they grow more decided the longer the experiment is continued. A little later the animal may be made to respond by vertical head movements

'Anatomy and Physiology of the Family Nepidae, Am. Nat., 1884, p. 364.

312 Journal of Comparative Neurology and Psychology.

when the light is passed back and fourth over the long axis of the body. When the light is in front the head is bowed down and when it is passed behind the body the head is tilted upward. Both the lateral and vertical movements are such that they tend to place the upper surface of the head at right angles to the direction of the rays. The vertical movements, like the lateral ones, are at first slight, and increase in vigor the longer they are caused to continue. The animal performs these movements with machine-like regularity and precision and without showing the least activity in any other part of the body. If the light is moved around the body in a circle the head will follow it with a corresponding rotary motion. If the light is in front and to the right of the animal the head will be tilted over to the right and at the same time pointed downward in front ; or if the light is to the left and behind the animal the head will be tilted over to the left and raised up in front. For each position of the light there is a corresponding attitude of the head. The orientation of the head is remarkably precise. By carefully watching the tip of the beak with a lens as the light is moved it may be observed that a change of only two or three degrees in the direction of the rays produces a corresponding change in the direction in which the head is pointed.

After some time the movements of the head are accompanied by movements of the breathing tube. When the light is in front of the body the tube is lowered ; when it is carried behind the body the tube is raised. These movements are at first slight but they become more decided and more regular the longer they are continued. Lateral movements of the breathing tube in response to light do not occur.

The next movements to appear are those of the limbs. The animal, after a time, shows irregular twitchings of the leg muscles, and soon afterwards slowly and unsteadily raises itself upon its legs and stands as if undecided whether to walk away or subside again into a state of repose. If now the light is passed over the body from side to side the creature will perform swaying movements each time the position of the light is changed. If the light is on the right, the back of the insect is tilted

Holmes, 77/^ Reactions of Ranatra to Light.


over towards it, the right legs are flexed and the left ones extended (See fig. i). Pass the light over to the left side of

Fig. I. Attitude of Ranatra when the right side is toward the light.

the body and the back tilts over to the left, the left legs become flexed and the right ones extended. For a short time after coming out of its feint Ranatra will usually sway back and forth as the light is moved over it without attempting to walk, but soon it begins to follow the light, at first with slow and unsteady steps, but later with more and more vigor, until finally its efforts to go towards the light become almost desperate, and it becomes oblivious to everything else.

Besides the lateral swaying movements which Ranatra performs when light is passed over the body from side to side there are equally pronounced longitudinal swaying movements when the light is passed back and forth along the axis of the body. When the light is placed behind the animal the body is raised up in front and the head held high in the air. Now place the light in front, and immediately the body is lowered, and the head bowed down, the grovelling attitude contrasting almost

314 Journal of Comparative Neurology and Psychology.

comically with that assumed when the light is in the rear. Bymoving the light around the animal in a circle all combinations of lateral and longitudinal swaying movements may be produced, the body following the light by twisting about in a most curious fashion. With the light to the right and behind the animal the body is raised up in front, tilted over to the right, the legs on the right side flexed and those of the opposite side extended, the head turned to the right and the beak held high in the air. If the light is passed in front of the animal on the left side the head and body are both bowed down and tilted to the left, the left legs flexed and the right ones thrown in a state of extension. As with the head, so with the body, there is a certain attitude assumed for each particular position of the light.

^/[g'. 2. The lower figure represents the position of Ranatra when the light is behind the body. The upper figure represents the position assumed when the light comes from in front.

Ranatras may be made to follow the light in any direction in the most slavish manner. By keeping the light to one side and a little behind the middle of the body the animal may be kept wheeling about in one spot, often, however, falling over on one side in its eagerness to get around. One reason for its awkwardness in this case is that as the light comes from the

Holmes, TJie Reactions of Ranatra to Light. 315

rear the anterior part of the body is carried high in the air and the creature consequently easily loses its balance.

At any time after Ranatra has ceased to feign death it can be made to perform swaying movements in either direction, although these may be combined with efforts at locomotion. By carefully controlling the light, however, the tendency to locomotion may be largely checked. By moving the light around the animal in a circle the tendency to turn back when the light is behind may be made to balance the tendency to go forwards when it is in front, and the animal simply sways around in its tracks.

Light seems to dominate entirely this creature's behavior when the phototactic reactions are once started. It does not manifest any fear or awareness of any object in its environment save the light which it so strenuously seeks. Its excitement increases the longer it is operated with, and after a time it may be picked up without feigning death, or with only a momentary feint. Not content with walking as rapidly as possible towards the light, the insect begins to fly towards it, always doubling up its fore legs in a curious manner before spreading its wings. The wings are never used, however, until the creature has sought for some time to reach the light by the ordinary method of locomotion. Their employment marks the attainment of a high pitch of excitement in which the insect seems animated by an uncontrollable frenzy which lasts until it is checked by approaching exhaustion.

If a Ranatra is placed on its back it often has considerable difficulty in righting itself, and if near a light it will often walk towards it without turning over. Locomotion under these circumstances is effected mainly by the anterior legs which are flexed dorsally at the middle joint. These legs are not employed in ordinary locomotion, but when the creature is in an inverted position they are used with considerable dexterity. By bending the legs dorsally the anterior end of the body is elevated, and by the alternate movement of these appendages the insect walks along in a tolerably efficient manner. It will follow the light around in this way in any direction. When out

3i6 Journal of Comparative Neurology and Psychology.

of orientation it reaches over to one side with the anterior leg nearest the light and pulls the body over until it is parallel with the rays.

2. The Negative Reaction. While under ordinary circumstances Ranatra is positively phototactic it may in certain conditions become strongly negative. If Ranatras which have been kept for several hours in the dark are brought back again into the light they usually show a marked negative reaction, but this negative phototaxis is never so violent as the positive sometimes becomes. Instead of appearing to be the result of reflexes which are comparatively direct and involuntary, the behavior of Ranatra gives the impression of being caused by the desire to escape from a situation which is unpleasant. It is perhaps remotely analogous to the action of a man when, after having been for some time in the dark, he instinctively turns away from the sudden glare of a strong light. The behavior of specimens after having been kept in the dark may be illustrated by the following experiments :

Twenty-five Ranatras which had been in a dark room for twelve hours were placed, one at a time, in a glass trough through which light was passed from an incandescent lamp situated a foot from one end. Each specimen was placed in the water in the center of the trough at rivht angles to the direction of the rays and let go in that position. If the specimen swam to the negative end, and within one minute returned to the other end when the light was changed it was classed as negative; and a similar criterion was employed for the positive reaction. If a specimen swam to one end of the trough and did not return to the other end within one minute after the light was changed it was classed as doubtfully positive or negative according to which end it first reached. Seventeen of the specimens proved to be negative; one was positive; four were doubtfully negative and two doubtfully positive. Out of the twenty-five specimens employed only one was unmistakably positive in its reaction, and this one swam to the positive end of the trough upon change of the light several times in succession. One spectmen did not swim to either end for five minutes. The Ranatras were then left exposed to the light of an incandescent lamp placed three inches from the end of the trough for one hour and forty minutes. At the end of this time all of the specimens were positive, The light was then held eight feet from the trough and the positive reaction still continued.

One of the specimens which showed a very marked negative phototaxis was singled out. During the time the others were being experimented with it was exposed to the light, and when again studied (after about a half hour's exposure) it showed a very evident negative phototaxis, but not so strong as before. After a time the negative reaction became so faint that it was scarcely distinguishable.

Holmes, Tlie Reactions of Rajiatra to Light. ^ly

The specimen was then exposed to the strong light from a projection lantern. It showed at first a weak positive phototaxis which grew stronger the longer it was exposed, until it finally became almost violent. When exposed to the much weaker illumination from a i6 candle power lamp it still showed a marked positive phototaxis, but not so strong as when exposed to the light of a projection lantern.

At another time a Ranatra was placed in the trough which was exposed to the light of an incandescent lamp placed two feet from one end. The specimen immediately swam to the negative end of the trough. When the lamp was placed two feet from the other end the specimen quickly returned. The light was then changed repeatedly several times and each time the Ranatra would swim to the negative end of the trough, usually within four or five seconds. After a time its responses became slower and less definite. The light was then held close to the end of the trough and the responses became as prompt as before. It was changed from one end of the trough to the other forty-two times, and each time the insect within four or five seconds went to the negative end. Then it was left several minutes, after which its negative responses became less marked. When it was exposed to the strong light of a projection lantern it was still negative. It was then taken out of the water and laid on the table. In a few minutes it came out of its feint and showed the usual swaying movements when an incandescent lamp was moved near it, but it was very reluctant to walk. After 15 minutes of hesitation it became more active and showed an unmistakable positive reaction, and several times flew towards the light. When placed in the water again it still showed a strong positive phototaxis. The next morning (it was not in the meantime exposed to light, the room being darkened) it was markedly negative, and was driven from one end of the trough repeatedly by changing the position of the light. Then it was taken out of the water and placed on the table and an incandescent lamp was moved about near it to bring it out of its feint. It was a long time in awakening and for several minutes afterward it was disinclined to walk. When it did so it began to follow the light and soon became so excited that it would fly towards it repeatedly when four or five feet away, When put back into the water it was still positive and would follow the light in every direction. It was then exposed to the light from a projection lantern and became more strongly positive than ever, moving wildly towards the light even in the intense glare of the focus. Brought back to an incandescent lamp again it showed only a comparatively feeble response.

The negative reaction is associated with a condition of lowered phototonus. It is rarely shown except when the animal is in a condition of comparative sluggishness. When in great excitement, when its movements take place with quickness and vigor, Ranatra always shows a positive response. It never flies away from the light. Whenever it is wrought up sufficiently to use its wings, its reaction is invariably positive. The negative movements are slow and stealthy, often giving one the impression that the insect is attempting to sneak away unob

3i8 Journal of Comparative Neurology and Psychology.

served. The promptitude and decision of its negative movements may increase up to a certain point if the insect is kept close to the light, but when the movements begin to become vigorous there is a transition to the positive type of reaction. Often the advent of positive phototaxis is accompanied by a marked accession of energy as if a strong dormant propensity had suddenly been awakened.

The causes that produce the negative reaction are, as a rule, those which lead to diminished activity and excitement. Cold, exposure to darkness, the quieting effect of contact stimuli lead to a condition of lessened excitability and, perhaps as a result of this, to a negative reaction to light.

^. Head and Szvaying Movejnents in Negative Phototaxis. It would naturally be expected that the leg movements in negative specimens are the reverse of those in positive ones, and to a certain extent this is true. It was thought not improbable that the movements of the head would be reversed as well ; but it was found that in all cases in which Ranatras moved away from the light the head reflexes take place exactly as in individuals that are positive. Both in strong and in weak light, under a variety of different conditions the head reflexes are the same in kind regardless of the general sense of the response.

The swaying movements in negative phototaxis are readily observed in Ranatras that have been resting quietly in the dark for some time in a glass dish of water. If an incandescent light be brought near them they are at first irresponsive. By moving the light slowly about them the head reflexes are first induced. Then there may be very slow and at first scarcely perceptible movements of the legs, the second and third pairs being very lazily extended on the side towards the light, and flexed on the opposite side, the back being rolled over so that it more nearly faces the light. On placing the light on the opposite side of the body there is a very gradual extension of the legs previously flexed and a flexion of those previously extended, the body rolling over at the same time so that its dorsal surface lies more nearly at right angles to the rays. The animal may be made to repeat this performance many times in succession. After a

Holmes, TJie Reactions of Ranatra to Light. 319

time its movements become more vigorous and it turns to walk away from the light ; it can then be driven about in any direction at will.

The difference in the swaying movements of positive and negative specimens is mainly brought about by the different movements of the legs at the femoro-tibial joint. In negative specimens the legs on the side toward the light are extended at this joint while those on the opposite side are flexed ; in positive specimens the reverse relation occurs. The rolling of the body is the same in both cases so that the muscles extending between the legs and the body are similarly affected in both kinds of reaction. The same relation probably obtains with the muscles between the joints of the upper parts of the leg. The vertical swaying movements of the body which involve the employment of these muscles are the same in both kinds of response. When light is behind a negative specimen the anterior part of the body is held high in the air as the insect walks away. When the light is held in front of the insect the anterior end of the body is lowered as it is turned from the light.

Only a part of the organism is subject to a change in the sense of its phototactic response. The head reflexes and swaying movements of the body are always such as to bring their upper surfaces more nearly at right angles to the direction of the rays, whether the organism as a whole is going towards or away from the stimulus. No matter how strong or how weak the light, or whatever may be the condition of the anirrial, these responses, if made at all, always occur in the same way.

^. The Effect of Contact on Phototaxis. As shown by the following records of experiments, Ranatra may be made negatively phototactic by means of contact stimuli. In one experiment seven specimens that were swimming against the side of the dish towards the window were picked up by their breathing tubes and dropped back into the water. At first they remained quiet but soon showed a negative reaction, swimming vigorously against the side of the dish away from the light. In about twenty minutes all but two had become positive again. They were all picked up by the breathing tube a second time

320 Journal of Comparative Neurology and Psychology.

and dropped back into the water without touching any other part of the body. Soon all became negative without exception. After a few minutes they began to cross to the positive side of the dish one by one, and it was but a short time before every individual was positive. When I returned after an absence of an hour and a half all of the specimens were negative, although the light to which they were exposed had increased in intensity. Warm water was then added so as to bring the temperature of the medium up to 30° C. Four of the specimens soon became strongly positive. These were picked up by the breathing tube and dropped back into the water ; all became markedly negative. As specimens came over to the light side of the dish they were picked up and dropped as before, with the result, in almost every instance, of producing a marked temporary negative reaction. The next morning at 10 o'clock all of the specimens in this dish and in another that was beside it were swimming towards the light. They were all picked up and dropped back into the water when, without exception, they became negative. Soon they began to come over to the positive side of the dish and in about half an hour they were all positive again. They were all picked up and dropped a second time. All but two became negative. At 2:40 in the afternoon the specimens in each dish were aggregated into a dense bunch at the negative end. When stirred up some seemed positive and some negative, but their reactions were not decided. The temperature of the water in one dish was increased to 32° C. when about half of the specimens became positive in an unmistakable degree. When picked up and dropped into the water they quickly became negative. When they became positive again they were handled under the water without taking them out ; as soon as released they showed an unmistakable negative reaction.

I have tried handling positive specimens under water repeatedly. The effect is, in nearly all cases, to produce a change in the sense of the phototactic response.

The effect of contact and disturbance is very marked also on specimens while out of the water. This as well as other

Holmes, The Reactions of Ranatra to Light. 321

features of interest is shown in the following experiments performed upon one individual :

The specimen was taken out of water from a darkened room where it had been placed the day before and placed near a light on a table. Soon it awoke and began to turn away from the light. The light was moved around it in different directions and, although the insect seemed at first dazed and walked about with no very decided tendency to go either towards or away from the light, it soon began to show a more pronounced negative reaction. It was then picked up by the breathing tube and placed at right angles to the rays from an incandescent lamp four feet away. During eight successive trials in which the right and left sides were presented alternately to the light to eliminate any tendency to turn to a particular side that might be due to habit, it turned in each case away from the light. Its movements were at first slow and stealthy. At the ninth trial it turned slightly towards the light, but reached the edge of the table before it had turned very far. At the tenth and several subsequent trials it turned towa»ds the light and went up to it rather quickly. Its movements now became much more rapid. It was then placed ten feet from the light and still showed a positive response. It was then placed on the floor 22 feet away from the light, which was laid on the floor at the other end of the room. Although the light to which it was exposed was relatively very dim, the insect traveled to the light across the whole length of the room in nearly a straight line. When brought near the light the insect became more and more strongly phototactic and in about twenty minutes its eff^orts to reach the light became almost frantic. After a time it became apparently exhausted and settled down to rest.

It was then immersed in water and laid down on the table. Its movements were very sluggish and its responses to light slow. When placed at right angles to the rays it would slowly and stealthily creep away. It did this eight times in succession when the right and left sides were alternately placed towards the light. At the ninth and several subsequent trials it went towards the light. Then it was caused to follow the light about for a few minutes and soon it became quite excited. It was picked up and stroked but it could not be induced to feign death and as soon as released it made for the light, which was four feet away. It was then held in water for several seconds, but as soon as liberated it showed an unmistakable though not very strong positive reaction. It went up to the light, touched the bulb with its anterior legs, jerked back quickly as if burned, then stopped for a moment and walked away from the light. As the light was moved about, the insect would flee from it repeatedly as if it feared a repetition of its disagreeable experience. Soon, however, its responses became weak and indefinite ; its movements were sluggish. After a time it showed a positive reaction and began to follow the light all around the table. A little later it became much excited and flew towards the light repeatedly. The light was placed in a cylinder of water and the insect struggled to go towards it for ten minutes. It was then held in cool water for a short time and placed again on the table. After this experience it showed a marked negative response. The light was moved about the insect in various directions so as to keep it close to the body for about twenty minutes when positive phototaxis was again induced. When its reaction became very strong, the insect was picked up and stroked, but it feigned

322 Journal of Comparative Neurology and Psychology.

death only momentarily and then resumed its efforts to go to the light.

A repetition of the experiment was followed by essentially the same result. When dipped in water again it showed a faint negative reaction. It would go towards the light when it was placed in front of the body or away from it when it was placed behind. Its general behavior was sluggish and it would perform only scarcely perceptible swaying movements when the light was moved over its body. Soon its negative reaction became more pronounced and it would turn away from the light every time it was placed to one side of the body. The insect was then placed under a bell jar near the light and upon my return after a two hours absence it was markedly positive. When dipped in water it showed a suggestion of a negative reaction, and for some time was apparently indifferent to the light. Repeated dippings failed to make the specimen more negative, and after a time a sluggish positive response began to appear. The light was then moved around it and finally the creature became very violent in the eagerness of its response and flew towards the light several times. When it was dipped in water it became sluggish. When placed at right angles to the rays it turned away from the light and started to do so a second time, but turned towards the light and went up to it. The same experiment was repeated three times in succession, and each time the insect turned at first away from the light and then towards it before having proceeded more than a few inches. After thirty minutes of positive reactions it was dipped in the water again. It went slowly towards the light but passed by it, and in several subsequent trials went away from the light ; soon, however, it became weakly positive and in a short time its positive response was strong. After three hours of exposure to strong light it was still positive. When dipped into the water it showed at first a faint negative response but soon turned and went up to the light. Subsequent dippings failed to evoke a negative response.

It is not the effect of water in itself that changes the response, but the experience of being dipped in water. If the Ranatras are allowed to remain in the water they soon show a positive reaction. If then they are lifted out and put on the table they almost at once become negative if they are not thrown first into the death feint. Curiously enough, dipping into water is more effective in changing the sense of the response than handling or stroking the specimens in the air. Specimens which have been handled so much that they no longer respond to that treatment either by feigning death or by showing a negative response to light may usually be rendered negative after dipping into water.

As handling positively phototactic specimens usually causes their reaction to become negative, unless the experiment is repeated too often, it is probable that the change produced by dipping them in water is due to the influence of contact stimu

Holmes, TJie Reactions of Ranatra to Light. 323

li. It takes place independently of differences of temperature, and it cannot be satisfactorily accounted for by attributing it to chemical or osmotic changes produced by the surrounding media. Anything which makes towards the peculiar nervous condition which accompanies the death feint tends to produce the negative response. Dipping Ranatras into water usually throws them into condition of quiet when other influences fail, and the negative reaction is doubtless a result of the nervous state thus brought about. When a Ranatra is either placed in the water or removed from it, every portion of the surface of the insect is stimulated, and, although the stimulus upon no part is strong, the general effect may well be considerable. That the general integument is very sensitive is indicated by the fact that decapitated specimens often respond very strongly to the slightest breath of air.

5. The Effect of Temperature on Photoiaxis. Raising the temperature tends to accentuate the positive phototaxis in Ranatra and lowering it tends to produce the negative reaction. In several experiments two dishes containing Ranatras were set before a window so as to receive the same amount of light. As the specimens had been previously kept in the dark, they showed a negative reaction. Into one dish warm water was poured raising the temperature from about 20° C to nearly 30° C. In a few minutes the specimens in the warmer dish became positive, the ones in the cool water still showing a negative phototaxis. Ranatras transferred to the cooler dish soon became negative while those which were picked up in the same way and dropped back into the warm water from which they were taken soon resumed their positive reaction. On the other hand, transferring negative specimens from cool to warm water produced in a short time a positive response. In cool water there is a marked tendency to form a dense cluster in the negative end of the dish. In warmer water the insects become more active and the groups are more apt to be broken up. Water at a temperature of 30° C usually stimulates them to very energetic movements. That negative phototaxis in Ranatra is induced by a reduction of temperature affords an illustration of

324 Journal of Comparative Neurology and Psychology.

the general fact that circumstances which reduce the excitability of the insect tend to produce the negative reaction.

6. Phototaxis Leading to Fatal Results. Wishing to ascertain if Ranatra would continue to be positively phototactic if it were thereby led into a situation which exposes it to stimuli having injurious, if not fatal, effects I performed the following experiment.

A strong arc lamp was placed on a table so that the focus or space between the carbons was about five inches from the top, this elevation being chosen so that the specimen could not walk directly into the luminous arc although it was free to move about beneath it. As the lamp gave out a large amount of heat the insect in approaching the focus would he brought into a region sufficiently heated to produce a fatal effect if it remained there long. A Ranatra placed on the table, moved toward the light until it came under the focus, then starting off again, only quickly to return. It soon became wildly excited and made repeated dashes toward the light. Several times it flew towards it, but luckily escaped coming between the carbons. After a time it gave signs of being overcome with the heat, but whenever removed from the light it would quickly return. Its movements became weaker, although its efforts to go to the light were no less persistent. It became unable to raise its body off the top of the table as it walked, but used its legs to slide its body towards the light, and it would quickly re-orient itself when placed obliquely to the rays. Even its very last piteous efforts were devoted to pushing its body a little nearer the light. When no longer able to move it was placed in cool water, but it did not revive.

When a moth flies into a flame it is probably because it does not have time to check or change the course of its flight after it has drawn near enough to experience the injurious effects of the heat. The suicidal conduct of Ranatra, however, cannot be accounted for in this way. The movements of the insect are slow and deliberate enough, especially when it becomes weakened, so that it need not be carried by its momentum into a region from which it would otherwise flee. It reacts positively at every step, even when nearly overcome by the heat. Essentially the same phenomenon is seen in Talorchestias which when exposed to direct sunlight, keep jumping towards the light until the heat overcomes them and they die.

7. Inhibitiofi of Phototactic Responses by other Activities. The phototactic responses of Ranatra which usually occur with such regularity and precision are sometimes checked when the insect is engaged in performing some other function. Specimens that have been following the lieht for some time often

Holmes, TJie Reactions of Ranatra to Light. 325

stop to rub their eyes with their first pair of legs, using their claws in what seems to be an effort to scrape off some foreign object from the cornea. If a light is moved over an insect when it is engaged in this operation the swaying movements of the body will no longer be performed. The head reflexes are also sometimes inhibited, but usually the head can be kept moving about at the same time the insect is rather ineffectually attempting to rub its eyes The swaying movements are likewise inhibited when Ranatra stops to rub its wings or any other part of the body. These actions may be caused by daubing asphalt varnish upon the insect, when efforts are made to get rid of the offending substance. As soon as the cleaning movements are over the insect promptly reacts to light as before. It generally makes no attempt to do two things at once.

The phototactic response may also be inhibited by efforts to obtain food. Ranatras which are swimming towards the light can often be caused to discontinue their phototactic efforts if several small insects are placed near them. If the phototactic activities are very lively and vigorous it is more difficult to divert the attention of the insect to the capture of prey. When attention is once directed to seizing the smaller insects the light is disregarded. When the prey has once been captured and the Ranatra is engaged in sucking out its juices little attention is paid to the light. The repast being finished the insect may resume its positive response.

- Efforts to go towards the light are frequently inhibited by contact stimuli. When several individuals are put into a dish of water near a window they commonly cease, after a time, to swim towards the light and form a dense cluster in which they lie at all possible angles to the direction of the rays. If now the individuals are placed in separate dishes they soon show a positive phototaxis. When placed together again they quickly form a group as before. Contact stimuli not only inhibit positive phototaxis but they produce a negative reaction as we have already seen : the latter tendency however, is often held in check by the same cause by which it is brought about.

Phototactic activities may also be checked by the sudden

326 Journal of Comparative Neurology and Psychology.

appearance of a large object in the field of vision. Ranatras that are swimming towards the light generally check their movements and lie perfectly quiet for some minutes as soon as they perceive one's approach, but if no movement is made near them they soon resume their phototactic activities. This inhibition of movement recalls that which in higher forms is often brought about by fear, but of the usual manifestation of fear in the efforts to escape from enemies by flight Ranatra evinces no sign.

Control in Ranatra probably amounts to nothing more than supplanting one instinctive tendency by another. The reactions of the insect to light seem to take place inevitably unless some circumstance calls into play some other equally stereotyped form of instinctive response.

8. The Effect of Hemisecting the Brain. Cutting the brain of Ranatra through the middle was accomplished by means of a fine needle ground down to a sharp edge. Only a small opening need be made, and there does not follow the profuse bleeding which results from making larger incisions. After hemisection of the brain specimens are easily kept alive for several days. Their behavior may be illustrated by the following records of experiments :

The brain was hemisected in three specimens^at 8:45 A. M. In one minute or less they came out of their death feint; in five minutes they were picked up and stroked and then laid on the table. None feigned death for more than two or three seconds. At 9:25 they were still very restless and none of them paid the least attention to the light that was held near them. They all performed circus movements to a greater or less extent, due, doubtless, to the fact that the brain was not cut exactly through the middle. At 11 A. M., one flew out of the dish . They were all picked up aud stroked again, but none feigned for more than a few seconds. When the light was moved about near them they would show no head reflexes or other decided response, although the light seemed to stimulate them in an indefinite way. During the next two or three days they continued the same restless movements and could be induced to feign death only for a few seconds. Definite responses to light failed to return.

The brain in three other specimens was hemisected at 12 M. They all soon came out of the death feint that was induced by handling them during the operation. At i:20 P. M., when they were again observed, they were very restless. When picked up, stroked, and laid on the table they feigned death for one minute, four minutes, and ten minutes respectively. The next day they could not be induced to feign for more than a minute and they showed no definite response to light. They were also tried on each of the three following days and their be

Holmes, The Reactions of Ranatra to Light. 327

havior was essentially the same. Two specimens seemed to respond when the light was moved near them but when the rays were passed through water before reaching them no response could be evoked. It is probable, therefore, that their movements were the result of stimulation by heat. Ranatras with the brain cut through the middle, like those with the brain removed, are very sensitive to all sorts of stimuli, and they are set into action by causes which would produce no manifest effect in a normal individual.

Bethe^ found that hemisecting the brain of Carcimis caused the phototaxis of the animal to disappear, although most of its other responses took place in a normal manner. I have found the same in several species of insects^. In the amphipod TalorcJiestia longicornis, which has a remarkably strong positive phototaxis, hemisection of the brain is followed by a complete loss of the power of orientation. Sensitiveness to light, however, is not entirely destroyed. If when a specimen is resting quietly in a shaded spot a beam of light filtered through an alum cell is thrown upon its eyes it usually responds by a few irregular movements. Since a large part of the fibers of the optic nerves cross in the brain, hemisection of this organ cuts off the main path of the impulses concerned in orientation to light.

g. The Effect of Covering the Anterior Half of the Eyes. Ranatras with the anterior surface of both eyes blackened over walk with the head strongly upturned and the anterior end of the body high in the air. Sometimes they stand nearly vertically, and several times I have seen them fall directly over backwards. Even when going towards the light the anterior part of the body is elevated, but not so much so as when the light is held above or behind the insect. When the light is moved backward and forward above the insect the body sways to and fro, and the head responds with the usual vertical reflexes. When the light is behind the insect the head and front part of the body are much elevated ; if now the light is carried to the front the creature bows down only for a short distance instead of assuming the grovelling attitude of a normal individual in the same situation.

^Archivf. mik. Anat. 1897, 50, 617. ■Am. Jour. Physiol. 1902, 5, 211.

328 Journal of Comparative Neurology and Psychology.

When Ranatra is compelled to walk on a glass plate while the light is held beneath the body it still carries the head and anterior part of the body high in the air. The head and body are lowered somewhat if the light is placed below and in front of the insect, but not nearly so much so as in a normal specimen. If the light is passed to the rear beneath the glass plate the head and front part of the body are raised up. Lateral head and body movements are performed in the usual way when the light is passed transversely over the posterior part of the body, but as the light is carried forwards these movements become less marked. Similar results are obtained if the light is moved beneath an insect which is placed on a glass plate ; the responses become less evident and precise as the light is carried in front of the body.

Ranatras with the anterior surface of their eyes blackened over are still able to follow the light when it is in front of them, but their movements are hesitating and their orientation inaccurate. This is a very natural result since the insects must be guided by the light which enters the posterior sides of the eyes.

10. The Effect of Covenng tJie Posterior Half of the Eyes. Blackening over the posterior surfaces of both eyes produces effects opposite to those observed when the anterior surfaces are blackened over. The insect walks with the anterior part of the body lowered and the head inclined slightly downward. If a light is held behind the insect the head and anterior part of the body are elevated, but not so much so as in a normal individual. When the light is in front the body is lowered anteriorly and the head bowed down. The same effects are produced when the Ranatra is placed on a glass plate and the light moved beneath the body. The light is followed very readily when it is kept in front of the body ; if, however, the light is moved to the rear the insect experiences considerable difficulty in orienting itself and frequently travels for some distance directly away from the light before finally turning around. Lateral movements of the head and body are readily brought about when the light is moved from side to side in front of the insect, but they become less pronounced when it is carried further back.

Holmes, The Reactions of Ranatra to Light. 329

//. The Effect of Destroying or Covering One Eye. If one eye of Ranatra is blackened over or destroyed the insect in most cases no longer walks in a straight line but performs more or less decided circus movements towards the normal side. Under the stimulus of light the insect assumes a peculiar attitude ; the body leans over towards the normal side and the head is tilted over in the same direction. When a light is held opposite the normal eye the insect leans over towards it, and holds the legs nearest the light in a flexed condition while those on the other side of the body are extended. When the light is passed over the body transversely swaying movements are performed as long as the light is opposite the normal eye, but when it is passed to the other side of the body the insect sways back only to the middle position or slightly farther. The longitudinal swaying movements of the body and vertical movements of the head are performed when light is moved forward and backward above the insect. These movements become less decided, however, as the light is held over towards the blind side.

Fig. J. Head of Ranatra fusca showing the almost stalked condition of the eyes.

If a light is held directly in front of a Ranatra with one eye blackened over the insect frequently does not travel straight towards it, but veers over towards the normal side until it comes to go in a quite different direction from that in which it

330 Journal of Comparative Neurology and Psychology.

started out. In one experiment a Ranatra with the the right eye blackened over was placed at a distance of four feet from a l6 candle power lamp. In all of the twelve trials that were made it started on a nearly straight course towards the light as it was originally pointed. In all cases it veered to the left and naturally diverged from the straight path more and more the farther it proceded. When a little more than half the distance to the light had been covered it corrected its course and proceeded towards the light again in a more direct path. This took place by two methods and usually occurred when the normal eye was looking away from the light. In this situation the insect would frequently stand for some time as if undecided which way to turn. In four instances it turned towards the right and proceeded again in the direction of the light. In eight cases it continued turning to the left until it had gone completely around in a circle after which it went up to the light.

Specimens with one eye blackened over present, however, marked individual differences of behavior. In some cases the insect walks toward the light in a nearly straight line. Ranatras which were taken late in the fall or early in the winter in most cases went nearly straight to the light while those experimented with in the early fall usually performed circus movements. This difference may have been due to accidental individual differences in the specimens, or it may have been the result of differences in age, the older Ranatras being better able to correct their course than the younger ones. The tendency to turn away from the blind side is manifested to a greater or less extent in all cases, but in many specimens it does not go so far as to produce decided circus movements.

If the whole of one eye and all but the posterior surface of the other is blackened over the insect in many cases is still capable of following a nearly straight path towards the light, A tendency to perform circus movements is more or less manifest and many specimens will go around in a circle repeatedly, especially when the light is near them. When all but the posterior surface of one eye is blackened over the insect is able to guide itself to the light nearly as well as when one eye is entire

Holmes, TJie Reactions of Ranatra to Light. 331

\y exposed. It may even follow the light by turning towards the blind side as the following experiment shows :

The right eye and all but a small part of the posterior surface of the left was thickly blackened over. When the specimen came out of its death feint it walked nearly straight to the light. The light was then moved over toward the right side of its path several times, and the course of the insect was changed so as to continue going towards it. The light was then held to the right and behind the body and the insect circled around to the right. By keeping the light in the proper position the insect could be caused to keep circling around to the right in the same spot. The head and body in these movements were tilted over slightly to the right, but not so strongly as they are inclined to the left when the light is held on that side. The insect would perform circus movements to the left more readily than in the opposite direction, but it would make sharp turns to the right whenever it was necessary to maintain a position of orientation. When the light was placed behind the body the insect would usually turn around to the left to reach it. If, however, it was placed behind the body and a little to the right the specimen would turn around towards the right side.

12. Reactio7is of Specimens with only Smallli Part of the Lateral Surface of One Eye Exposed. The right eye of a specimen was completely blackened over and all but a small part of the lateral surface of the left.

The insect was at first sluggish. When the light was held on the left side the head and body were tilted over towards it, and the specimen turned slowly to the left. When the light was placed two feet ahead of the specimen it moved very slowly and with the greatest hesitation, turning this way and that as if seeking to get its bearings, but going, nevertheless, nearly straight to the light. Although a decided tendency to make circus movements to the left was manifest, deviations in that direction were corrected by a direct turn towards the right side. When the light was held on the right side of the body the insect seemed confused and undecided which way to turn. When the light was held to the front and to the right so that the rays struck the body at an angle of about 45° the insect would not turn towards the light but moved about aimlessly until the left eye was presented to the light when it would turn to the left and go up to it. By keeping the light nearly in front of the body the insect would follow it all around the table. The experiment was then tried of holding the light farther over towards the right. The insect could then be caused to keep turning slightly towards the right in order to follow it. When the light was held still further to the right the insect still followed it. After a while it would follow it when held in a position to which it would not turn at the beginning of the experiment. After having got into the way of turning to the right side it would turn in that direction much more abruptly than at first.

A second specimen was treated in the same way, but it showed at first a pronounced negative reaction. It would pay little attention to the light except when it was held opposite the left eye when the insect would move away from it sidewise, much like a crab, by extending the legs on the left side and flexing

332 Journal of Comparative Neurology and Psychology.

those on the right. When the light was held on the blind side the insect showed no inclination to turn away from it. The movements of the insect were slow, but, after about twenty minutes, it suddenly began to turn vigorously to the light and became much more active, following the lamp around with eagerness and rapidity. In several trials in which it was placed three feet away from the lamp it went towards it in nearly a straight line, correcting its deviations from a straight course, whether to the right or the left, by the appropriate turn. It soon became quite excited and several times attempted to fly. It would follow the light when it was in front and to the right side by turning directly to the right. If, when the light was reached, it was changed to some other part of the table the insect would turn around and make for it in nearly a straight line. After moving about in this way for some time it settled down so as to rest its body against the top of the table. The light was then placed near the specimen, but it paid no attention to it until it was moved opposite the left eye, when it began to move slowly away from it by crawling sideways. The whole character of the insect's behavior now became suddenly changed. It became sluggish and stealthy in its actions instead of vigorous and excited. It would respond to the light by moving sidewise instead of straight ahead, and only when the light was placed in a certain posi. tion. After moving the light around the specimen its negative reaction became more decided and then suddenly changed to positive. It would follow the light around either to the right or to the left, but if the light were held too far to the right side it would not turn directly towards it. The specimen had several other fits of reacting negatively when its conduct was essentially as described above. It was caused to become positive each time by keeping the light near the left eye. Its negative reaction would in each case grow more vigorous and then suddenly give way to the positive response. The appearance of the positive response seemed to indicate the attainment of a certain degree of excitement.

The differences between the negative and positive responses of the specimens experimented with are very striking. The peculiarites of the negative reaction are probably due to the fact that only when the light shines directly into the left eye is the stimulus strong enough to produce action. When out of the direct glare of the light the insect is content, and it gets away by the most direct method. In the positive reaction the insect is attempting to get more of the stimulus ; it is in a state of increased phototonus, and each movement that brings it nearer the light tends to increase its activity.

/^. Phototaxis as Modified by Experience and Habit. That Ranatra is able to modify its reaction to light as the result of experience was shown by several experiments. The attempt was made to determine if Ranatras which perform circus movements when one eye is blackened over would in time come to

Holmes, The Reactions of Ranatra to Light.


travel to the light in a more direct course. A Ranatra with its right eye blackened over was placed on the top of a table three feet away from an incandescent lamp. Each time the specimen reached the lamp it was picked up by its breathing tube and placed in its original position facing the light. The course over which it traveled was in each case followed by a pencil. The paths taken in successive trials are represented in the accompanying diagram (Fig, 4).

Fig. 4. Paths taken by a Ranatra in successive trips to the light. The position of the lamp is indicated by a dot at the upper side of each path. The cross lines indicate places where the insect took wing. When it flew directly to the light its course is represented by a dotted line. At A the experiment was discontinued for fifty minutes, after which the remaining tracings were taken.

The specimen showed a strong tendency to fly as it drew near the light, and after the tenth trial its wing covers were stuck down by asphalt varnish. After a few ineffectual attempts to raise the wings the insect gave up trying to use them and walked to the light more quickly and with fewer pauses. The

334 Journal of Comparative Neurology and Psychology.

points on the path where the specimen took flight are indicated by a short cross Hne, and the position of the lamp is represented by a dot.

In the first trial the insect veered over constantly to the left, passed by the lamp and went off from the table before it turned around. In the following trials a marked tendency to turn to the left is also shown ; frequently the insect makes one or more complete circus movements to the left before reaching the light. At the eleventh trial its course is corrected for the first time by a turn to the right side, but, instead of going straight up to the light, it performed a complete circus movement to the left before reaching it. The next time the course was corrected by a sharp turn to the right and the circus movement was dispensed with. At the next trial the course was corrected in the same way, and at the fourteenth attempt the insect deviated only slightly to the left side and then turned to the right to reach the lamp. In the following ten trials it reached the light by a nearly straight path. Whenever it began to turn away from the light to the left it corrected its course by a direct turn in the opposite direction instead of going around in a complete circle as at first. During the latter half of the experiment the insect became more excited and walked to the light more rapidly and with less hesitation and wavering. The experiment was discontinued at 12:10 P. M. and resumed at i P. M. The insect had not forgotten in the meantime how to reach the light by the most direct means. In eleven trials it went to the light in nearly a straight course, as is shown by the second series of tracings.

The specimen was then placed with its left side towards the light. In each of its successive trials it turned sharply to the left and went up to the lamp. It was then placed with its right or blind side towards the light. In the first two trials it turned sharply towards the light and went up to it. In the third trial it first turned towards the light, but soon veered over to the left and went away from the light instead of towards it. When placed again with its left side to the light it no longer went towards it as before. It manifested a decided negative phototaxis in fifteen successive trials. When placed facing the light, as in the first experiment, it turned sharply to the left and went away from it for three times in succession. Then a fit of positive phototaxis began to manifest itself again. In the fourth and fifth trials it reached the light by an irregular course. After this it went nearly straight to the light fourteen times. It was now allowed to rest for an hour and a half, when it was experimented with again. At the first trial it turned to the left, but corrected its course by a turn in the opposite direction and reached the lamp. In the second trial it went to the lamp in nearly a straight line. At the third attempt, after it had gone nearly to the light, in an almost straight line it turned away from it and went off in another direction. A fit of negative phototaxis now began to assert itself and in the next thirty trials it turned to the left and went away from the light. It vras now nearly exhausted and could scarcely raise its body from the table. When near the light it would move in an excited manner in an effort to get away from it, but when it reached a darker region its movements became much slower, and it would soon come to rest. The experiments were then discontinued until the following day, when the specimen proved to be still strongly negative. After fifteen trials wera made, in which it turned away from the light, except during a short fit of

Holmes, The Reactions of Ranatra to Light. 335

positive phototaxis, when it turned twice towards it; the specimen was put back into the water. At one o'clock on the day following, it was taken out again. It still showed a decided negative reaction and turned to the left and went away from the light twenty-five times in succession. In nearly all cases after it had turned its back upon the light it went away from it in nearly a straight line.

After a rest of an hour and a half the asphalt varnish was removed from its eye. When placed on the table, three feet from the light, the specimen went towards the light and passed by it, but it turned around before proceeding very far and went up to it. After this it went to the light eleven times in nearly a straight line. The left eye was then blackened over. In the first two trials the insect veered slightly to the left, but reached the lamp by a turn to the right. In the third trial it turned sharply to the right and went off from the table. In the fourth trial it went straight to the lamp. After this another fit of negative phototaxis apparently seized it, for it turned from the light and walked away from it in numerous trials. It then became fatigued and the experiment was discontinued. It is worthy of note that while the insect, after one eye was first painted over, came to take a direct course to the light only after several trials, it went to the light, when the opposite eye was painted over in nearly a straight line at the first attempt. Unfortunately the fit of negative phototaxis which supervened prevented observation on this point from being carried out as fully as would be desirable. Owing probably to fatigue or frequent handling, the specimen became more prone to react negatively the longer it was worked with, until it no longer served the purpose of the experiment.

Other experiments showed a similar straightening of the course after a number of trials. As experiments on this subject were undertaken late in the season, most of the inviduals worked with went nearly straight to the light at the first trial. In one case a specimen went nearly straight to the light six times in succession ; it was then taken up and its right eye given another coat of varnish to make sure that it was entirely covered. When put down again it seemed confused and performed a number of circus movements to the left. After several attempts, however, it came to travel to the light in nearly a straight line.

Similar experiments were tried on Notonecta with much the same results. These insects when placed on a table near a lamp travel towards the light with much eagerness. If one eye is blackened over they perform circus movements towards the normal side. After quite a number of trials, however, the Notonectas come to take a nearly straight course to the light. When they become deflected from their course they orient themselves by a direct turn in the appropriate direction. If the

336 Journal of Comparative Neurology and Psychology.

lamp is kept over towards the blind side they may be made to perform circus movernents in a direction opposite to that in which they would otherwise tend to go.


J^i^. J. Tracings of successive paths towards the light taken by a Notanecta, which had the left eye blackened over. After a number of other trials the specimen came to travel towards the light in nearly a straight line.

i^. Fortnation of Habits of Turning. In working with Ranatras I have several times noticed what appeared to be a tendency to fall into habits of action, and I was naturally led to test the matter further by experiment. A Ranatra with its right eye blackened over was placed so that its head pointed directly away from a lamp three feet behind its body. As was to be expected, the Ranatra turned around to the left and went to the light. It was then picked up and placed in the same position ten successive times. In each trial it turned to the left and went up to the light as at first. In these and all subsequent trials the insect, when picked up, was turned alternately to the right and to the left before it was put down again. In this way the possibility that the results might be due A to compensatory motions was eliminated. f^g- 6. After the first ten trials the insect was placed

obliquely (at an angle of 45°) to the rays with its right side towards the light (position B in Fig. 6). In each of the ten trials that were made the insect turned towards the right or bUnd side, and went to the light. It was then placed

C «

Holmes, The Reactiojis of Ranatra to Light. 337

at right angles to the rays, the Hght falHng on its right side as before (position C). In each of the ten trials made it turned towards the right as in the previous experiment. Then it was placed with its head pointing obliquely away from the light (position D), when it still turned to the right in each of ten successive trials. Finally, it was put back into its original position, A, with its head pointing directly away from the light. Instead of turning to the left, as it did at first, it turned to the right as in the experiment immediately preceding, and it repeated the performance in the same way ten times in succession. It was then placed in a slightly oblique position so that its left side was exposed to the light. Twice it turned to the right as before, but in the third and several subsequent trials it turned to the left. It would then continue to do so when placed back in its original position, or even with its right side slightly turned towards the light.

The experiments afford unmistakable evidence that when Ranatra has turned in any direction once it tends, ipso facto, to turn in the same direction a second time. The habit thus formed may even overcome the tendency to turn towards a particular side which is caused by blinding one eye.

IV. General Considerations on the Phototactic Response.

The behavior of Ranatra in relation to light is, I believe, not without interest in relation to the general theory of the phototactic response. Many features of the phototaxis of this insect seem to afford strong support to the reflex theory. The perfectly definite and regular movements of the head in response to changes in the position of the light and the fact that these movements take place under all conditions in exactly the same way indicate that they are as machine-like as the most devoted partisan of the reflex theory of tropisms could wish. There is no evidence of choice, properly so called, in the performance of these actions. It is true that during the death feint they are no longer performed, but this affords no ground for regarding them as in any sense voluntary, since, as is well known, undoubted

338 Journal of Comparative Neurology and Psychology.

reflexes are no longer performed in certain conditions of the nervous system.

The swaying movements of the body in response to changes in the position of the Hght seem almost as mechanical as the head reflexes. They are more easily inhibited, however, since they are often checked by efforts to rub the eyes, to clean the body, or to raise the wings after they have been fastened down, but they are immediately resumed as soon as the energies of the insect are no longer diverted to other actions. The swaying movements of the body are the result of a strong and definitely directed tendency to action which, when the insect is in a certain condition of tonus, it seems powerless to control. It is only when other instinctive responses are brought into play that the insect is able to overcome its phototactic activities.

The swaying movements of the insect vary, however, according as it reacts to light in a positive or a negative manner, but there is little evidence of choice as regards which mode of response is followed. In one nervous condition the insect is always negative ; in another condition it is always positive. A Ranatra that is lethargic and sluggish reacts to light in a negative manner, while one that is highly wrought up and excited is always strongly positive. In intermediate states of excitement there is often a hesitation between the two modes of response, and the insect may exhibit alternate fits of positive and negative phototaxis of short duration. The power of one instinctive response to inhibit or overcome another may be considered as the first step towards the voluntary control. Choice as manifested by Ranatra is determined by which of two instinctive tendencies to action gains ascendency.

The mechanical nature of the phototactic response is further evinced by the effect of blackening over different parts of the eyes. When the posterior sides receive light the legs are extended and the anterior part of the head and body are held high in the air. When only the anterior sides receive light the head and body are bowed down. If light comes in only at the side of one eye the legs are flexed on the side towards the Hght and extended on the side away from it. It would almost seem as

Holmes, The Reactions of Ranaira to Light. 339

if different areas of the compound eyes have special connection with particular sets of muscles such that when a given area is stimulated the muscles are set in action which bring about a particular attitude of the body.

We should bear in mind the possibility of interpreting the phototaxis of Ranatra as due to the fact that the light is sought because the insect derives from it an agreeable stimulus. We might regard the creature as so constituted that it derives pleasure from light and reacts toward it so as to get as much of this stimulus as psssible. Its behavior would then come under the head of what Baldwin calls the " circular reaction." The movements of the head and body would be interpreted, according to this conception, as efforts to place the body in such a position that it can receive more of the desired stimulus. Whatever the position of the light Ranatra reacts towards it so as to place the upper side of the head more nearly at right angles to the direction of the rays. Both the head movements and the swaying movements of the body conspire to this end. And in this position it is probable that more light is received by the eyes than in any other.

Whatever may be the explanation of the process in physiological terms, it is evident that animals perform many actions simply because they derive pleasure from so doing. Such actions are, I am inclined to believe, not quite the same as simple reflex acts, even in such a creature as Ranatra. Simple reflex acts may, however, be accompanied either by pleasure or by pain, although neither of these states nor their nervous correlates determines the nature of the reaction. But in most cases of the pleasure-pain response the character of the nervous analogue of these states somehow determines whether movements producing them shall be continued or inhibited. Movements which bring agreeable stimulation are persisted in, while those which produce painful effects are checked. If any stimulus brings a pleasant sensation an animal is apt to make efforts to get more of that stimulus. If the creature possesses any power of association, movements bringing an increase of the stimulus are repeated and come to be performed with ever increasing

340 Journal of Comparative Neurology and Psychology.

readiness. In view of the fact that insects are capable of forming quite complex associations and of guiding their action by their previous experiences there is no reason to balk at the admission of the possibility that an insect may seek the light because it thereby derives an agreeable feeling. Whether the phototaxis of Ranatra is to be interpreted as a form of pleasure seeking, or as a matter of direct reflexes, a sort of behavior forced upon the creature from without, is a question not to be decided without careful observation and experiment. That an organism travels towards the light in the direction of the rays, even though it goes from a brighter area into a darker one, is no decisive proof of the latter theory. We might interpret the orientation of an animal going towards the light as due, not to forced reflexes, but to a more or less voluntary effort to retain a maximum of stimulation. If a creature has reached that plane of psychic development at which it acquires a capacity of reacting with increasing readiness to stimuli which produce an agreeable effect, and of discontinuing those reactions which give rise to unpleasant results, it seems probable that it would not stupidly continue to orient its body and go towards the light if this were not associated with some sort of gratification. If the organism be capable of profiting by experience, we should expect such meaningless behavior sooner or later to be stamped out.

The impulses which are concerned in the reactions of Ranatra to light pass through the principal centers of the nervous system. The responses of the creature to light, if we consider them as reflex acts, cannot, therefore, be reflexes of a very simple kind. The impulses involved must pass from the eye to the brain, thence through the sub-oesophageal ganglion and the ganglia of the ventral nerve cord to the nerves supplying the legs. There are several links in the chain of neurons between the reception and the discharge of the stimulus, and there is abundant opportunity afforded for modification of the reaction through impulses coming from associated parts of the nervous system. As a matter of fact, we have found that the light reactions of Ranatra are modified in this way, as well as conditioned by the general nervous state of the animal. And

Holmes, TJie Reactions of Ranatra to Light. 341

whatever changes in the nervous system may be due to the formation of associations or the effect of habit afford a further element in determining the nature of the phototactic response.

Now, while the light reactions of Ranatra take place to a large degree in a stereotyped and mechanical way, there is much to indicate that the insect seeks stimulation by light much as a bit of food or other object of interest is sought by a higher animal. A dog seeing a piece of meat on the other side of a fence makes a variety of efforts to get the prize. He runs up and down looking for a hole where he can get through or perhaps a low place where he can jump over. We cannot class his actions as direct reflexes in response to outer stimuli. His conduct is determined by the effort to secure a gratifying experience and various methods are employed to gain that end. A Ranatra seeking the light is much less resourceful than a dog trying to get on the other side of a fence, but, to a certain extent, the behavior of both animals is determined by interest in the object sought. As we have seen, Ranatra is able to go towards the light despite obstacles of various kinds. Even when one eye is totally blackened over and all but a small part of the posterior surface of the other is covered, many Ranatras are able to go towards the light in nearly a straight line. How is this done ? A Ranatra in this condition is in much the same situation a man would be if one eye were blindfolded and he were attempting to reach a light by walking backwards. A man would accomplish this, if he had not previously seen where the light was situated, in a very indirect manner. By moving his head from side to side he could determine the darkest part of his visual field, and by facing in that direction he would finally be brought by walking backwards to the light. The darkest part of the visual field in this case is a rather indefinite thing as anyone may readily assure himself by trying the exexperiment. A man's course towards the light would probably be very indirect, and it is not a little remarkable that so benighted a creature as Ranatra can reach the goal by so direct a course.

The two sides of the compound eyes must function differ

342 Journal of Comparative Neurology and Psychology.

ently in one respect, since in going towards the light the anterior side is turned so as to face the brightest part of the field of vision while the posterior surface of the eye is kept facing the darkest part. The impulses from the two sides of the eyes do not antagonize each other. The experiments of blackening over the anterior sides of the eyes show that the sides looking away from the light, as well as those looking towards it, play a part in orientation, although they must respond, so far as seeking intensity of stimulus is concerned, in opposite ways.

The method of trial and error plays, I think, only a subordinate role in the phototaxis of Ranatra, although in some situations it undoubtedly comes into play. There is a certain amount of random movement in the behavior of this form, but deviations from the direct path to the light are usually corrected by an appropriate turn, and not by making a lot of trial movements and following up the successful ones. Ranatras often show periods of hesitation between two directions of turning. Specimens that perform circus movements when one eye is blackened over usually manifest a decided hesitancy when they have turned so that the normal eye looks approximately away from the light. They often stop in their course, turn this way and that, often many times, and occasionally settle down to rest, as if in despair over the situation. Sometimes they turn directly towards the blind side and go to the light ; at other times they reach the light only after performing a complete circus movement to the left. When past the critical point their movements usually take place with little hesitation.

Specimens with only the posterior half of one eye exposed often turn slightly this way and that during the first part of their course, as if attempting to get their bearing. If they deviate either to the one side or the other they frequently stop, as if they perceived something to be wrong, turn back and forth several times, and then proceed nearly straight towards the light. In nearly all specimens thus treated one can detect a tendency to veer over towards the normal side, but there are equally obvious efforts to check deviations that are made from the direct path. There is an uneasiness which appears much like impa

Holmes, The Reactions of Ranatra to LigJit. 343

tience when a Ranatra that shows a strong positive reaction has deviated considerably from a straight path towards the Hght. When the normal eye looks away from the light the specimen may sway from side to side, back off, then go ahead again, growing continually more excited until finally it takes to flight. If the wings are fastened down to prevent their being used, the insect frequently spends several minutes before getting out of its dilemma. Sooner or later a fortunate movement is made which brings the creature into a situation such that the tendency to turn to a particular side meets with little opposition. Out of many trials made in this state of perplexity a successful one is finally made and followed up. Individuals vary greatly in their conduct in such a situation. Many correct their course before getting far out of orientation ; others, when they get out of line, turn back again with comparative readiness.

We might explain the function of the posterior side of the eye in orientation on the supposition that a movement which brightens the visual field brings about a reflex that causes a turn in the opposite direction. In this way deviations from the position that kept the darkest part of the field of vision in focus would be checked and the insect would, therefore, travel towards the light. It is more difficult, however, to explain the orientation of an insect in which a small area of the lateral surface of one eye is the only part exposed. If the insect moved so that the eye would look toward the darkest part of the visual field it would place its body with the blind side towards the light. If, on the other hand, it moved so that the eye would receive the maximum amount of stimulus, the opposite side would be brought toward the light. It is obvious that the insect moves so that the eye receives neither the maximum nor the minimum amount of stimulus, but is kept exposed to light of an intermediate degree of intensity. The degree of intensity varies, moreover, in every step of its course, so that there is no justification for explaining the orientation of the insect through the effort to keep a certain intensity of light constantly before the eye. It is remarkable that Ranatras so treated frequently go towards the light in nearly a straight line. Moreover, if the light is moved

344 Journal of Comparative Neurology and Psychology.

during their progress they change their direction of locomotion so as to continue going towards it. If the light is not carried too far to one side the insect may be made to follow it around in either direction. In the specimens employed care was taken that every other part of the eyes was thickly covered except a small portion of the surface of one side, so there is no doubt that the movements of the insect were directed only by light entering at this point. The side of the eye in this case doubtless functions as it does in a normal individual, but it is difficult to explain satisfactorily the orientation of the specimen either through direct reflexes, or by the method of trial and error. Were the insect so constituted as to respond to an increase of light entering the left eye by a turn to the left and to a decrease of light by a turn to the right, we can understand how, when once pointed towards the light, a straight course might be preserved. If the insect turned towards the right there would be an increase of light entering the left eye which we might suppose stimulates the insect to turn in the opposite direcrion. Deviations to the left would cause a diminution of light entering the left eye, which we might suppose acts as a stimulus to turn to the right side. The right eye may be supposed to act, vmtatis mutandis, in a similar manner. The numerous cases of reaction to shadows {Scattenempfindlichkeit) which are found among several groups of animals show that diminution in the intensity of light may act as a stimulus as well as an increase in intensity. If it be the variations in the intensity of light which afford the stimuli for turning in the one or the other direction, we can attribute to all parts of the eye essentially the same function. If the posterior side of the left eye is all that is exposed and a Ranatra that is facing the light turns to the right side the exposed part of the eye receives an increased stimulus which we may suppose brings about a turn to the left. A turn to the left of the median position, up to a certain point, would probably (owing to the body intercepting a part of the rays) diminish the light entering the posterior side of the eye, thus causing a turn to the right. In this way the creature might be supposed to maintain a straight course towards the light. A

Holmes, The Reactions of Ranaira to Light. 345

similar explanation might be applied to the orientation of a Ranatra with only the anterior half, or in fact any other part of the eye exposed. The difficulty with this explanation is that an insect that started with an oblique orientation to the direction of the rays would tend to continue in that position, since a departure from it towards either side would be followed by a compensatory movement. If, however, light produces a constant effect upon the muscular tonus of the body, irrespective of affording stimulation by variations in intensity arising from turning towards different sides, we can better explain the fact of orientation. Take the case of a Ranatra with only the lateral surface of one eye exposed. Light entering the eye tends to increase the action of the flexors on the same side, and that of the extensors on the opposite side of the body. Through this action alone the insect, so far as light directs its movements, would continue to go around in a circle indefinitely. But as a matter of fact, as its circus movements bring the eye away from the light, they become checked and are followed by a turn in the opposite direction. What is the stimulus to this turn? Obviously there is a diminution of light received by the eye as it turns away, and we may regard this diminution as a stimulus to a movement in the opposite direction. This stimulus may be conceived, then, to overcome the tendency to the performance of circus movements, and thereby bring about an approxi mate re-orientation of the creature to the direction of the rays. That there is a conflict of impulses as the insect turns away from the light towards the side with the functional eye is evinced by the hesitancy, the swaying this way and that which often occurs when the creature has reached a position such that the amount of light received by the eye is materially diminished. The effect of light upon the tonus of the muscles of which the behavior of Ranatra gives so much evidence, fails to account for the fact of compensation in the movements of the insect. The responses to variations in the intensity of light, on the other hand, do not adequately account for the preservation of efforts to attain a parallel orientation to the direction of the rays. Each of these factors may, however, supplement the other in such a way as to

346 Journal of Comparative Neurology and Psychology.

cooperate in the maintenance of a direct course towards the light.

But even if we invoke the aid of both the factors mentioned, we are not able completely to bring the phototaxis of Ranatra under the category of direct, mechanical reflexes. Consider the behavior of the Ranatra with its right eye blackened over, placed at right angles to the rays with the normal eye looking away from the light. The effect of light upon the muscular tonus of the creature would tend to make it turn towards the left side. But either at first, or after a few trials, the insect turns directly to the right and goes to the light. This is done with greater readiness after several trials are made ; soon the insect turns immediately after being set free. In these experiments the specimens were turned about, first one way and then the other, before being set down on the table, but this made no difference in the directness with which they turned towards the light. In many cases the insects would struggle to turn towards the blind side to get to the light before they were liberated from my hand. So far as could be observed, there was no preliminary feeling about, no employment of the method of trial and error; the insect seemed to retain awareness of the position of the light, since it immediately made for it by the shortest route as soon as it was liberated. The non-mechanical character of the response is further evinced by the fact that habits of turning towards a certain side may be persisted in in situations in which turning would otherwise take place in the opposite direction. The fact that Ranatras and Notonectas which have one eye blackened over come, after several trials, to dispense with circus movements towards the normal side and correct deviations from their course by a direct turn in the right direction affords further evidence for the same conclusion. We have seen that Ranatras which at first will turn only a slight way towards the blinded side when the light is carried over in that direction will make sharper turns after they have become accustomed to following the light around towards that side. These features of the phototaxis of Ranatra indicate that seeking the light has an attractiveness or interest much like that which catching prey has

Holmes, TJie Reactions of Ranatra to Light. 347

for a higher animal. The phototactic movements of the creature are not merely stereotyped reflexes which the insect is involuntarily forced to perform. To get to the light is an end which is arrived at if not by one method, then by another.

The phototaxis of Ranatra comes, to a considerable degree, I believe, under the pleasure-pain type of response. Why a creature should be so constituted as to derive satisfaction from so stupid a performance as wildly chasing after a strong light is a subject that need not concern us in our present quest. The behavior of Ranatra presents the essential features of the pleasure-pain reaction of higher forms, and we are justified, I believe, in classing it under this heading. The fact that Ranatra will continue to follow the light even when it is brought thereby into a situation that produces a fatal effect does not necessarily exclude its conduct from this category. While it is true that animals tend to continue reacting towards stimuli that produce a beneficial effect and away from stimuli that bring about deleterious results, the rule is by no means absolute. In a state of nature Ranatras probably are rarely, if ever, exposed to conditions that produce as strong positive phototaxis as they show under artificial conditions in the laboratory ; and there seems to be no benefit that possibly can be derived from their strong propensity to seek the light. This propensity, like that of human beings for certain stimulants and narcotics, has probably not been evolved by natural selection for any useful purpose, but is an incidental result of the way the creature is constituted. Whether there is any connection between pleasure-giving acts which tend to be repeated and acts which secure some benefit to the organism closer than that which would naturally be established through selection may well be doubted. Neither an animal's direct reflexes nor its attempts to seek some source of gratification infallibly lead to securing some benefit; and the fact that a certain kind of behavior is persisted in until it brings about fatal effect does not ipso facto enable us to decide under which of these categories it falls.

It is the apparent telecity in the efforts of Ranatra to reach the light which it is difficult to understand according to a purely

348 Journal of Comparative Neurology and Psychology.

reflex theory of phototaxis. This is a feature of the creature's behavior which, I am inclined to believe, we shall not be able to understand until we can explain the physiology of the process whereby certain stimuli when they have been received one or more times are sought either directly, or indirectly, by a more or less round-about method, while other stimuli when they are experienced one or more times come to be avoided. Did light afford a stimulus of an unpleasant nature, it is probable that the positive phototaxis of Ranatra would soon be inhibited. If the phototaxis of Rantra falls to a certain extent under the category of reflex action, as much in its behavior indicates, the reflexes concerned are in line with a strong instinctive interest of the animal in seeking the light. This interest may lead to successful attempts to get to the light in situations in which purely reflex responses alone would fail. It acts as a sort of regulatory agent in the conduct of the insect, bringing its actions to a successful issue, which could not be attained by a purely machine-like mode of response.

It is through instinctive interests in certain things rather than by simple or even complex reflex acts that the conduct of higher animals is mainly guided. The play activities of higher animals, for instance, are performed, not because they are reflex responses to particular things in the environment, but because the animal is so constituted that it derives satisfaction from their performance. An animal interest may be chained, by virtue of its organization, to certain ends, such as the capture of a certain kind of prey, or the construction of a particular kind of habitation ; but at the same time its conduct may show considerable plasticity as regards the methods by which these ends are attained. In instinctive behavior, as in the morphogenic processes which lead to the establishment and maintenance of the normal form of the body there is an apparent effort to reach a certain end result, despite obstacles and unusual conditions. The explanation of this pecuHarity of animal behavior is a problem of fundamental interest. If we attempt to resolve highly complex modes of behavior into simple direct reflexes, we soon find ourselves at the end of our tether. Even in so apparently

Holmes, The Reactions of Ranatra to Light. 349

mechanical procedure as the phototaxis of Ranatra we encounter peculiarities which indicate that we have not struck bottom in our analysis of the phenomenon.

In many ways the phototaxis of the Ranatra seems to be intermediate between purely reflex conduct on the one hand and conduct of the pleasure-pain type on the other. These two kinds of behavior seem to be harmoniously combined in many instincts, if not in most of the more highly involved modes of instinctive action. Just what the steps are which lead from the one to the other we do not know. We are still in want of a satisfactory explanation of the pleasure-pain type of response. When we are able to supply one we shall be in a position to give a more adequate interpretation of the phototaxis of higher forms than can be supplied at the present time.


Verhandlungen der Anatomischen Gesellschaft auf der Achtzehnten Versammlung in Jena. Anat. Anz. Erganzungsheft zitm 2§ B. 1904.

Schultze, O. Ueber die Entwickelung des peripheren Nervensysteme. 3-7. The growing nerves in the embryo are syncytial in structure. The sheath nuclei are derived from the nuclei of this nervous syncytium. The plexus nervosus profundus of amphibian larvae (Czermak) is a sensory syncytium which is, and has arisen, in continuo with the developing nerve, and is not derived by a fusion of independent units. Schultze's interpretation is in opposition to the neurone theory.

Koelllker, A. Ueber die Entwickelung der Nervenfasern. 7-13.

This paper is directly opposed to Schultze's and in favor of the neurone theory. Each axone grows out as a process of a single nerve cell and the sheath nuclei are mesodermal in origin. This mode of development, the author asserts, is followed in Vertebrata, Arthropoda and MoUusca. The process may be simpler in lower forms.

In the discussion which follows these two papers, Froriep argues for the ectodermal origin of the sheath nuclei by a migration from the central system along the nerve roots. Retzius, Benda, Ballowitz, Harrison and Disse cite various instances and observations in favor of the neurone theory. The necessity of experimental work upon the regeneration of the peripheral end of the severed nerve fiber is emphasized by Roux and Barfurth.

Joseph, H. Ueber eigentUmliche Zellstrukturen im Zentralnervensystem von Amphioxus. 16-26.

The cells in question are the large cells in the anterior region of the nerve cord which v. Kupffer called "dorsale Ganglienplatte." The author finds similar cells in the corresponding position in the caudal region of the cord. All of these dorsal cells, contrary to other authors who differ among themselves, are unipolar, and, in structure, correspond exactly with the "Sehzelle" of Hesse. They are not, however, capped with the pigment cell which is characteristic of the typical "Sehzelle." Their axones pass towards the periphery of the cord and probably enter the posterior roots. In the caudal region of the cord "Sehzelle" occur without the pigment cap, in which case they are identical in structure with the dorsal plate cells of the same region.

Literary Notices. 3 5 i

Hochstetter. Ueber die Nichtexistenz der sogenannten Bogeiifurchen an den

Gehirnen lebensfrisch konservierter menschlicher Embiyonen. 27-34. Schaper, A. Zur Frage der Existenzberechtigung der Bogeiifurchen am Gehirue menschlicher Embryonen. 35-37Both of these papers on the embryonic fissures agree with the recent work of Retzius, Mall and Goldstein, who find no true fissures in the wsU preserved human brain during the period to which His assigned his "Bogenfurche."

Ramstrom. Ueber die Innervation des Peritonaeums der vorderen Bauchwand. This study was made upon I'n^ra vitam methylene bhie impregnations of the psritoneum and associated tissues. In the mouse, the peritoneum on one side the median line was ludunted entire. No fibers from the i)hrenic were found entering this region. The innervation is wholly from the intercostal nerves, through a complicated system of plexuses.

Lubosch, W. Uber den Bau und die Entwickelung des Geruchsorganes von Petromyzon. 67-75.

This paper bears upon Blaue's theory of the origin of the olfactory epithelium by a process of fusion of primitive cutaneous sense organs. The author conceives the olfactory buds of Petromyzon as representing a protracted ontogenetic process. The fuiidament of the olfactory organ is in the form of a bud-like differentiation of the integument. The organ develops by a repetition of this process. Lubosch considers that Blaue's theory may be tenable upon the hypothesis that in the olfactory organ the nerve cell has maintained the primitive position, which it originally held in both olfactory and taste organs. Perfectly closed follicles are found in the nasal cavity of ten-centimeter larvae.

G. E. c. Rutter, Cloudsley. Natural History of the Quinnat Salmon. Bull. U. S. Fish Commission for 1902, 65-142. 9 Pis., 13 Figs. 1904. The objects of this investigation were to determine when young salmon should be liberated from the hatcheries, to discover a method for removing and fertilizing the eggs left in the fish after artificial spawning, and to fix the site for a new hatchery. Other matters of interest discussed are the activity of spermatozoa after the milt is mixed with water, the fertilization of the ova, care in handling embryos, alevins and their enemies, the fry and their food, parasites, and migrations, the food of young salmon and their period of growth. The adult salmon are treated in relation to migrations, changes after entering fresh water, sexual differentation, natural propogation, injuries and diseases, and death which ensues after once spawning. This contribution is of considerable scientific importance as well as being of practical value.


352 Journal of Comparative Neurology and Psychology.

Kerr, J. Graham. On Some Points in the Early Development of Motor Trunks and Myotomes in Lepidosiren paradoxa (FiTZ.). Trans. Roy Soc. EdinbitrgJt, 41, Part i {No, 7), pp. irg-i2S, 6 plates, igo.f.

The author's observations support the conclusion that the motor spinal nerves of Lepidosiren first appear as strands of soft granular protoplasm, extending between the spinal cord and the myotome. These strands are not cellular. They later become fibrillated and invested by mesenchymatous sheaths. c. J. ii.

Coggi, Alessandro. Le ampolle di Lorenzini nei Gymnofioni. JMonitorc Zoologico Italiano, 15, No. 24g-j6.

On the basis of the figures and descriptions of the vSarasins, the author compares the ^' Nchcuolircu,^' or accessory lateral line organs on the head of Ichthyophis with the ampullae of Lorenzini of elasmobranchs, concluding that the organs are strictly homologous, thus strengthening the current belief in the archaic character of the Gynmophiona.

It will be recalled that the reviewer has suggested (this Journal, vol. 13, p. 135) a similar homology between the ampullae of Lorenzini and a type of "small pit organs" found in the skin of the North American siluroid fishes, though the evidence for this relationship is not regarded as conclusive c. j. h.

Freidenfelt. T. Ueber den feineren Bau des Visceralganglions von Anodonla. Lituds Universitets Arsskrift, Bd, 40, Afd. 2. Nr. 5. Lund, 1905.

A detailed description of the neurones of the visceral ganglion and their connections, after methylene blue preparations, illustrated by four good plates. The author finds no true net-work (protoplasmic continuity of the older authors) in the neuropil, but refrains from expressing an opinion on the question of continuity of the neuro-fibrils of Apathy and Bethe. c j. h.

Tumors of the Cerebellum.

Under the above title, the A. R. Elliott publishing Co., of New York, has re-printed six papers by Drs. jNIills, Frazier, de SchweinITZ, Weisenburg and Lodholz. While the clinical and surgical details do not fall within our limits, we call attention to the cases reported and to the summaries of literature as important contributions to the theory of cerebellar functions. c. j. h.

Banchi, Arturo. Di un cervello senze commessure e con hinzioni apparentemente normali. Archivio di Fisiologia. Vol. I, pp. 614-618, 1904.

A brief statement of a remarkable case of cerebral malformation where the mental life seemed perfectly normal. An extended account of the case will shortly appear in the Archivio Italiano di Anatomia e di Embriologia j. c. b

Literary Notices. 353

Gross, Hans. Kriraiual-Psychologie. Leipzig, Vogd, 1905. xii -|- 701. Zweite Auflage. Although intended primarily for those who have professional interest in juristic ]jroblenis and in the scientific side of criminology, this hook of Geoss' contains much that is of value to the comparative psychologist. In reading it one is likely to be impressed with the obvious need of psychological knowledge in the treatment of criminals and defectives. But, at the same time there is opened up, in thoroughly scientific systematic fashion, a vast realm of abnormal psychology which has been almost entirely neglected by scientific investigators, namely, the psychology of crime. The work is authoritative, and in this second edition it has been considerably improved iu form and increased in value. R. m. y.

Streeter, George L, The Development of the Cranial and Spinal Nerves in the Occipital Region of the Human Embryo. Am. Jour. Anat., 4-, 83-116. 1904. A study of numerous human embryos from the collections of Hrrtwig, Mall, His and Golci. The period, 20-81 days, involves all the changes in form and relation from the time when the peripheral nerves are first ca])able of reconstruction to the time when the adult condition is practically established. The ganglionic crest in the cerebral region is from the first distinct from the trunk ganglia, though it is connected with the latter by a thin sheet of loosely arranged cells. The ganglia petrosum and nodosum are apparently adherent to thickened patches of the epidermis. The ganglionic crest of the tenth and eleventh nerves at first shows no signs of segmental arrangement. The latter nerves are found to belong to the same complex and alike have root ganglia and motor and sensory roots. In the course of development the sensory element comes to predominate in the cephalic portion of the complex and the motor in the caudal portion. The hypoglossal is like the ventral cervical roots in position and segmental arrangement. The ventral spinal nerves are developed earlier than the dorsal roots. g. E c.

Carlson, A. J. Contributions to the Physiology of the Ventral Nerve Cord of Myriopoda (Centipedes and Millipedes), foil)-. Ex. Zool. 1. 269-288. 1904.

In the ventral nerve cord of Scohpcndra the conduction rate is greater in the antero-posterior direction. This antero-posterior differentiation, is most marked in the short types. Experimental studies on the localization of centres of control of important reflex functions.

G. E. C.

354 Journal of Comparative Neurology and Psychology.

Mettler, L. Harrison. A Treatise on Diseases of the Nervous System, 9S9 pages, Cleveland Press, Chicago, 1 905.

" The Neurone Doctrine is now an accepted fact." After an introduction of 88 pages devoted to classification, etiology, symptomatology, etc., the writer describes first the neuronic diseases, beginning with the neuroses, hysteria, vertigo, neurasthenia, epilepsy, and all the other neuroses, the sympathetic included, and the system diseases — locomotor ataxia, the systemic motor palsies (inclu;!ing myasthenia), and the mixed afferent and efferent types.

The non-neuronic diseases (p 457-957) include the changes starting from the glia, vessels and sheaths, i)ractically all the diseases of the peripheral nerves and diffuse and focal diseases of the nervous system. Alcoholism and some other toxic and infectious disorders with prominent participation of the nervous system come last (j).

959-971) Notwithstanding this somewhat peculiar arrangement, Mettler gives a very readable presentation of the chief facts, general and clinical, in most respects superior to several of the books which are in the hands of many students In view of the excellency of the matter it is rather a pity that an essentially doctrinal issue should be at the top of every page — the distinction of neuronic and non-neuronic diseases, in the somewhat arbitrary sense in which Mettler makes his subdivision.

The illustrations are well chosen and very instructive. The book deserves recommendation. Lapses are relatively rare, as on p. 486, under Burdach's column: " astereognosis, loss of reflexes and every physiological act that involves sensation, directly or rcllexly, arc diminished or abolished," when it is diseased. a. m. Child, C. M. Studies on Regulation. The Relation of the Central Nervous System and Regeneration in Leptoplana : Posterior Regeneration. /oiir\ Ex. Zoology, 1, 493-512. 1904.

The presence of about half of the cerebral ganglion has both quantitative and qualitative influence on the reaction of pieces, and the efficiency of the brain in this respect is not localized. This efficiency is affected by the amount of nervous tissue present in the piece and not by the presence of a particular part of the nervous system. The influence of the nervous system upon posterior regeneration is quantitative only, and is not " formative" but indirect through the influence of functional conditions. In regeneration, also, the influence of the nervous system depends upon theamoinit present and not upon the presence of a particular part : a condition which argues against the idea of functional centers. c;. e. c.

The Journal of Comparative Neurology and Psychology

Volume XV 1905 Number 5



{From the Physiological Laboratory of the University of Kansas.)

As a contribution to the study of the comparative physiology of the brain, it seemed important to undertake a more detailed investigation of the functions of the different parts of the frog's brain than had hitherto been made.

The work done by me under the supervision of Dr. Ida H. Hyde — to whom I am under great obligations — can, at best, constitute only a small link, in the long and complex chain of comparative physiology. The experiments extended over a period of two and a half months, and the material operated on consisted of more than fifty frogs of the species Rana temporaria.

I employed Goltz's method, i, e., that of excising various parts, or aggregations of parts, of the central nervous system, and observed the resulting phenomena, particularly the deficiency phenomena. Both uni- and bilaterial excisions were resorted to. The animals were anaesthetized with ether until they would not readily turn over on the ventrum when put on the dorsum, but would respond only with convulsive jerks of the muscles of the limbs. For operations on the medulla the anaesthesia was carried further to prevent movements during the operation After the frog had reached this stage of anaesthesia, it was tied with its ventrum upon the frog board. A sponge moistened with ether and placed in front of the nares was sufficient to keep the animal in the desired stage of anaesthesia

356 Journal of Comparative Neurology and Psychology.

during the operation. The skin ov^er the cranium of the frog was first washed with a weak solution of bichloride of mercury, which was immediately removed with sterilized cotton and then Avith 95 ^ alcohol. The instruments and the hands of the operhad been previously sterilized with 5 ^r carbolic acid, followed by 95 % alcohol.

An incision was made through the skin of the cranium, following the mid-dorsal line. The skin was held apart by a pair of weighted hooks, and an opening was made with a small trephine through the cranium over the part of the brain to be removed. Since the two occipital arteries take their course laterally, this opening was made as near as possible on the mid-dorsal line of the cranium. With a delicate pair of forceps small pieces of the cranial bones were carefully broken away to make the primary opening of sufficient size for the operation. Great care was taken not to injure the meninges and the vessels of the choroid plexus therein. A i ^, solution of adrenelin was used with great success whenever injury to the choroid plexus, or even the internal carotid and branches caused profuse hemorrhage.

A small pointed knife and a very fine needle were used for excising or burning the different parts. Ringer's solution was used to moisten the wound after the operation. No antiseptics were used in the wound, which was closed with sterilized silk sutures. The animals were kept isolated in moist moss, which was frequently changed. At first they were observed two or three times a day, but, after several days, once a day, and after seven to ten days, once every two days was sufficient. Most of the frogs were fed after the operation; a few — depending upon the nature of the operation — were soon able to catch flies. The observations were tabulated, dated, and later the post mortem findings were added. Two animals were used for each operation, and the reactions compared with one another and with the reactions of normal frogs.

I seldom noted the immediate effects within one-half hour after the operation, since they were obscured by the effects of the ether. In some instances where it was quite evident that

LoESER, Functions of the Frogs Brain. 357

the effects of the operation were present before one-half hour after the operation, I called attention to this fact.

Although I used a lens during the operation and postmortem, I made no microscopical preparations of the specimens; therefore that degree of exactness which is obtainable by means of microscopical preparations was not reached.

By the term deficiency phenomena — as I use it in this article — I mean those phenomena which are normally present; but which are permanently — that is, for two and a half months — absent after excision of a part of the central nervous system, or permanently exhibited in a degree different from that in the normal animal.

Operations and Results.

I. The first operation consisted in removing one cerebral hemisphere. As a result, the animal displayed muscular weakness on the crossed side during the first three days. Thus the frog, when jumping, turned slightly toward the unimpaired side. Four days after the operation this temporary deficiency had permanently disappeared.

II. In the second operation I removed both cerebral hemispheres, and noticed as a result bilateral muscular weakness lasting about three days and accompanied by subnormal skin reflexes. No deficiency phenomena were noticeable the fourth day after the operation.

III. The third operation consisted in burning out the thalamus opticus of one side. The results were fewer voluntary movements ; bilateral defective sense of touch, especially in the fore-limbs ; almost complete loss of sight on the crossed side ; and weakness in the crossed legs. The first, second and fourth difficiencies improved slightly during the first two weeks, but the sight defect remained unimproved.

IV. After burning both optic thalami with hot needles, sight was almost completely abolished on both sides, and the animal was even more defective in the sense of touch and voluntary movements than after operation III. All three deficiencies showed practically no change during the two and a half months after the operation. The marked parallelism between the loss of touch-sensibility and voluntary movements was even more noticeable than after operation III. Flies put into the cage after operations III and IV were never caught.

V. The fifth operation consisted in unilateral burning out of the lobus and thalamus opticus. The result was a complete and permanent loss of sight on the crossed side ; forced movements straight ahead, which gradually improved ; transient retention of urine, and reduction in voluntary movements and touchsensibility, exactly as noticed after removal of one thalamus opticus. Nearly all of the animal's movements were at first forced — straight ahead — during which it jumped higher and apparently with more vigor than normally. Stimulating the frog caused it to jump so powerfully that it frequently injured itself. It assumed,

358 Journal of Comparative Neurology and Psychology.

moreover, abnormal positions of the limbs, for instance, adduction of both forelegs until they almost touched each other.

VI. The sixth operation consisted in burning both lobi and thalami optici. It caused total blindness on both sides ; reduction in voluntary movements and touch-sensibility, as noticed after bilateral removal of the thalami optici and the other defects of unilateral excision of the thalamus and lobus opticus, but in a much more marked degree. The abnormal retention of urine was one of the first phenomena to disappear ; three weeks after the operation it was no longer noticeable. The forced movements also improved, and were not noticeable after two weeks unless the animal was greatly excited. The sight was permanently abolished on both sides.

VII. The seventh operation — consisting in unilateral burning of the lobus opticus — caused slight defect in sight on the crossed side, straight forced movements, and abnormal retention of urine. The last two phenomna were as indicated in operation V. Sight was normal on both sides after two weeks.

VIII. After the next operation in which both lobi optici were removed, the trifling defect in sight was bilateral, the straight forced movements were stronger, and the retention of urine more pronounced than after unilateral excision of the lobus opticus. The forced movements and the abnormal retention of urine improved as indicated in operation VI, and the slight bilateral defect in sight improved within ten days so as not to be detected.

IX. The ninth operation, which consisted in removing the anterior half of the two lobi optici, showed, during the first two weeks, the abnormal retention of urine and slight bilateral loss in sight, as noted under bilateral removal of the lobi optici. Later, until two and a half months after the operation, sight was normal on both sides, and urinary retention gradually disappeared during three weeks after the operation.

X. In the tenth operation the posterior half of both lobi optici was removed, and the result was that the forced movements became apparent as noted in operation VIII. Later these were present only when the frog was extremely excited.

XL By removing both optic thalami and the anterior half of the lobi optici all the phenomena as noted under bilateral removal of the thalami and lobi optici, with the exception of the forced movements, were noticed.

XII. By burning the dorsal half of both thalami and lobi optici sight was only slightly injured for two weeks, whereas the abnormal retention of urine, the forced movements, the depression of voluntary movements and touch-sensibility were injured according to the amount of nerve tissue removed.

XIII. After bilateral removal of the cerebellum, the jumps became very unsteady and were weakened in force but not much reduced in rate. Most of the time the frog lay stretched in contact with the floor of the vessel and covered with moss. The unsteady and weakened condition of the limbs persisted in a marked degree two and a half months after the operation (especially when the frog was greatly excited). The frog's hibernating tendency persisted two and a half months after the operation and it did not succeed in catching flies put into its cage.

XIV. Unilateral excision of the cerebellum caused a number of phenomena which also appeared more pronounced after unilateral excision of the medulla an

LoESER, Functions of the Frog s Brain. 359

terior to the origin of the vagus group; the limbs of the crossed side — especially the posterior one — were abnormally extended and abducted, locomotion was always towards the homonymous (operated) side. The spine was slightly bent, concave towards the homonymous side in the thoracic region. Though the forced position of the limbs of the crossed side and the curved state of the spine improved considerably in one week after the operation, yet they, together with the squatted position of the body, and locomotion, remained quite characteristic two and a half months after the operation.

XV. Bilateral excision of everything anterior to the medulla leaves the croak, the turning-over, the swallowing and heart-action reflexes, the sense of equilibrium and respiration intact. The croak and turning-over reflex could hardly be elicited during the first few days, but they were normal after one week. If the anterior portion of the medulla was also removed, apnoic pauses alternated with periods of normal respiration. During such a pause, the animal remained quite motionless for about five minutes. To avoid injury to the medulla, it is best to burn off the cerebellum after everything anterior to it has been removed.

XVI. After unilateral excision of the right side of the medulla anterior to the calamus scriptorius abdominal respiratory movements on the homonymous side stopped, whereas the narial and oral as well as swallowing movements were still normal.

The animal assumed a flat position and the head and homonymous side were lower than other parts of the body. The legs on the crossed side, especially the posterior one, were far extended and abducted ; and the homonymous legs, especially the front one, were superflexed and adducted. Forced rotary movements around the sagittal axis of the body and towards the homonymous side were present as the sole movements on land and in water. The eye reflex to pressure on the homonymous side of the body was abolished, but was normally present on the crossed side. The crossed eye was turned up and outward, and the homonymous one, down and inward. The homonymous eye was turned up synchronously with inspiration, and down with expiration. The nictitating membrane of this eye was evidently beyond the control of the frog since it covered the eye, especially when it moved downward. The spine and neck were curved as after unilateral cerebellar excision and tremors were present in the limbs on both sides, but were more pronounced in the weaker legs of the homoymous side. A weakened croak reflex still persisted.

These defficiencies were almost entirely confined to the homonymous side. Sight was perfect on the crossed side, and although the eye reflex to pressure was abolished on the homonymous side, the animal still saw with this eye. The rotating table reactions were complex and will therefore be described in greater detail. One half to five hours after the operation the animal did not react on the rotating table with either head or body movements. Twenty-four hours later, when the table was rotated in a clock-wise direction, i. e., towards the operated side of the frog, the animal held its head in anti-clock-wise direction during rotation ; after rotation, the head was brought back beyond the sagittal axis of the body in a clock-wise direction. On rotating the table in anti-clock-wise direction, no head movement was produced during and after rotation, but a slight circular movement of the body in anti-clock-wise direction was seen after rotation. Forty-eight hours afterward, while rotating the table in a clock-wise direc

360 Journal of Comparative Neurology and Psychology.

tion, a slight movement of the head, as previously noted, was seen during and after rotation of the table ; and sometimes a circular movement of the body in clock-wise direction, after rotation of the table. While rotating the table in anticlock-wise direction, no head movement was seen, but a slight circular movement of the body in anti-clock-wise direction appeared after rotating the table.

Abdominal respiration ceased during rotary movements, but subsequently it was resumed on the operated side. Fifteen days after the operation, by means of vivisection under very slight anaesthesia and exposure of both lungs, I observed that only the crossed lung respired. In the meantime, the characteristic position of body and head had but slightly improved, and was as pronounced as ever when the frog was excited. The rotary movements around the sagittal axis, which began about ten minutes after the operation, at the end of two days had diminished, but were aroused by the slightest stimulation, even by sight or movement, on the rotating table toward the crossed side ; they could also be inhibited in part or completely by touch and movement on the rotating table toward the homonymous side.

After three days the position of head and body was temporarily normal and the superflexed or superextended legs respectively were also temporarily brought back to their normal position. After five days the rotary movements showed themselves only when the frog was jumping ; the animal usually made a complete circuit around its sagittal axis during the jump. Three days later the frog sometimes landed on the dorsum after a jump ; and the croak reflex became stronger. By the tenth day the rotary movement was manifested by the fact that the animal landed on the side after a jump. The jumps never became as strong and skilled as before the operation. The position of the head, body and legs had improved ten days after the operation, although still quite abnormal. The position of the eyes and the movements of the homonymous eye and nictitating membrane had improved very little during the fifteen days following the operation. In fact, no appreciable improvement was noticeable after ten days. The defects were least noticed while the animal was at rest.

XVII. Unilateral excision of the right side of the medulla, anterior to tiie origin of the vagus group left the croak reflex and abdominal, narial, and oral respiration movements intact on both sides. It caused rotatory movements around the sagittal axis, the peculiar flat position of the front part of the body and head, the characteristic position of the limbs, the eye reflex, the position of the eyes, the movements of the homonymous eye and nictitating membrane and the rotating table reactions exactly as stated in operation XVI. The apnoic pauses in the abdominal respiration after rotary movements around the sagittal axis, were less pronounced.

XVIII. Unilateral excision of the right half of the medulla from a point posterior to the origin of the vagus group down to the calamus scriptorius, left the croak reflex intact. The abdominal respiration was slightly injured on the homonymous side and intact on the crossed, as was proved by vivisection and exposure of both lungs under very slight anaesthesia. The eyes were normal on both sides, excepting that the pupil of the homonymous eye was extremely contracted. The characteristic position of the body, head and legs was present in a lesser degree than after operation XVII. The rotatory movements around the sagittal axis were absent, but the animal landed on the homonymous side, after

LoESER, Functions of the Frogs Brain. 361

jumping, in a position at right angles to the direction in which it started. The rotating table reactions were normal in both directions. The abdominal respiration remained weakened on the homonymous side during the fifteen days following the operation ; after which the post-mortem was made. The position of the body, head and legs had gradually improved at the end of the first week so that they were almost normal. From that time until the fifteenth day no further changes were observed in these or in the unskilled and weak action of the homonymous limbs. The homonymous eye remained permanently constricted.

XIX. A unilateral excision of the middle third of the medulla, including the origin of the vagus group, abolished abdominal respiration on that side but left a weakened croak reflex. The pupil of the homonymous eye was contracted, otherwise the eyes appeared normal. The position of the head, legs and body, and also the locomotion agreed with those noted under operation XVIII. The rotating table reactions were normal and no rotary movements around the sagittal axis were exhibited. The croak reflex was quite normal two weeks after the operation.

XX. Bilateral excision of the middle third of the medulla, including the origin of the vagus groups abolished the croak reflex, turn-over reflex, and the abdominal respiration on both sides. The hea»t, though beating strongly ten minutes after the operation, became quite feeble two hours later and stopped at the eighth hour after the operation. The respiratory movements of the nares and mouth remained bilaterally present, increased very much in rate and force and exhibiting asphyxia symptoms toward the end. It was impossible to keep these animals alive longer than eight or twelve hours even when the water surrounding them was frequently changed. After this operation a complete inversion of the stomach and oesophagus, with prolapsus into and even outside of the mouth occurred. The pylorus was the most anterior portion of the prolapsus. I placed a dead fly on it and was interested in seeing it surrounded by slow peristaltic movements. The pancreas, stomach, duodenum, and the adjoining small intestine were also involved in the prolapsus. The stomach was distended to twice its normal size. The peculiar position of the head noticed after the different operations on the medulla was most conspicuous after this operation. The head pointed into the ground and formed an obtuse angle with the body. The urostyle was always less prominent after medulla injury.

XXI. Excision of everything anterior to the spinal cord abolished the respiratory movements, sense of equilibrium, eye, sight, croak, swallowing, heart, and the turning-over reflexes. A stimulus was immediately answered by co-ordinated movements. It was answered sooner and with greater certainty than when the higher portions of the brain were intact. The heart action gradually faded away during the five hours following the operation.

XXII. The excision of everything down to and including the anterior portion of the spinal cord almost to the origin of the brachial plexus caused all of the deficiency phenomena enumerated under operation XXI and in addition caused the co-ordinated movements to fade away. They failed as soon as the anterior limbs were incapable of supporting the body and helping the posterior limbs in their efforts.

362 Journal of Comparative Neurology and Psychology. Discussion and Conclusions.

The removal of one or both cerebral hemispheres proved that the motor centres of the cerebral cortex of the frog are of minor importance and the sensory centers scanty. The fact that a frog deprived of both cerebral hemispheres caught flies proves that the reflex connection of the optic nerve with the innervation of the snapping and swallowing apparatus is not located in the cerebrum. Although Goltz^ observed a loss of muscular sense for several days in his dog deprived of both cerebral hemispheres, no such defect was visible in my frogs from which both cerebral hemispheres had been removed. The decrease in voluntary movement and sensitiveness to touch which existed for a few days, I attribute to disruption of association fibers and not to destruction of motor or sensory areas.

In connection with my results, I may mention that Bickel,' using chemical agents, and both kinds of electrical stimulations, was unable to produce paralysis in the pigeon or frog unless the current was excessively strong. Bickel further concludes that in turtles the cerebrum does not regulate movements but has a motion-inducing (bewegungsanregend) influence. AdamKiEWicz^ also states that an animal deprived of the cerebral cortex carries on all bodily movements normally if it is artificially induced to perform them. Bary* determined that the new-born child lacks the principal group of motions which are controlled by the will. Moreover, Schrader'^ demonstrated that frogs deprived of the cerebrum behave, after several days, like normal frogs, and Goltz showed that his dog deprived of the cerebrum was essentially normal, excepting that the higher cerebral functions — memory, deliberation, intelligence, pleasure and envy — were absent.

1 Goltz, F. PJlugcrs Arckiv, 51, 570, 1892.

  • Bickel, A. PJlugers Archiv, 72, 190, 1S9S, and Archiv fur Anatomie

und Physiologic, 52, 1 901.

3 Adamkiewicz, a. Neurologischcs Zentralblatt, 22, 12, 546.

  • Bary, a. Archiv filr Anatomic tmd Physiologic, 341, 1898.

^ SCHRADER, M. PJOigcr's Archiv, 41, 75, 1887.

LoESER, Fimctions of the Frogs Brain. 363

Opposed to my observations are those of Steiner^ who states that the frog deprived of the cerebrum does not take nourishment spontaneously or carry on willed movements, and that the region presiding over these functions is in the posterior portion of the cerebrum.

The operations on the optic thalami indicate that these, more than any other portion of the frog's brain contain the primary nuclei of the nervus opticus, and that this nerve joins their ventral, rather than their dorsal portions. The permanent reduction of voluntary movements and tactile sensibility noted after excision of these bodies may be regarded as an important deficiency phenomenon, and I assume that the principal reflex ganglion of the sense of touch is located in these bodies. The reduction in voluntary movements must be attributed more to the presence of a motor nucleus in these bodies than to a destruction of associative connections which they have with other motor ganglia, for this phenomenon lasted during two and a half months.

My results and conclusions in regard to the optic thalami strengthen those obtained by Schrader" who states that frogs deprived of the optic thalami show a decrease in voluntary movements and sensitiveness to touch and complete blindness. My observations and conclusions do not agree with those of Steiner who admits that there seem to be primary elements of sight in the optic thalami of the lizard, but show instead that this is contrary to what is found in the frog; and they do not agree with the results of Sellier and Verger'^ who found that in the dog injury to the optic thalami causes transient disturbances of the muscular and touch sensibility.

The operations on the optic lobe showed that the anterioventral quarters of these bodies are in close association with the retinal part of the optic thalami, but that the chief center of the

^ Steiner. Die Functionen des Central Nervensystems und ihre Phylogenese, Abtheilung 4.

2 ScHRADER, M. PJiuger's Archiv, 41, 75, 1887.

^ Sellier and Verger. H. C. R. Sodetes de Biologie, Mai, 522, 1898.

364 Journal of Comparative Neurology and Psychology.

optic nerve is not in the optic lobes. The fact that the sHght defect in sight disappeared two weeks after the operation, indicates that its cause was inhibitory stimuH. By excising the lobi and thalami optici unilaterally it was proved that the nervus opticus of the frog crosses completely at the chiasma. The abnormal retention of urine and the straight forced movements — always noted after excising the anterior and posterior half of the lobi optici respectively seemed to be caused by abnormal stimuli due to direct loss of association fibers. It indicates that the anterior part of the lobi optici is in close connection with the sacral urinary ganglia and the posterior, with some important motor nucleus.

My observations and conclusions regarding the optic nerve agree with those of Singer and Munzer\ Wiedersheim- . and ScHRADER, who found that in fishes, reptiles and amphibia the decussation of the second nerve seems to be complete, and that the centres of the nervi optici are located in the thalami optici. My results agree neither with those of Donaldson^, who states that the frog loses the power to avoid obstacles when the optic lobes are removed nor with those of Landois^ who says that destruction of the optic lobes in birds, amphibia and fish is followed by blindness.

The unsteady and weak jumps, flat position of the body and the tendency to abnormal hibernation, following bilaterial cerebellar excision, I believe to be due to the weakened condition of the limbs, to a loss of muscular tone, as well as to a loss of timely sequence or co-ordination and gradation according to the strength of the contractions, i. e., sensory ataxia. It seems that these deficiency phenomena are to be explained by the absence and modification of normal impulses proceeding from or going to the cerebellum. Anatomically, the explanation might

1 Singer and Muzner. Detikschriften der mathematik-natunvissenschaftlicheH Classe der Kaiserlichen Akademie der IVissenschaften, 55, 1888.

  • Wiedersheim. Vergleichende Anatomic der Wirbelthiere, and Edition,

335, 1886.

' Donaldson, H. H. American Textbook of Physiology, 708, 1897.

  • Landois, L. Textbook on Human Physiology, 805, Ig04.

LoESER, Functions of tJie Frog s Brain. 365

be offered that the neuron-complex, consisting of an afferent path to and an efferent path from the cerebellum and the ganglionic structures intervening, had lost important elements and that thus the function of the whole neuron-complex had been destroyed. This represents absence of normal impulses and may explain the defects after cerebellar excision. It is, however, impossible for the remaining neurons to make connections with other surrounding normal neurons and thus take up part of the functions formerly possessed by the desfroyed neuron-complex. There evidently was a slight trace of this modification of impulses present two months after cerebellar excision, since the position of the body and the condition of the limbs had improved two months after the operation. The fact that the condition of the body and limbs became worse on stimulation, probably indicates that the modified paths were fairly able to master the situation during rest but were not capable of doing so during excitation. It has been observed in many instances that stimulation brings out to their fullest extent the defects which are in part hidden by the vicarious activity of the modified paths during rest.

After unilateral cerebellar excision, the bilateral defects were confined to the homonymous side. Uni- as well as bilateral excisions of the cerebellum have shown that it does not possess the function of maintaining the equilibrium or any intellectual functions. The former function is located in the medulla, and the latter — the small amount that the frog possesses — probably in the scantily developed cerebrum and in the optic thalami.

My conclusions regarding the cerebellum do not agree with those of Steiner' who states that symmetrical removal of the brain down to the medulla in the frog produced no changes in the motor activity of the extremities.

Bilateral excision of everything anterior to the medulla shows that, although the croak and turning-over centers seem

' Steiner, J. Untersuchungen iiber die Physiologic des Froschgehirns. Braunschweig, 1885.

366 Journal of Comparative Neurology and Psychology.

to have associative connections with the anterior portions, they are located posterior to the cerebellum. The inhibitory influences noticed especially during the first two or three days after the operation, I attribute to the loss of association fibers. The centers for swallowing, the sense of equilibrium, and heart action seem to have only slight associative connections with these anterior parts. It appears that the greater portion of the functions performed exclusively and in a more specialized degree by the cerebellum of the dog is performed by the anterior third of the medulla of the frog. It was interesting to note after unilateral excision of the medulla, that the abnormal position of the spine and the rotatory movements were closely related in time of appearance, and also that the position of the body and legs were closely related to one another and to the two former phenomena. The rotatory movement around the sagittal axis was the first phenomenon to appear and also to disappear, whereas the sensory ataxia and the atonia became apparent only when the forced rotatory movements became less vigorous. I attribute the forced rotatory movements principally to the direct acute disturbance caused by the excision of an important component of the neuron-complex pertaining to the geotropic sense. The fading away of the forced rotatory movements accordingly would mark the disappearance of this acute disturbance owing to the function vicariously assumed by paths formed by the union of the intact neurons of the broken neuron -complex with other neurons surrounding them. When these new paths were fully formed and the acute symptoms had in part subsided, the atonia and sensory ataxia became more prominent.

The absence of abdominal respiratory movements on the homonymous side after unilateral medullar excision anterior to the calamus scriptorius proves that the centre for abdominal respiration is present in the medulla above the calamus and that its centrifugal fibers do not appreciably cross, moreover, that it does not control the laryngeal and narial respiratory movements, since these still persist bilaterally and normally. The loss of the eye reflex to pressure on the homonymous side is to

LoESER, Functions of the Frog s Brain. 367

be attributed to the excision of the origin or section of the fifth, the sympathetic, or the third nerve, or of all three. This phenomenon also proves that these nerves do not appreciably cross. I attribute the upward and outward position of the crossed eye to the fact that its superior oblique muscle is no longer innervated, since the nervus trochlearis crosses, and its nucleus, as well as the nuclei of all the nerves that govern the muscles of the eye on the operated side have been excised in the operation. Consequently, the superior oblique is the only muscle of the eye on the operated side that is still innervated and this fact explains the downward and upward position of the homonymous eye. Since this eye was covered by the nictitating membrane, it is evident that the downward movement of the eye is correlated with upward movements of the nictitating membrane. The persistence of the swallowing and croak reflex after unilateral medullar excision, shows that these are bilaterally quite independent in the frog and that one side may perform the function. This operation in conjunction with operation XVII proves that the sense of equilibrium is lost when the origin of the eighth nerve is destroyed.

The unilateral excision of the medulla anterior to the origin of the vagus group, in conjunction with the unilateral excision of the medulla posterior to the origin of this group, adds the fact that as long as the origin of the vagus grovip is intact, abdominal respiratory movements persist on the homonymous side. It is apparently easier to injure abdominal respiration by excisions posterior to the external origin of the vagus group, than anterior to its origin, owing probably to the fact that the deep origin of this group is more posterior than anterior to the external origin. The contracted pupil of the homonymous eye, noted after unilateral excision of the medulla posterior to the origin of the vagus group, was probably due to the destruction of the path of the sympathetic.

Unilateral excision of the middle third of the medulla in fishes, according to Hyde,^ removes the ganglia for respiratory

1 Hyde, I. H. American Journal of Physiology, 10, 5, 1904.

368 Journal of Comparative Neurology and Psychology.

movements, namely those of the seventh, ninth, and tenth nerves, and indicates that the narial and oral respiratory apparatus innervated by the seventh and ninth, can continue their respiratory movements independently of those of the abdominal ; moreover, that no appreciable crossing of the efferent fibers of these ganglia takes place. Furthermore, it corroborates the result found after unilateral excision of the medulla, namely, that the croak ganglia are located in the origin of the vagus group and are functionally independent, although probably closely connected with one another by association fibers.

The excision of everything anterior to the spinal cord corroborates a multitude of facts found in previous operations, and proves that co-ordinated movements do not cease with the removal of all the nervous structures anterior to the spinal cord. It moreover shows that apparently some inhibitory influence is taken away from the spinal cord with the removal of the structures anterior to it.

My experimental results on the medulla agree with Schrader\ who states that frogs cease to catch flies if the optic thalami are injured ; that the reactions on the rotating table are reflex in character, the stimulations acting upon the semicircular canals and passing to the centre by the acoustic nerve ; that frogs with injured medulla show a loose position of the feet when resting, and that their heads point toward the ground. Furthermore, that the medulla, down to the calamus scriptorius may be removed in the frog without destroying co-ordinated movements and that centres for respiratory movements, swallowing and croaking are in the medulla. My observations also corroborate Schrader's statement that after lesion behind the vagus group there is often complete inversion of the stomach and oesophagus, accompanied by prolapsus into the mouth, and that the frog after section of the medulla even behind the trigeminus group is still able to turn over on the ventrum when placed on its dorsum. Schrader also states that if the section

' Schrader, M. Pfluger's Archiv, 41, 75, iJ

LoESER, Functions of the Frog s Brain. 369

is behind the trigeminus group and if the brain anterior to this group is removed, every reaction on the turning table ceases.

The excision of everything anterior to and including the greater part of the brachial plexus, proves, as Schrader observed, that co-ordinated movements stop only because the origin of the nerves of the front legs is destroyed ; that after removal of the medulla, co-ordinated movements still remain, and that the more nearer the section approaches the origin of the brachial plexus, the more is co-ordination impaired.

Table of Operations and Chief Results.

I. Excision of one cerebral hemisphere.

(a) During the first three days the animal jumps slightly toward the crossed


(b) From four days to two and a half months, no deficiencies were noticed.

II. Excision of both cerebral hemispheres.

(a) During the first three days, bilateral slightly defective strength of the

muscles and somewhat subnormal skin reflexes.

(b) Later, to two and a half months, no deficiencies were observed.

III. Unilateral removal of the thalamus.

(a) During the first two weeks,

(1) Voluntary movements rare.

(2) Touch sensibility very defective.

(3) Almost total blindness on the crossed side.

(4) Crossed legs are weaker.

(b) From three weeks to two and a half months afterwards, (i), (2), and (4)

are slightly, but (3) is not at all improved.

IV. Bilateral removal of the thalami optici,

Causes these permanent deficiencies : (i) Almost complete bilateral blindness.

(2) Touch sensibility very much impaired.

(3) Voluntary movements very rare.

V. Unilateral removal of the lobus and thalamus causes,

(a) From one half hour to two weeks,

(1) Total loss of sight on the crossed side.

(2) Forced straight-ahead movements. .

(3) Abnormal retention of urine.

(4) Reduction in voluntary movements and touch-sensibility.

(b) From third week to two and a half months, (l) is unimproved, (2) is

much improved unless the animal is excited, {3) is normal, and (4) somewhat improved.

VI. Bilateral removal of the lobi and thalami. (a) From one half hour to three weeks,

(i) Total blindness on both sides.

370 Journal of Comparative Neurology and Psychology.

(2) Other symptoms as noted under operation V, but in a more marked degree, (b) From three weeks to two and a half months, (i) remains unimproved, (2) normal, (3) other symptoms as noted under operation V (b).

VII. Unilateral removal of the lobus opticus.

(a) From one half hour to two weeks,

(i) Trifling defect in sight on crossed side.i

(2) Forced straight-ahead movements and abnormal urine retention as noted after operation V (a).

(b) From two weeks to two and a half months,

(i) Sight is normal.

(2) The other two symptoms as indicated in operation V (b).

VIII. Removal of both lobi optici.

(a) From one half hour to two weeks,

(i) Trifling bilateral defect in sight.

(2) The other symptoms as noted under operation VII (a), but stronger.

(b) From two weeks to two and a half months,

(1) Sight is normal on both sides.

(2) The other symptoms as noted under operation VII (b).

IX. Removal of the anterior half of the lobi optici.

(a) From one half hour to two weeks,

(1) The same symptoms as after operation VIII (a), excepting that the

forced movements were absent.

(b) Later, to two and a half months, the animal showed nothing abnormal.

X. Removal of the posterior half of the lobi optici,

(a) From one half hour to two weeks, forced straight-ahead movements.

(b) Later, to two and a half months, the forced movements are present only

when the animal is much excited. XL Removal of both thalami optici and the anterior half of the lobi optici.

AllJ symptoms are noted under operation VI excepting that the forced movements were absent.

XII. Removal of the dorsal half of both thalami and lobi optici.

All symptoms as noted under operation VI but in a lesser degree, whereas sight was only slightly and temporarily (two weeks) aftected.

XIII. Bilateral removal of the cerebellum.

(a) From a half hour to three weeks,

(i) Flat position of the frog.

(2) Jumps infrequent, weak, and unsteady.

(b) Later, to two and a half months.

Muscular weakness and unsteadiness is marked only when the frog is excited ; the animal is unable to catch flies, although it sees them.

XIV. Unilateral removal of the cerebellum, (a) From a half hour to one week,

(i) Limbs of crossed side, especially the posterior one was abnormally far extended and abducted.

(2) Locomotion is always toward the homonymous side.

(3) Head was lower and almost touched the ground on the homony mous side.

LoESER, Functions of the Frog s Brain. 371

(4) Spine was bent in the thoracic region concave to the homonymous side, (b) From second week to two and a half months,

(1) Locomotion improved very little.

(2) All the other symptoms had considerably improved in one week ;

but were far from normal, and were characteristically present after stimulation two and a half months afterward.

XV. Bilateral excision of everything anterior to the medulla.

(a) From a half hour to three days,

(i) Croak reflex is difficult to elicit.

(2) Turning over reflex is inhibited.

(3) In addition, all symptoms as after operation VI and XIII.

(b) Four days to one month after operation,

(1) Collective symptoms of operation VI and XIII.

(2) Croak and turn-over reflex are normal.

(3) Swallowing, heart action and respiratory reflexes, and sense of

equilibrium are intact.

XVI. Unilateral Excision of the right side of the medulla anterior to the calamus scriptorius.

The observations pertaining to this operation necessitate recording at shorter intervals. I shall use the same number for an observation throughout, and after noting a phenomenon once, shall not mention it again until at the end, unless it changes.

A. Ten minutes after operating.

(1) No reaction with head or body movements on rotating table.

(2) Body, especially anterior portion and head, are lower on homony mous side than on crossed side.

(3) Legs on crossed side, especially the posterior one, are far extended

and abducted.

(4) Legs on homonymous side, especially anterior, are superflexed and


(5) Forced rotatory movements around the sagittal axis of body toward

the homonymous side are the sole movements.

(6) Abdominal respiration on the homonymous side is stopped.

(7) Narial and oral respiratory movements are bilaterally normal.

B. One half to five hours after operating.

(8) Eye reflex to pressure on the homonymous side is gone but is

normal on the crossed side.

(9) Crossed eye is turned upward and outward.

(10) Homonymous eye is turned downward and inward.

(11) Homonymous eye is turned upward synchronously with inspiration;

and downward with expiration.

(12) Nictitating membrane of homonymous eye is evidently beyond the

animal's control ; it covers the eye, especially when this moves downward.

(13) Croak reflex elicited only in a depressed degree.

(14) Swallowing reflex is normal.

(15) Spine and neck curved as after operation XIV.

372 Journal of Comparative Neurology and Psychology.

(i6) Tremor on both sides ; but strongest on the homonymous side.

(17) Weakness of muscles in proportion to tremor.

(18) Sight is perfect on both sides,

C. Twenty-four hours after operating.

(la) Rotating the table in clockwise direction, i. e., toward the operated side, the frog moves his head in anti-clockwise direction ; after rotating, the head is brought back beyond the sagittal axis of the body in clockwise direction.

(lb) Rotating in anti-clockwise direction elicits no head movements ; but a slight, circular movement of the body in an anti-clockwise direction is noted after rotating the table. (5) Is increased by the slightest stimulation.

D. Two days after operating,

(la) Same as after one day but in addition, sometimes a circular movement of the body in clockwise direction after rotating the table, (lb) Same as after one day. (5) Diminishing.

E. Three days after operating,

(2), (3), and (4) were temporarily normal during short intervals while the frog was resting.

F. Five days after operating,

(5) Frog rotates only when it attempts to jump.

G. Six days after operating,

(5) Frog, on jumping, sometimes lands on dorsum. H. Ten days after operating, (13) Is normal.

(i)' (')> (3)) ^^'^ (4) ^rs still far from normal. (5) On jumping, frog sometimes lands on dorsum and sometimes on homonymous side. I. No appreciable change in any symptoms from tenth to fifteenth day.

XVII. Unilateral excision of the right side of the medulla anterior to the origin of the vagus group.

(a) The same phenomena as observed after operation XVI ; excepting that

(b) Abdominal respiration is intact on both sides.

(c) The croak reflex is normal immediately after the operation.

(d) The bilateral apnoeic pauses in abdominal respiration, after rotating

movements around the sagittal axis, are pronounced.

XVIII. Unilateral excision of the right side of the medulla posterior to the origin of the vagus group to the calamus scriptorius.

Causes from first to fifteenth day —

(i) Slight impairment of the abdominal respiration on the homonymous side.

(2) Contraction of pupil on the homonymous side.

(3) Weakness of limbs on the homonymous side.

(4) Position of legs, body and head as observed after operation XVI,

but in a lesser degree. They improved during the first week but did not regain normal.

LoESER. Functions of tJic Frog s Brain. 373

XIX. Unilateral excision of the right middle third of the medulla including the origin of the vagus group.

(a) Phenomena which showed no change from first to the fifteenth day —

(') Abdominal respiration on the homonymous side is completely

abolished. (2) Pupil of this side is contracted.

(b) Phenomina as observed after operation XVIII,

(1) Position of head, legs, and body.

(2) Defect in locomotion.

XX. Bilateral excision of the middle third of the medulla, including the origin of the vagus group.

(a) Abdominal respiration is abolished on both sides, also, the croak, turn over, and swallowing reflex is gone.

(b) The head points into the ground.

(c) Urostylic prominence is gone.

(d) Complete inversion of oesophagus and stomach with prolapsus outside of

the mouth.

(e) Heart becomes feeble two hours after operating.

(f) Eight hours after operating, respiration of nares and mouth is exagger ated, due to asphyxia.

(g) Frogs died 8 hours after operating.

XXI. Excision of everything anterior to the spinal cord.

(a) Respiratory, croak, swallowing, heart-action, and turn-over reflexes are

gone entirely.

(b) Sense of equilibrium is lost.

(c) Eye reflex and sight is lost.

(d) Stimuli are answered sooner and with greater certainty than when higher

portions of the brain are intact.

(e) Heart action gradually ebbs away during the five hours which the frog

usually lives after the operation.

XXII. Excision of everything anterior to and including the anterior portion of

the spinal cord almost down to the origin of the brachial plexiffe.

(a) All phenomena of operation XXI.

(b) Co-ordinated movements fade away during the first two to three hours

after this operation.

(c) Three hours after operating, the fore limbs cannot support the body

any more.

(d) The frog is usually dead five hours after the operation.


By C. JuDSON Herrick.

Studies from the Neu7-ological Laboratory of Dt-nison University. No. XVIII^. With Forty Figures in the Text.


Section I. Introduction.

Section II. The Perii'Heral Gustatory System in Fishes.

Section III. The Central Gustatory System of Cyprinoid Fishes.

1 . Primary Gustatory Centers.

2. Secondary Gustatory Tracts.

(i) Descending Secondary Gustatory Tract. (2) Ascending Secondary Gustatory Tract.

3. Superior Secondary Nucleus and its Connections.

Section IV. The Central Gustatory System of Siluroid Fishes. Section V. Summary and General Conclusions.

Table of the Gustatory Paths in Fishes.

Mammalian Homologies of the Gustatory Centers of Fishes.

General Morphology of the Gustatory System. Literature Cited.

Section I. Introduction.

The conception of the nervous system as a mechanism for putting the organism into correspondence with the external environment and, in higher animals, for coordinating the reacting apparatus itself (internal environment) may be said to give the key to its evolutionary history. These two factors have given direction to the differentiation of the nervous system into somatic and visceral systems respectively and the further subdivision of each of these.

' This study was awarded the Cartwright Prize for 1905 by the Alumni Association of the College of Physicians and Surgeons, Columbia University, New York. It is published simultaneously in the Journal of Comparative Neurology and Psychology and the Bulletin of the Scientific Laboratories of Denison University, pages 375 to 456 of volume XV of the fournal being severally identical with pages 35 to 116 of volume XIII of the Bulletin.

3/6 Journal of Comparative Neurology and Psychology.

The labors of the students of nerve components have given us for the peripheral nervous system a paradigm or schema which seems to hold for all vertebrates, though with infinite variation in its details ; and it now remains to correlate these peripheral components with the central conduction paths so as to give a detailed knowledge of the whole course of each reflex pathway.

In attacking this general problem there are obviously two general lines of procedure open to us : — (i) beginning with the simplest brains we may work out exhaustively for each critical species in the phylogenetic series the conduction paths as completely as possible by monographic treatment of types and thus in the end approximate to a reconstruction of the phyletic history of the nervous system. (2) Or we may take each sensori-motor reflex system as the unit and trace its phylogeny through the series of types. This second method has the obvious advantage that one can start with the system in question in some type where it attains maximum development and, having arrived at a thorough knowledge of its anatomy and physiology here, it will be easier to read this schema backwards to the more primitive animals, as well as forwards in its further evolutionary modifications. It is hardly necessary to call attention to the fact that the human nervous system is the least favorable starting point for this sort of a research except for the neo-pallium and its appendages.

Each method has its advantages. The monographic treatment of type brains is really far more difficult, even in the lowest vertebrates, because of the difficulty in interpreting such simple undifferentiated pictures and analyzing a complex where there are few salient features. But nature has effected the analysis for us in some of the more specialized types by the hypertrophy of isolated systems ; and if, as sometimes happens, the other functional systems are in a primitive or reduced condition, we have a favorable point of approacii for a monographic study of the exaggerated functional system (cf Judson HerRICK, '03).

The purpose of this study is to make such a detailed analy

Herrick, Gustatory Paths in Fishes. ^j'j'j

sis of a single functional system of neurones — the gustatorysystem — in types where it reaches its maximum development and is obscured as little as possible by a high development of other systems. These conditions are fulfilled perfectly in the cyprinoid and siluroid fishes, whose brains are uncomplicated by any pathways leading to or from the cortex of the fore-brain and are in the main merely reflex mechanisms, but whose peripheral gustatory pathways are more highly developed than in any other vertebrates.

Section II. The Peripheral Gustatory System in Fishes.

As is well known, taste buds occur freely scattered over the mucous lining of the mouth and gills of nearly all fishes from the lips to the oesophagus. These are innervated by the VII, IX, and X cranial nerves. Similar taste buds, commonly called terminal buds, occur in the outer skin of some fishes and these also have been recently shown to be gustatory in function (JuDSON Herrick, '04). They are, in all cases where the innervation is known, supplied by the facial nerve. All taste buds in the pharynx and back part of the mouth, then, are supplied from the vagus and glossopharyngeus, those in the front part of the mouth, lips and outer skin from a root of the faciahs which apparently corresponds with the portio intermedia of human anatomy.

Associated with each of these roots are unspecialized visceral sensory fibers ending by free arborizations in the mucous membrane of the mouth cavity, these being very numerous in the region of the lower vagal roots and diminishing cephalad. The central connections of these two elements have not as yet been clearly differentiated, and both are provisionally designated the " communis system " of nerves by students of nerve components. With the unspecialized fibers we are not here concerned. The specialized communis fibers related with taste buds, either within or outside the mouth, and the ganglion cells from which they are derived will be termed the peripheral gustatory system of neurones.

378 Journal of Comparative Neurology and Psychology.

The peripheral gustatory system has recently been worked up both anatomically and physiologically in a number of fishes, most thoroughly in the siluroids, or cat fishes (Judson Herrick, '01 and '04). From the paper last cited we copy the accompanying figure (Fig. i), which gives the peripheral distribution of the cutaneous branches of the communis root of the facial nerve to supply the taste buds in the outer skin, in the horned pout, Ameiurus.

Fj^-. I. A projection of the cutaneous branches of the communis root of the right facial nerve in Ameiurus ?nelas, Jordan and Cope. Natural size.

The branches of this nerve which supply taste buds within the mouth are not drawn. All of the branches shown are gustatory in function. From the Bulhtin of the U. S. Fish Commission for 1902 (JUDSON Herrick, '04).

The general form of these buds of Ameiurus and their relations to the skin are shown in Figure 2. Practically all parts of the skin in these fishes are provided with taste buds supplied by this root, these same areas receiving also general cutaneous nerves for tactile sensation from other nerve roots, and the barblets being especially richly supplied with both sorts of nerve endings.

It has been determined by an experimental study of a series of both fresh water and marine fishes (Judson Herrick, '04) that the cutaneous communis nerves are gustatory in function and that this function is absent from the skin in species where these nerves and their sense organs are not developed. When edible substances or sapid solutions were brought in contact with these cutaneous sense organs the fishes would react by the

Hekrick, Gustatory Paths in Fishes.


appropriate movements to approach and seize the food and in the case of the horned pout, Ameiurus, it was found possible to differentiate a gustatory reaction from a tactile reaction and to prove that both senses possess a "local sign ;" i. e., both the tactile and the gustatory reaction can be localized in space.

Fig. 2. Section tlirough tlie skin of the top of the head of Amiurus nielas, showing a terminal bud. From \.\\tt Journal of Comparative Neurology. (JUDSON Hekrick, 'oi).

At d\% the dermis which is raised into a low papilla under the sense organ and whose center is pierced by the nerve for the organ ; cl., clavate cells of Leydig ; fiiuc, mucus cells of the epidermis.

The peripheral pathways for these motor responses are now well known and the problem of the present research may be thus stated : Given the terminal relations of the peripheral gustatory neurones and the peripheral motor neurones involved in the known gustatory reflex movements of these fishes, what are the central connections between them ?

Experiment has shown that the taste buds in the outer skin function in substantially the same way as those within the mouth, though there are certain obvious differences, particularly the fact that the stimulus if applied to the cutaneous taste buds

380 Journal of Comparative Neurology and Psychology.

usually calls forth a movement of the whole body toward the food object, while this is unnecessary in the case of a gustatory stimulus received within the mouth. This is the probable reason for the differentiation of a special center in the medulla oblongata for the reception of the facial root of the communis system, the lobus facialis, or " tuberculum impar, " in those species which possess large numbers of terminal buds in the outer skin.

Fig. J. Two views of the l>iain of the buffaUj fish, Cnrpiot/cs vclifcr (Kaf.), (i) from above, (2) from the right side. X 2. After C. L. IIerrick.

The vagal lobes (/,. vg.) are relatively larger than in the carp and, with the overhanging cerebellum, completely conceal the facial lobe. In the lower figure the branches of the trigemino-facial root complex are marked Vi.d., Vi.7>., V2, VII, and the root of the auditory nerve VIII. Between the latter and the caudal tip of the optic lobe is a well-defined protuberance ventrally of the cerebellum. This is the lateral portion of the superior secondary gustatory nucleus f^Rindciilaiotcii, Mavsek). Caudad of this, immediately dorsally of the VIII root, is the small lobus lineae lateralis (Iohnston) from which the tuberculum acusticum and cerebellar crest extend caudad beneath the vagal lobe. In the upper figure, the median shaded part of the roof of the optocoele is membranous, the optic lobes being widely divaricated by the enormous valvula cerebelli. In both figures the membranous roof of the fore-brain is dissected away to show the lobules of the basal ganglia, and the olfactory buliis are cut away.

With the exception just noted, it is probably safe to assume that there will be found broad lines of similarity between

Herrick, Gustatory Paths in Fishes. 381

both the peripheral and the central gustatory paths in all vertebrate types and that fishes like the carp with enormous hypertrophy of the gustatory roots of the VII, IX and X nerves may safely be used as guides to point the way for further researches upon the gustatory paths of higher vertebrates where the system is less easily analyzed. Such being the case, our problem assumes a measure of importance when we remember that the central gustatory pathway is at present almost totally unknown in all vertebrates, including man, and that even the peripheral pathway is still in dispute among students of human anatomy.

It has been already pointed out (Judson Herrick, '04) that the siluroids (cat fishes and horned pouts) and the larger cyprinoids (carps, suckers, etc.) present a striking similarity to each other and contrast to other teleosts in both the peripheral and the central nervous systems, and that these features center about the gustatory pathways. In teleosts generally the gustatory pathway from the lining of the mouth in the branchial region, entering the brain by the communis roots of the IX and X nerves, and the gustatory pathway from the anterior part of the mouth and from the outer skin, entering by the communis VII root, terminate together in a single cerebral enlargement, the lobus vagi. But in these two groups of fishes we have a second tuberosity, as mentioned above, developed for the communis root of the facialis to provide for the unusually large number of taste buds in the outer skin supplied by this nerve.

In the cyprinoids, particularly the carp-like forms, both oi these lobes are enlarged, correlated with a high development of the taste buds in each of the corresponding peripheral regions, but particularly in the vagal region to supply the remarkable collection of taste buds on the palatal organ. In the siluroids, on the other hand, there is no hypertrophy of the gustatory organs in the branchial region and the vagal lobes, accordingly, are essentially similar to those of other teleosts. But the highly developed cutaneous gustatory organs innervated by the facialis have called forth an enlargement of the facial lobe greater than that of cyprinoids. Since in the siluroids all of the taste buds of the outer skin are supplied by the communis root of the

382 Journal of Comparative Neurology and Psychology.

facialis (Judson Herrick, '01) and since the palatal and other branches of the communis VII root which supply taste buds within the mouth are not enlarged, as compared with other fishes, it is clear that the enormous communis VII root and the lobus facialis in which it terminates are related mainly to cutaneous taste buds in these fishes.


Ft[i^. 4. Dorsal view of the brain of the yellow cat fish, Leptops olivaris (Raf.), X 2.

The olfactory bulbs are cut off ; also the membranous roof of the fourth ventricle, exposing the facial lobes (Z. F.) and vagal lobes (Z. V.). This ventricle is bounded behind by a transverse ridge containing the commissura infima Halleri (c. i.) and the commissural nucleus of Cajal. The tuberosity laterally of the cerebellum and facial lobe is the lobus lineae lateralis (/. /. /.), which is greatly enlarged and entirely conceals the superior secondary gustatory nucleus.

Section III. The Central Gustatory System of Cypri NOiD Fishes.

We shall now proceed with a description of the gustatory pathways in a selected series of teleosts, beginning with the larger cyprinoids where it attains its maximum development. The results here obtained will be controlled by an equally careful examination of the brains of siluroid fishes (particularly Ameiurus), whose gustatory centers differ greatly in detail from those of the carp-like forms. These minor differences will serve to bring out more clearly the points of fundamental resemblance, which are very striking.

The end-station of the gustatory neurones of the first order (peripheral neurones) is the nucleus of origin for the neurones of the second order, giving rise to secondary gustatory tracts, and these in turn to tertiary tracts. The nomenclature of these tracts and centers offers almost insuperable difficulties and some new terms will have to be introduced and old ones more nar

Herrick, Gustatory Paths in Fishes. 383

rowly defined. A summary of the nomenclature here employed for the gustatory system of these fishes is given in tabular form in Section V, to which frequent reference should be made.

/. Primary Gustatory Centers.

The general topography for the vagal and facial lobes of cyprinoid fishes has often been described, and the reader is referred for the details to the classic paper of Mayser ('82). and also to the figures of Bela Haller ('96), whose descriptions, however, I do not in all things confirm. I have found little in the extensive work of this author ('98) on the brains of Salmo and Scyllium which sheds further light on the secondary connections of the vagal lobe. However, I do not profess to have mastered the contents of this unsystematic and obscure paper, though I have dilligently studied it. Nor is there anything of importance from our present point of view in the recent dissertation on the vagal lobes of cyprinoids by Groth ('00), whose purpose was merely to test the accuracy of some of Haller's observations on nerve anastomoses in these fishes.

The vagal lobe of the cyprinoids, as compared with that of the siluroids and the teleosts generally, represents an enlargement of both the sensory and the motor centers of the vagus and glossopharyngeus. This enlargement is correlated with the development of the curious palatal organ in the mouth of these fishes, the sensory fibers being derived from the taste buds which cover the surface of this organ and the motor fibers going out chiefly to the small muscles which permeate its interior. The " motor layer " of the vagal lobe, from which these motor fibers are derived is a dorsal extension of the nucleus ambiguus. The latter nucleus has the typical position and relations, supplying the branchial musculature, and is not commonly regarded as a part of the vagal lobe.

We shall describe, so far as the material at command permits, the conduction pathways in the vagal lobe of the larger cyprinoids on the basis of sections of adult and young brains cut in various planes and stained by Delafield's haematoxylin and the methods of Weigert-Pal and Golgi. The illustrations

384 Journal of Comparative Neurology and Psychology.

given throughout the remainder of this paper are all camera drawings of single sections, except Fig. 20, which is a composite, and the schemata, Figs. 38,- 39 and 40.

First a few words by way of general orientation. The vagal and facial lobes are dorsal structures in the oblongata, but they do ?iot represent, as commonly taught, the morphological continuation of the dorsal horns of the spinal cord. These are represented in the spinal V tract and its associated substantia gelatinosa Rolandi, their ventro-lateral position in the vagus region of these brains being due to crowding by the more mesially placed gustatory centers. The latter centers, therefore, if morphologically related to anything in the spinal cord, must represent spinal structures lying dorsally of the canalis centralis and beneath the floor of the dorsal fissure. In the oblongata this fissure opens out and its floor becomes greatly extended to form the membranous roof plate (His), or velum medullare posterior, betieath which the lobes in question are developed

(cf. JUDSON HeRRICK, '99, p. 2 1 3).

Figure 5, taken midway of the vagal and facial lobes of the carp, illustrates how these structures are superposed upon the great longitudinal conduction paths which constitute the chief landmarks of morphological relationship. The spinal V tract, whose substantia gelatinosa at this level is reduced to a mere vestige, lies ventrally of both the sensory and motor vagus roots. Crowded into the space ordinarily occupied by the substantia gelatinosa Rolandi are the great longitudinal secondary gustatory paths, ascending and descending. M esially of these is the substantia reticularis, shown by Cajal and others to be the continuation of the ventro-lateral funicle of the spinal cord and to be composed of short paths, mostly sensory fibers of the

Fig. §. A transverse section taken through the middle of the vagal and facia} lobes of an adult carp stained by the method of Weigert-Pal. X i6.

The layers of the facial lobes and the vagus roots are designated at the right, a. s. X., ascending secondary gustatory tract from the vagal lobe ; com. r. VII, communis (gustatory) root of the facialis entering the facial lobe ; d. s. VII, descending secondary gustatory tract from the facial lobe ; Im., lemniscus (laterales Langsbiindel, MAYSER);jr/. V. spinal root of the trigeminus.

Herrick, Gustatory Paths in Fishes.


3 01














o _





-1 (Q





— c










» Si

■» o




3 •










386 Journal of Comparative Neurology and Psychology.

third order designed to distribute sensory excitations over a large field of motor nuclei of the oblongata. Close to the raphe in the median Ime is the fasciculus longitudinalis medialis, likewise composed largely of short paths and chiefly motor. Between this fasciculus and the substantia reticularis, there is ventrally the continuation of the ventral funicle of the spinal cord and dorsally the lemniscus (funiculus lateralis, Fritsch; laterales Langsbiindel, Stieda and Mayser).

The leninisais is very complex. It clearly is composed in the main of crossed ascending fibers from the primary sensory centers of the spinal cord and oblongata to the mid-brain. These fibers correspond closely with the lemniscus lateralis or lateral fillet of mammals. The lemniscus medialis, or direct pathway to the cerebral cortex, as found in mammals, is of course not present here and this whole ascending path I shall term simply lemniscus. It probably receives fibers from the whole length of the spinal cord with a large accession of similar fibers from the funicular nuclei and enormous numbers of fibers from the tuberculum acusticum. It receives no appreciable number of fibers from the vagal lobes or other visceral sensory centers and therefore may be considered a somatic sensory secondary tract. It terminates in the mid-brain beneath the tectum in the protuberance into the optocoele which is so characteristic of the teleosts and is termed the torus semicircularis (nucleus lateralis mesencephali, Edinger). This body, in fact, seems to be primarily the end-nucleus of this tract. Whether this bundle also contains descending fibers I do not know, but am sure that if present they are relatively few in teleosts. Such descending fibers would of course have to be excluded from the designation lemniscus.^

Some of the smaller cyprinoids (e. g., Notropis) exhibit but little enlargement of the vagal lobes. In the gold fish {Carassius aiiratus) the vagal lobe is greatly enlarged, but the

' The reader will note that this tract is designated lemniscus on account of its partial homology with the tract of that name in the mammals, and that it has nothing to do with the so-called lemniscus of Mayser and others, or tractus tectobulbaris et spinalis.

Herrick, Gustatory Paths in Fishes. 387

sensory roots of the IX and X nerves penetrate at once to its deeper layers and then turn outward to end near the surface much as will be described in Ameiurus. In Cycleptus the lobe is larger and the sensory roots penetrate but little before they spread out to end in a superficial layer of large cells which stain very pale in haematoxylin. And in the carp (Fig. 5), whose vagal lobes are still more enlarged, the greater part of the sensory roots do not penetrate, but spread out over the whole outer surface of the lobe (constituting Mayser's first layer), thence to pass in separate strands directly into a thick layer exhibiting alternate bands of crowded small cells and dense neuropil (Mayser's second layer).

This will be termed the layer of secondary Jieurones. Sections stained by Delafield's haematoxylin show within it seven concentric bands of cells and neuropil which are more or less sharply distinguishable, of which the two outer are of greatest importance. The first is a thin layer of crowded cells of relatively large size which take up a very pale stain — the chief secondary gustatory neurones. Immediately internal to and among these cells are smaller very closely crowded cells whose nuclei take up a very intense stain. These will be termed the smaller gustatory neurones. The other bands contain chiefly minute cells with deeply staining nuclei and occasionally larger, irregularly shaped cells among them, while the inner border of this layer is a band of neuropil containing a few large cells which take a very pale stain. Sections stained by the method of Pal show sensory root fibers entering this band at its inner border. These root fibers will be termed the deep communis root of the vagus, as distinguished from the larger superficial root (Fig. 5), The fibers of the deep root turn outward to end in the second layer along with those of the superficial root.

This layer of secondary neurones is very wide in Cyprinus and is traversed radially by unmedullated and delicately medullated fibers for the deeper layers of the vagal lobe. The large pale cells forming its outer boundary are the most important and characteristic cells of the lobe, for they give rise to the

388 Journal of Comparative Neurology and Psychology.

great secondary gustatory tracts which leave the lobe for distant centers. The narrow layer of smaller cells adjacent to these may share this function, but their neurites are mostly shorter. The small cells of the other bands of this layer have short neurites which terminate for the most part within the lobe and are therefore called intrinsic secondary gustatory neurones.

Immediately internal to the seventh band of the layer of secondary neurones is Mayser's third layer, the layer of secondary gustatory fibers, which pass ventrally to constitute the "secondary vagus bundle" of Mayser. These fibers are of small caliber with thin sheaths which take a very pale stain by the Weigert method.

Internally from the secondary gustatory tract is the motor layer (Mayser's fourth), containing large cells which give rise to motor root fibers of the IX and X nerves. The thick ependyma containing large blood vessels is Mayser's fifth layer. Some of its supporting elements are shown in Fig. 6.

The endings of the peripheral neurones which compose the superficial layer of root fibers of the vagal lobe of the cyprinoids are richly impregnated in many of my Golgi preparations. Those of the deeper layers impregnate with much more difficulty. They are seen in Fig. 6. The superficial fibers end by extensive terminal arborizations which ramify widely among the dendrites of the chief secondary gustatory neurones, forming a very dense neuropil in the more superficial parts of the layer of secondary neurones, though some penetrate through the whole thickness of this layer. These endings are shown in Figs. 6 and 15 and somewhat imperfectly in Fig. 8. Fig. 6 shows that the slender strands of root fibers pass from the superficial to the deep layers of root fibers throughout the whole extent of the vagal lobe.

The most caudal fibers of the vagus constitute the descending or spinal root of this nerve, which passes a very short distance caudad to terminate in the grey centers associated with the commissura infima Halleri and funicular nuclei as Cajal ('96, p. 46) describes the endings of the fasciculus solitarius in the nucleus commissurale of the mouse.

Herrick, Gustatory Paths in Fishes. 389

The general arrangement of the chief secondary tieiirones may be understood by reference to Fig. 7, drawn from a sagittal section of a young carp, where these cells are richly impreg

Fig. 6. Transection through the vagal lobe of Catastomus co/ttmersoni. GoLGi method. X 40 The section illustrates the endings of the superficial and deep layers of root fibers of the vagus and the position of the chief secondary gustatory neurones (c.s.g.n.). An intrinsic neurone of type II is shown with dendrites in relation with the termini of the deep root. Two neuroglia cells of the superficial series are drawn and several of the deeper series (some of the latter having been drawn from an adjacent section), ^.t'., blood vessel; w.r., motor root of the vagus ; s.g.t., secondary gustatory tract ; sp. V, spinal V tract.

390 Journal of Comparative Neurology and Psychology.

nated. The layer of root fibers is not impregnated in this preparation, but its relations are suggested in Fig. 8 and more clearly in Fig. 6. In Fig. 8 root fibers are stained in the ventral part

f^-'-'-secl/// / / '"/^'-motoy la\jer

Fig 7. Section through the vagal lobe of a young carp, 5 cm. long. GOLGI method. X 40.

The section is approximately sagittal, but strongly inclined so that, while the caudal end (at the left of the figure) is near the median line, forward it passes toward the right. It includes, therefore, the motor nucleus of the right vagal lobe and the right side of the facial lobe, only the extreme caudo-ventral border of which is indicated. The preparation gives a typical view of the chief secondary gustatory neurones of the vagal lobe. One of the smaller secondary neurones («) is shown at the extreme ventral (lower) side of the vagal lobe (cf. Fig. II). A few scattered neurones of the deeper portions of the layer of secondary neurones are imperfectly impregnated. A few thick fibers are seen to pass from the layer of secondary neurones into the motor layer to arborize widely among its cells. Their origin could not be determined, but is probably trom some of the deeper cells of the layer of secondary neurones. Asc. sec. X., ascending secondary gustatory tract ; mot. X r., motor root of the vagus.

of the vagal lobe, but only secondary neurones in its dorsal part. In preparations where both are impregnated together the peripheral layer of secondary neurones presents an indescribably intricate complex of interlacing fibers. Fig. 9 shows a single one of the chief secondary neurones of Fig. 8 drawn separately on a larger scale.

Hekrick, Gustatory Paths in Fishes.


These neurones are densely crowded in radial arrangement over the entire periphery of the vagal lobe and are of very peculiar form. They are unipolar with irregularly elongated cell body and dendrite springing from the inner end. The

fuberculum acusticum


sec )(

Fig. 8. Sagittal section through the vagal lobe of a young carp. GoLGl method. Drawn from the same series as Fig. 7, but farther toward the right side, so as to cut tangentially the extreme right borders of the vagal and facial lobes. X 40.

The figure illustrates a few gustatory neurones of the vagal lobe and (imperfectly) endings of the gustatory root fibers, also two neurones of the deeper portion of the layer of secondary neurones whose processes reach the motor layer (cf. Fig. 7). These cells appear to be of the same type as the large cell lying in the layer of secondary tracts in Fig. 13. The neurone in the facial lobe is sketched in from the section lying next mesially, It occupies the area intermedia of the facial lobe and spreads out over the whole of the lateral aspect of the latter, its neurite not being shown. It is evidently of the same type as the one figured in Fig 18 at a.

peripheral end of the cell body is drawn out to a point which seems usually to reach the external limiting membrane of the vagal lobe. In these young fishes the external layer of root fibers is relatively thin. In older carp this layer is much thicker and there is no well defined external limiting membrane. The bushy dendrite breaks up at once into a very dense mass of

392 Journal of Comparative Neurology and Psychology.

exceedingly fine branchlets among which the arborizations of the root fibers end. From one or more of the terminal branches of this dendrite a very delicate neurite arises and goes by the shortest path directly into the layer of secondary tracts by which it leaves the vagal lobe to enter the longitudinal secondary gustatory tracts of the oblongata, meanwhile taking up a

Fig. g. A chief secondary gustatory neurone from the vagal lobe of the carp — the neurone designated a on Fig. 8. X 250. The black mass at the left of the neurone is precipitate deposited in the external limitating membrane of the vagal lobe.

very delicate medullary sheath. In its passage through the layer of secondary tracts it sometimes gives collaterals into the motor layer of the vagal lobe (Fig. 10).

Upon casual examination these chief neurones might suggest the appearance of neuroglia elements. The latter, however, are of totally different form (Fig. 6) and there is no opportunity for confusion. The gustatory neurones have exceedingly fine and densely branched dendrites, while the branches of the neuroglia are thicker and of the peculiar mossy or pulverulent appearance so characteristic of the supporting elements. The neuroglia, moreover, as a rule impregnates with a dark brown instead of black color.

Besides the chief gustatory neurones just described, there are closely associated with them just internally the smaller gustatory neurones shown in Figs, ii and 12. These have the same general form and arrangement, but are much simpler. The cell body is smaller and usually rounded and the dendrite less profusely branched. The slender neurite arises from the tip of the dendrite and enters the layer of secondary tracts.

Herrick, Gustatory Paths in Fishes. 393

Some of these fibers or their collaterals ente/ directly the motor layer and there arborize. Others end by simple arborizations in the substantia reticularis and others turn caudad in the descending secondary communis tract. Whether any go cephalad

Fig. 10. Transverse section through the vagal and facial lobes of the spotted sucker, Minytrema melanops (Raf.) from a fish 8 cm. long. Golgi method. X 40 On the left at a a single chief secondary gustatory neurone is impregnated whose neurite sends collaterals into the dorsal part of the motor layer of the vagal lobe. On the right the area marked by the crosses is filled with a rich impregnation of similar cells which are not sketched in. At 3 is a single secondary neurone of the type shown in Fig. I2, and at c imperfect impregnations of deeper secondary neurones of the intrinsic type. On the right one cell of the motor layer is impregnated whose dendrites reach secondary gustatory tracts partly within the vagal lobe and partly in the substantia reticularis.

aj-if. J- .^.A, ascending gustatory tract ; com.r.VII, communis (gustatory) root of the facialis ; desc.s.g.i., descending secondary gustatory tract, entering substantia reticularis ; /./.w., fasciculus longitudinalis medialis; Int., lemniscus; mot.r., motor root of the vagus ; s.g.t., secondary gustatory tract from the vagal lobe ; sp. V, spinal V tract; sup.r., superficial gustatory root of the vagus.

394 Journal of Comparative Neurology and Psychology.

in the ascending secondary tract I have not determined.

These neurones clearly are of the same type as the chief gustatory neurones, but their neurites have a shorter course as a rule. Cells of this type in my preparations are impregnated chiefty in the more superficial parts of the layer of sec

facial lobe

motor /ai/er

subs! retiCulans'^A'-^


Fig. II. Transverse section through the vagal and facial lobes of a young carp 5 cm. long. The section is obliquely inclined with the left side a little farther caudad than the right and the dorsal side much farther cephalad than the ventral. GoLGi method. X 40 The section passes through the extreme cephalic end of the vagal lobe. None of the chief gustatory neurones are impregnated in this preparation, the superficial cells drawn being of the smaller variety. Among them are two intrinsic neurones of the vagal lobe. There is impregnated also a single large neurone in the dorsal part of the substantia reticularis one of whose dendrites reaches the layer of secondary gustatory tracts and whose neurite passes toward the ventral commissure., fasciculus longitudinalis medialis ; /;//., lemniscus ; mot.r., motor root of the vagus ; s.g.t., secondary gustatory tract ; s. V, spinal V tract.

ondary neurones ; but sections stained with haematoxylin show very numerous nuclei of the same appearance throughout this layer and occasional Golgi impregnations show that they are in fact scattered less freely throughout the thickness of the layer. There are frequent impregnations, usually very imperfect, of larger cells of different kinds in the deeper parts of this layer.

Herrick, Gustatory Paths in Fishes. 395

as shovv^n in Figs. 7, 8, 10 and 13 ; but I have not been able to trace the connections of all these types of cells. There are many indications that they reach the motor layer of the vagal lobe (Fig. 14) and substantia reticularis of the oblongata immediately below the vagal lobe. In most of these cases the dendrites spread irregularly among the other cells of this layer

\ \



\ I I

Fig. 12. A single neurone of the smaller gustatory type from the right vagal lobe of the carp. Drawn from the same series as Fig. ii, two sections farther caudad. GoLGl method. X 2-0.

The main axis of the dendrite is directly prolonged to form the neurite. Terminal arborizations of the gustatory root fibers occur among these dendrites. The broken line represents the outer border of the vagal lobe.

and the processes which penetrate the motor layer do not differ essentially from the others. There is one type of cells, however, among these whose dendrites extend tangentially over a very wide area (Fig. 16, b and Fig. 18, b). In one case (Fig. 13) a neurite from a cell was seen to enter the secondary gustatory tract; but its further course is unknown. Fig. 13 also shows one of the smallest of these intrinsic cells whose thorny dendrites are sharply contrasted to the very delicate much branched neurite (neuropodium), giving the cell the characteristic form of GoLGi's type II.

This is apparently the most frequent type of these intrinsic cells, though the neurite is rarely impregnated in recognizable form. Fig. 6 illustrates a different form of this type of cell, with cell body and dendrites among the endings of the deep layer of root fibers and neuropodium between this layer and the layer of secondary tracts.

396 Journal of Comparative Neurology and Psychology.

Stannius ('49, p. 82) has shown that mechanical stimulation of the vagal lobe of the carp produces movements of the intrinsic muscles of the palatal organ. Since the nucleus ambiguus of the cyprinoids does not differ from that of other

mot. fayer

Fig. ij. Portion of the vagal lobe from a transverse section of the brain of Minytrema melanops (Raf.) 8 cm. long. GoLGi method. X 1S7.

This is drawn from the same series as Fig. lo, the sections being adjacent. For ease of orientation the position of the neurone marked a on Fig. 10 is designated a in this sketch. The large neurone at the bottom of the figure in the layer of secondary gustatory tracts sends one of its two chief dendrites outward into relation with the arborizations of the peripheral gustatory neurones, the other is cut off. The neurite is not shown. The intrinsic neurone to the left of it is a characteristic type II cell. The intrinsic neurone at the top of the figure is drawn in from the adjacent section. Its neurite enters the secondary gustatory tract {s.g.t.)

teleosts, in which it is known to innervate the branchial musculature, it is extremely probable that the cells of the motor layer of the vagal lobe give rise to the nerve fibers for these intrinsic muscles, for these two structures (the palatal organ and motor layer of the vagal lobe) are always developed proportionately with each other.

Herrick, Gustatory Paths in Fishes.


mot. )(. r.

Weigert sections show very clearly that fibers of the motor roots of the vagus and glossopharyngeus arise not only from the nucleus ambiguus but also from the motor layer of the vagal lobe, as has been well known since Mayser.

My GoLGi preparations very rarely show good impregnations either of the cells of the motor layer or of the nucleus ambiguus. The latter have the form and arrangement which have been often figured, the main dendrites spreading out in the substantia reticularis and reaching the extreme ventro-lateral border of the oblongata. Other smaller dendrites spread out within the nucleus itself and the neurite usually springs from the base of the main dendrite. The dendrites of the cells of the motor layer in some cases spread out within this layer, where they come into relation with the radial fibers which enter it from the layer of secondary neurones. Others reach out into the layer of secondary tracts and into the substantia reticularis (Fig lo), in both of which places they reach collaterals from the secondary gustatory tracts.

It appears that in no case do the peripheral motor neurones come into direct relation with the peripheral gustatory neurones, but at least one intermediary element is always interposed. These elements may be the intrinsic neurones of the vagal lobe, for short reflex arcs, or the chief secondary gustatory neurones for more complex reflex connections reaching beyond the vagal region of the brain.

Fig. 14. Transverse section of the vagal lobe of the right side of a young carp, 5 cm. long. GoLGl method.

X soIllustrating two intrinsic secondary neurones of the vagal lobe whose processes extend inward to reach the motor layer, s.g.t., secondary gustatory tract ; sp. V, spinal V tract.

398 Journal of Comparative Neurology and Psychology.

The substantia reticularis is clearly the chief medium of communication between the sensory and motor centers for simple reflex paths here, as in higher vertebrates. This reticular substance receives gustator\- fibers both from the secondary gustator}- tracts and from the intrinsic neurones of the vagal lobe. We shall see beyond that it also receives fibers from the facial lobe.

fobus fac/a//&W


Fig. jj. Section through the right lobus glossopharyngei of a young carp 8 cm. long. Golgi method. X 40 The section is approximately transverse, but very oblique, so that the right side and the ventral surface are much farther cephalad. It cuts the lobus IX at its widest part and shows a strand of the communis root of the facialis (a) arching over its dorsal side and apparently entering both this lobe and the cephalic part of the lobus vagi. Terminal arborizations of the communis root of the vagus are al50 shown ending in the layer of secondary neurones of the lobus vagi.

com.r. VII, communis ^gustatory) root of the facialis; n. IX. mot., motor nucleus of the IX nerve ; r.IX.mot., motor root of the IX nerve; r.IX.s., sensory root of the IX nerve ; r.X.s., superficial gustatory vagus root ; s.g.t., secondary gustatory tract ; sp. V, spinal V tract ; V, ventricle.

Herrick, Giistator)' Paths in Fishes.


Fig, i6. Sagittal section

throagh the brain of a voung

carp, 5 cm. long. Golgi

method. X 40.

The section is somewhat

oblique, the caudal end being

arboHyati^r,- ^f ,\. ' farther mesial. The terminal

nated Tw ^""'"-' '°°' °' ^^^ ^"^^^ — -^ richlv impreg on of JZ^eTT r ^\°"° •" ^'^ ^^^^^"^^ '^^P^^^- P-' °f the'lobus IX,

lieinaboutThe " ^'^^^^^r"" ^" ^'^ °^°^°^ ^^' -^^-^- ^hese neurone he m about the transverse level indicated in Fig. x; and at the level of r.IX.s.n.

rZt ! f '"' " ^"^"^ ^^ ^"'^tatorv root fibers of the IX nerve

nat d n X T / 'l^^^'^"'^" °' ^'^ ^^^^^ ^^^ ^ ^^^ °- ^^-^^-^^ ^-ig acu ticL V. / ; T""' ^ ^'^ ^-' --'^^^ "P °^ tuberculu;

acusticum , tr.sp.l ., spmal V root.

400 Journal of Comparative Neurology and Psychology.

Before proceeding with the description of the long secondary gustatory tracts from the vagal lobe we shall describe the internal structures of the glossopharyngeal and facial lobes, as the long tracts from all these centers can best be described together.

Between the vagal and the facial lobes of cyprinoids is a small tuberosity which, as pointed out by B. Haller ('96, p. 93 and Fig. 12), receives the glossopharyngeus nerve. It may be termed the lodus glossopharyngei. It is very distinct in forms like the gold-fish, Carassius, whose glossopharyngeus nerves are far separated from the vagus, the sensory IX nerve entering its dorsal side and the motor IX its ventral. It appears in the carp at a level somewhat cephalad of that shown in Fig. 5 along the line of union of the vagal and facial lobes. In this type it appears to receive from behind some fibers from the vagus roots ; at any rate the IX and X roots are somewhat confused in this region. In both the carp and the gold-fish it receives from in front filaments of the communis root of the facialis. These filaments are very clearly shown arching over the dorsal side of the lobus IX in an oblique Golgi section (Fig. 15, a)\ a certain part of the facial root apparently also reaches the cephalic end of the lobus vagi. It would be interesting to learn whether these facial fibers which separate from the facial lobe to end in connection with gustatory fibers from the mouth cavity coming in by the IX and X nerves are derived from the palatine and other facial branches which also supply taste buds within the mouth.

The lobus glossopharyngei has essentially the same connections as the lobus vagi. Fig. 16 gives an impregnation of two of the intrinsic neurones of the extreme cephalic part of the lobus IX in sagittal section of the brain of a young carp. The sensory (gustatory) fibers of the IX nerve end among the dendrites of these neurones. The neurone at the extreme left does not show its neurite, but the neurite of the one marked a is shown completely, arborizing in the motor IX nucleus, thus completing the simplest gustatory reflex arc for the IX nerve.

The structure oi t\\Q facial lobe (tuberculum impar, "lobus

Herrick, Gustatory Paths in Fishes.


trigemini," Mayser) of the carp is considerably simpler than that of the vagal lobes. It is in general similar in internal organization to the facial lobes of Ameiurus, though its cells are of quite different form and it differs externally in that there is a complete fusion in the middle line of the right and left lobes with no clearly defined external or internal boundary between them.

^' facial lob

Fig. ly. Transverse section through the facial lobe of Minytrema 7nelanops (Raf.). GoLGi method. X 4° The section passes through the cephalic part of the facial lobe and includes the extreme cephalic end of the vagal lobe and the caudal end of the tuberculum acusticum. It shovi^s five chief secondary gustatory neurones at the surface of the lobe and internally at a an imperfect impregnation of a cell of the intermediate zone of the same type shown in Fig. 18, a.

asc.s.g.t., main ascending secondary gustatory tract ; f./.iii., fasciculus longitudinalis medialis ; VII, communis (gustatory) root of the facialis ; r.IX. mot., motor root of the glossopharyngeus ; 5tr. F///, secondary fibers from the tuberculum acusticum to ventral commissure ; s.g.t. VII, secondary gustatory tract from the facial lobe ; sp. V, spinal V tract ;, tuberculum acusticum.

The center of the lobe is filled with very small intrinsic neurones arranged in dense clusters or rosettes of from 10 to 100 cells each. In Fig. 5 the clear spaces in the neuropil of the

402 Journal of Comparative Neurology and Psychology.

mot. I/// r. \ f.Lrn. /


n mo

'^..s.g.l/// tr

facial lobe "^'^'-^^^




Fig. iS. Horizontal section through the base of the vagal and facial lobes of Minytrema i/ielanops (Raf.). GoLGl method. X 40.

The section is slightly oblique, the left side being farther ventral. It passes through the ventral part of the facial and vagal lobes, illustrating a neurone (a) of the intermediate zone of the facial lobe. The dendrites spread throughout the lateral part of the lobe and the neurite passes laterally to enter the substantia reticularis of the oblongata (s.r.), here coming into relation with dendrites of the motor VII nucleus, as shown in the figure, and with other motor nuclei. The section shows the position of the dorsal tip of the motor VII nucleus (n.mot. VII. d.) and the motor VII root springing from it. The motor VII dendrites which are shown ramifying in the substantia reticularis [n.mot. VII.v.) spring from cells of the motor VII nucleus lying ventrally and laterally of those figured and out of the plane of this section. The neurone h is drawn in from the section lying next ventrad in the same series.

Her RICK, Gustatory Pat J is in Pis lies. 403

facial lobe represent the positions of some of the larger of these rosettes. Superficially there is a layer of chief gustatory neurones of the same general form as those of the vagal lobe (Fig. 17), whose neurite springs from the tip of a dendrite and passes at once into the secondary gustatory tract. The peripheral gustatory neurones enter by the communis root of the facial nerve (= dorsal geniculate root of the trigeminus, Mayser = portio intermedia of Wkisberg in human anatomy) and pass back to the lobe as a huge tract of heavily meduUated fibers close to the median line, constituting the pre-vagal portion of the fasciculus communis of Osborn (^ fasciculus solitarius of mammals). Entering the lateral border of the facial lobe, it spreads out in many strands in the superficial layer among the dendrites of the chief gustatory cells (cf. Figs. 5, 16 and 20). These endings are thick, loosely branched arborizations of wide extent. One of the simpler forms of these endings is impregnated in Fig. 18 en the right side.

There is no well defined motor layer of the facial lobe corresponding with that of the vagal lobe, but the ventro-lateral portion along the line of attachment to the oblongata is of different structure from the rest and will be termed the intermediate zone, or 7incleus intennediiis facialis. It contains rather large neurones whose dendrites spread widely through the whole lateral part of the facial lobe and whose neurites reach the substantia reticularis (Figs. 18, a and 17, a and 8). Here they reach directly the dendrites of the motor VH and less immediately those of the motor V, IX and X nuclei. This is the main pathway for short reflexes from the facial lobe to the oblongata. The absence of a motor layer in the facial lobe which gives rise directly to motor root fibers, like that in the vagal lobe, is to

a, neurone of the intermediate zone ; asc.s.g.t., main ascending secondary gustatory tract ; com.r. VII, communis (gustatory) root of the facialis ; f.Lm., fasciculus longitudinalis medialis ; n.mot. Vll.d., dorsal part of motor VII nucleus ; n.viot. VII. v., ventral part of motor VII nucleus ; mot. VII. r., motor root of facial nerve ; r. IX. sen., sensory root of the IX nerve; r.X.viot., motor root of the vagus ; r.X.s., gustatory root of the vagus ; s.g.t.IX-\-X, secondary gustatory tract from X and IX lobes ; s.g. VII. L, secondary gustatory tract from the facial lobe ; s.r., substantia reticularis ; sp. V, spinal V tract.

404 Journal of Comparative Neurology and Psychology.

be correlated with the fact that its gustatory impressions come mainly from the outer skin and so give rise to more general body mov^ements, such as turning and seizing, rather than to movements of the intrinsic musculature of the palatal organ, for which the root fibers of the motor layer of the vagal lobe are mainly designed. The substantia reticularis is known to be related not only to the cranial motor nuclei of the branchial or visceral type, as mentioned above, but also to contain cells whose neurites reach the fasciculus longitudinalis medialis and other paths with somatic motor connections (see Cajal, '96, p. 129 and Fig. 16), so that we have here a very direct mechanism for producing the movement of eye-muscles and trunk-muscles necessary for locating and approaching a sapid substance which has been perceived by contact with taste buds of the outer skin.

2. Secondary Gjistatory Tracts,

The secondary connections of the vagal, glossopharyngeal and facial lobes are of two general types: — (i) short paths, by way of the intrinsic secondary neurones or of collaterals of the neurites of the chief cells, directly to the motor layer of the vagal lobe and to the substantia reticularis of the oblongata ; and (2) long paths to regions above and below the oblongata, arising from the chief gustatory neurones. The connections of the first type are diffuse and largely unmedullated ; they have been briefly described above. The connections of the second type are compact well defined tracts of meduUated fibers which will be termed the ascending and descending secondary gustatory tracts.

A word of further explanation may be necessary here in justification of the term gustato}y as applied to these tracts. We have seen above that the communis roots of the VII, IX and X nerves in fishes generally contain fibers from taste buds in the mucous membrane of the mouth or in the outer skin and also fibers which end peripherally by free arborizations unrelated to any specialized sense organs — undifferentiated visceral endings. In view of the fact that these types of fibers, so distinct peri

Herrick, Gustatory Paths in Fishes. 405

pherally, have not as yet been clearly separated at their central terminations, these roots must for the present retain the name "communis system," pending the time when they can be separated into specialized (gustatory) and unspecialized visceral components. With the secondary pathways and centers, however, the first step in this analysis may already be taken, thanks to the advantages of the comparative method.

In forms like the mammals, where the gustatory system is reduced, the analysis of these two elements in the fasciculus solitarius and its associated grey and secondary tracts will probably be impossible save by a degeneration method. The same applies to the amphibia, where the term fasciculus communis was first applied to the homologous structure. In fishes, again, we have some forms with reduced organs of taste, where the problem offers the same difficulties as in the cases cited, and also forms with enormous hypertrophy of the peripheral gustatory system with no appreciable change in the unspecialized component. In the latter cases whatever enlargement of the primary and secondary centers has taken place may clearly be assigned to the gustatory and not to the general visceral system. This is the case presented by the cyprinoid and siluroid fishes particularly, where the whole of the facial lobe and its connections and the greater part of the vagal lobe and connections have clearly arisen in response to the demands of the enormous peripheral gustatory system. While the diffuse connection of these centers with the substantia reticularis is known to have a parallel in the similar connections of the fasciculus solitarius of mammals and therefore doubtless in part pertains to the unspecialized visceral component, the long secondary paths which we here term the secondary gustatory tracts are highly developed only in those fishes possessing elaborate peripheral gustatory systems. This seems to constitute sufficiently good evidence that these long paths are mainly and perhaps wholly gustatory in function.

The neurites of the chief secondary gustatory neurones of the vagal and facial lobes, as we have seen, pass out in thick bundles to the ventro-lateral border of the oblongata, where they

4o6 Journal of Comparative Neurology and Psychology.

turn to take a longitudinal course parallel with and often almost completely enclosing the spinal V tract. This secondary tract is largely composed of small and feebly medullated fibers, so that in Weigert sections it appears paler than the spinal V and the other great longitudinal conduction paths of the oblongata. Part of its fibers turn caudad and part cephalad. We therefore have descending and ascending secondary gustatory tracts (tractus gustus secundus descendens et ascendens) from both the vagal and facial lobes.

(i) Descending Secondary Gustatory Tract.

In the carp, as shown by Weigert sections of the adult, the descending tracts from both the vagal and facial lobes are very extensive. From the facial lobe they constitute a massive aggregate of medullated fibers in the "secondary vagus bundle" complex. But the descending tracts from the vagal lobe are largely unmedullated fibers scattered through the substantia reticularis grisea of the oblongata.

The descending secondary facial tract arises, as in Ameiurus, chiefly from the cephalic portion of the lobus facialis. The fibers pass ventro-lateral to gather in a compact bundle mesially and dorsally of the spinal V tract, farther caudad enclosing this tract on all but its external aspect (Fig. 5, d.s. VII). There is no obvious addition of descending fibers to this descending bundle in the region of the lobus vagi. It can be easily distinguished from the ascending secondary gustatory tract from the vagal lobe (Fig. 5, a.s. X) by the fact that the latter tract is much more feebly medullated at this level.

In the caudal part of the vagal lobe the motor zone is greatly thickened as is also the substantia reticularis grisea, which lies ventro-Iaterally of it and in intimate relations with the layer of secondary gustatory tracts of the vagal lobe, the descending secondary facial tract and the spinal V tract. This whole area contains descending secondary vagal fibers. These are chiefly very fine unmedullated or feebly medullated fibers passing from the layer of secondary gustatory tracts of the vagal lobe caudad and mesad through the substantia reticularis grisea

Herrick, Gustatory Paths in Fishes.


Fig. ig. Longitudinal section through the left vagal lobe of a young carp, 5 cm, long. GoLGi method. X 40.

The section is approximately horizontal, but strongly inclined so that the left side is farther dorsal. It shows the descending secondary gustatory tract [desc. s.g.t.) from the vagal lobe to the region of the nucleus funiculi. At the periphery of the vagal lobe in the regions marked by the crosses there is a very rich impregnation of cells among which the secondary fibers arise, which are omitted so as not to confuse the drawing, since they are too densely impregnated to permit an accurate determination of their relations to the fibers. These cells 'are partly chief secondary neurones like those of Fig. 7, and partly the smaller type as shown in Fig. ii. A very few fibers of the ascending secondary gustatory tract (asc.s.g.i.) from the vagal lobe are shown. This tract appears voluminously a few sections ventrad in the position indicated by the arrow. The median line is indicated by the position of the left and right fasciculi longitudinales mediales {/.L and/.r.). The lemniscus on both the left (//«./.) and right (/m.r.) sides is shown ; also the position of the funicular nucleus and commissural nucleus of Cajal (

4o8 Journal of Comparative Neurology and Psychology.

in diffuse formation to reach the region of the nucleus funiculi and commissura infima Halleri (Figs. 19, 1 1). The indications are that it springs mainly from the smaller secondary neurones of the lobe (cf Fig. 11), though I have no conclusive evidence that the chief secondary cells do not also participate in its formation.

In the region of the funicular nuclei the substantia reticularis grisea, related above with the vagal lobe, enlarges. This I term the infenor seco?ida)y gustatory nucleus (nucleus gustus secundus inferior). It receives the diffuse tracts from the vagal lobe just described, and also the greater part of the descending secondary facialis tract. A portion of the latter tract, however, continues into the spinal cord caudad of the funicular nuclei in the ventro-lateral tracts.

This inferior gustatory nucleus is intimately related with the termini of the spinal V tract, funicular nuclei, commissura infima Halleri and commissural nucleus of Cajal. It is clearly a coordination center between gustatory and tactile sensory impressions. In fishes with feebly developed gustatory system it is of insignificant size. Its great development in the cyprinoids and, as we shall see, in the siluroids is correlated with the known feeding habits of these fishes. It has been shown (Judson Herrick, '04) that fishes with taste buds in the outer skin taste with these organs and habitually localize the food stimulus partly by the sense of taste and partly by tactile stimulation of the same cutaneous areas. It may safely be inferred that the descending secondary gustatory connections of these fishes are especially adapted to serve this correlation of the senses of taste and touch as employed in localizing peripheral stimuli in the search for food and to call forth the bodily movements necessary to seize the food when its position has been determined.

The detailed description of these connections involves the discussion also of the secondary connections of the general cutaneous (somatic sensory) system of neurones and must be reserved for a later contribution.

(2) Ascending Secondary Gustatory Tract,

This tract, which takes the greater part of the fibers aris

Herrick, Gustatory Paths in Fishes. 409

ing from the chief secondary gustatory neurones of both the vagal and facial lobes, is one of the most distinctive features of the teleostean brain. In the two groups here treated, particularly, it attains enormous dimensions, constituting the "Sekundare Vagus-Trigeminusbahn" of Maysek. whose excellent description ('82, p. 318) is fully confirmed. It is included in the "system y" of the descriptions of Goronowitsch ('88, and 'q6). The curious notion of B. Haller ('96, p. 92) that this is a descending tract from the cerebellum and other higher cerebral regions to the sensory centers in the vagal lobe need not detain us, for he gives no evidence for such a view. The common designation of this tract as "secondary vagus bundle" is inadmissible, since in all fishes it contains secondary glossopharyngeus and facialis fibers also, the secondary facialis fibers preponderating in siluroids. Moreover, there are other secondary vagus tracts for general cutaneous and other types of fibers, besides the descending and short gustatory paths described in the preceding section.

The ascending or central gustatory tract takes its position ventrally and mesially of the spinal V and descending secondary facialis tracts and partly enclosing them (Fig. 5, a.s.X.). It ascends to the level of the superficial origin of the sensory trigeminus root from the oblongata, where it turns mesially and dorsally to enter its own terminal nucleus (superior secondary gustatory nucleus. Fig. 20). The axial portion of this nucleus was termed by Mays er "Rindenknoten" and by Johnston ('01) secondary vagus nucleus. It is included in the "Uebergangsganglion" of Mayser's descriptions, this latter group of cells being very complex and not as yet fully understood. Its homologies are considered in the final section of this paper.

It lies in the dorsal part of the isthmus ventrally of the line of fusion of the valvula cerebelli with the body of the cerebellum. Typically it lies near the median line and projects into the aquaeductus, but in some cyprinoids it is so greatly enlarged as to reach also to the lateral surface of the brain at the isthmus and there project as a distinct tuberosity between the caudal end of the tectum opticum and the tuberculum acusticum

410 Journal of Comparative Neurology and Psychology.

(Fig. 3, lower figure). It lies cephalad, dorsad and mesad of the motor V nucleus, the long dendrites of some of whose cells pierce it and come into relations with some of the termini of the gustatory fibers (Figs. 2i and 23). The motor V nucleus is very large in these fishes and extends far cephalad and dorsad in intimate relations with the termini of the secondary gustatory tract. The dendrites of these motor cells spread out widely in every direction and are not directed ventro-laterad so generally as in the case of the motor nuclei of the VII, IX and X nerves. In fact, the greater part of them take the opposite course, as figured, into the caudal part of the secondary gustatory nucleus and the adjacent substantia reticularis grisea.

This fact, together with others of similar import, suggests the morphological interpretation of this secondary terminal nucleus ; viz. , that it is merely a specialization from the substantia reticularis. Cajal ('96, p. 128) has confirmed the view

Fig. 20. Parasagittal section through the brain of the spotted sucker, Minytrema melanops (Raf.). Golgi metlrod. X 12.

The sketch is designed to illustrate the course of the ascending secondary gustatory tract and the connections of its terminal nucleus. The plane of the section is slightly oblique so that the caudal end and the ventral side are nearer the median line than are the cephalic and dorsal borders. The figure is a composite, made by outlining one section with the camera lucida and filling in the details from this section and the three sections of the same series on each side immediately adjacent, omitting irrelevant detail. The features introduced are schematized as little as possible. The whole course of the ascending secondary gustatory tract from the facial lobe is shown. The origin of the tract from the vagal lobe lies farther lateral.

b., tract between secondary gustatory nucleus and n. lateralis valvulae ; com.h., commissura horizontalis, Fritsch ; coin.r. VII, communis (gustatory) root of the facialis; desc.sec.X, descending secondary gustatory tract from the vagal lobe ;/"./.;«., fasciculus longitudinalis medialis ;, lateral lobule of inferior lobe (hypoaria, C. L. Herrick) ; inf.l0b.7n., median lobule of inferior lobe {mammillare, C. L. Herrick) ; n.cort., nucleus corticalis, Fritsch ; n.IX, motor nucleus of the glossopharyngeus ;, nucleus lateralis valvulae ; n.rot., nucleus rotundus, Fritsch ; w. J/"., nucleus subthalamicus, C. L. Herrick; n.V., motor nucleus of the trigeminus; n.VII, motor nucleus of the facialis; r.X.s., sensory root fibers of the vagus; sec.gust.t., ascending secondary gustatory tract from the facial and vagal lobes ; tr.Lb., tractus lobo-bulbaris ; tr.t-c, tractus tecto-cerebellaris ; tr.t.lob., tractus tecto-lobaris, Johnston (commissura ventralis, C. L. Herrick). The area marked n./unicttli conx.&in'i also the inferior secondary gustatory nucleus.

Herrick, Gustatory Paths in Fishes.


412 Journal of Comparative Neurology and Psychology.

of KoLLiKER that the substantia reticularis alba is composed of sensory bundles of the second order and that the substantia reticularis grisea is a sensory field of the third order, designed to distribute sensory excitations from the V, IX and X nerves (and also, it should be added, from the VII nerve) over a large

cerebellar tracts

Fig. 21. Portion of a sagittal section of the brain of Minytrema melanops (Raf.). GOLGI method. X 4° The section is quite oblique, the cephalic end and the dorsal side being inclined toward the median line. It passes through the ascending secondary gustatory tract at the point where it enters its terminal nucleus under the cerebellum (the cephalic end of the figure being at the left). The spinal V tract [ is cut close to its superficial origin from the periphery. A single neurone of the most dorsal part of the motor V nucleus is impregnated, one of its dendrites crossing the mesial side of the gustatory tract to ramify in the most ventral and caudal part of the layer of chief tertiary gustatory neurones, a single one of which is impregnated. The tertiary gustatory path does not lie in the plane of this section (cf. Fig. 23).

field of motor nuclei of the oblongata. This definition, it will be observed, carries also the secondary gustatory nucleus, save that in addition to direct relations of the secondary termini with the motor nuclei, like the trigeminal connection just described, there is here, as we shall see, a much more extensive development of tertiary neurones for connections of a higher order in

Her RICK, Gustatory Paths in Fishes. 413

the floor of the thalamus (central tertiary tract). The dorsal and mesial position of this derivative of the substantia reticularis is easily explained by the topographic features of the isthmus. Here the ventro-lateral region is occupied by the great conduction paths between the oblongata and the mid-brain — the tractus tecto-spinalis, tractus lobo-spinalis, lemniscus, etc. The dorso-lateral region is occupied by the tuberculum acusticum and its cerebellar connections. The enlarged secondary gustatory terminal is prevented from growing ca\idad by the great cerebellar crest and tuberculum acusticum. It must, therefore, grow upward, inward and forward into the optocoele. In this position it appears typically in all teleosts. But when still more enlarged, as in cyprinoids, further growth in this direction being prevented by the valvula cerebelli, which is also very large in these fishes, it is forced to grow outward until it appears as a superficial eminence cephalad of the tuberculum acusticum and dorsally of the great ventro-lateral conduction paths just referred to (Fig. 3).

The secondary gustatory nucleus does not, however, comprise the whole of the substantia reticularis grisea of this region of the isthmus. For caudo-mesially of this nucleus at the level where the secondary gustatory tract enters it from the lateral side of the oblongata is another considerable cellular area which represents a less highly specialized portion of the same sensory field. This also borders the ventricle, in some types forming a considerable projection into it from the lateral wall immediately caudad and ventrad of the commissure of the secondary gustatory nuclei. Its anomalous position is brought about by the same forces which were discussed above in connection with the secondary gustatory nucleus. It is in very intimate relation with the motor V nucleus which lies ventro-laterally of it and it apparently is the chief medium of communication between the various sensory pathways and that nucleus. Its neurones, in other words, are like the others of the substantia reticularis grisea in being of the tertiary sensory type and discharging into a motor field, in this case the V nerve. It may therefore be termed the substantia reticularis grisea ttigemini.

414 Journal of Comparative Neurology and Psychology.

The position of this area of the carp is indicated in Fig. 22. It is reached, as shown in the figure, by a slender branched process of one of the chief tertiary neurones, which may be a dendrite, but more probably is a collateral neurite, though it does not spring from the main neurite (;/), which enters the tertiary tract separately. The neurone marked a lies in the adjacent section and is drawn enlarged in Fig. 25. From the base

Fig. 22. Section taken through the superior secondary gustatory nucleus of the carp. GoLGl method. X40 The section is approximately transverse, but strongly inclined so that the dorsal and right sides are father caudad. A single chief tertiary gustatory neurone is shown, whose neurite («) enters the tertiary gustatory tract. Its main dendrite passes out of the plane of the section (cf. Fig. 25). A much more slender process is completely impregnated, running dorsad and caudad to reach the substantia reticularis grisea trigemini {s.r.g.). At a is shown the position of the neurone drawn in Fig. 25, which lies in the section adjacent to the one here drawn. The secondary gustatory nucleus is bordered on the dorsal and mesial sides by the fibers of the tertiary tract.

of its neurite is given off a collateral, only a part of which is shown, which is probably of the same type as the one here figured. Contiguous sections of the same series show fibers passing from this area of substantia reticularis directly into the axis of the secondary gustatory nucleus and there arborizing, which probably represent dendrites of the cells of the substantia reticularis which are not impregnated.

Herrick, Gustatory Paths in Fishes. 415

These fragmentary data are sufficient to show that we have in addition to direct connections of dendrites of the motor V nucleus with secondary gustatory fibers, a similar but indirect functional connection via the substantia reticularis grisea trigemini. The numerous other connections of the latter area need not now concern us.

One of my Golgi preparations of the carp shows dendrites of a very large neurone lying in the vicinity of the nucleus of origin of the IV nerve sending dendritic branches ventrad into the region caudad of the commissura ansulata and other large branches farther caudad and laterad into the same portion of the secondary gustatory nucleus which is reached by dendrites of the motor V nucleus. The impregnation is so imperfect that it is impossible to be sure whether this neurone belongs to the motor nuclei of the eye muscle nerves or to the fasciculus longitudinalis medialis or to some other neighboring structure.

J. Superior Secondary Nucleus audits Connections.

The superior secondary gustatory nucleus presents, broadly speaking, the same general arrangement as the primary endstation in the vagal and facial lobes. That is, the secondary neurones end in relation with two types of tertiary neurones (i) intrinsic neurones, filling the interior of the nucleus ("Rindenknoten," Mayser) and (2) the chief tertiary neurones in a dense layer around the periphery. The chief difference between the arrangement of the primary and secondary end-stations lies in the fact that the latter is connected with its fellow of the opposite side by a broad commissure, the commissure of the secondary vagus nuclei of Johnston. As was recognized by Mayser, this commissure contains fibers of at least two sorts (i) neurites of the intrinsic tertiary neurones, (2) terminals of a portion of the secondary gustatory tract. From (2) it follows that the secondary tracts end partly on the same side and partly in the secondary nucleus of the opposite side, the former portion being much the larger. The uncrossed portion ends by free arborizations within the secondary nucleus and also extensively in the cortical

4i6 Journal of Comparative Neurology and Psychology.

portion among the chief tertiary cells (Fig. 23). The details of the endings of the crossed fibers I have not been able to observe, but there is every reason to believe that they are essentially similar.

I have no complete impregnations of the intrinsic commissural neurones. The evidence is that their delicate, feebly medullated neurites gather in the interior of the nucleus and terminate after crossing among the dendrites of the chief tertiary neurones of the opposite side. The layer of these chief cells envelopes the secondary nucleus on all sides except where it is interrupted by the secondary and commissural tracts. The neurites of these cells pass directly outward and become medullated as they enter the chief certiary tract for the inferior lobe. Just external to the layer of tertiary neurones the gustatory nucleus is encapsuled by a dense layer of heavily meduUated nerve fibers. These are partly the tertiary fibers, but chiefly cerebellar tracts. The position of the secondary gustatory nucleus in the isthmus is such that the cerebellar penduncles, p.issing downward from the cerebellum and valvula, almost entirely envelop it. It is this peculiarity which suggested to Mayser the name "Rindenknoten." Surrounding this fibrous capsule is an aggregate of nuclei with very diverse connections to which Mayser applied Stieda's name, "Uebergangsganglion."

This region is perhaps the most intricate and difficult of analysis in the teleostean brain, as it contains several large centers and numerous important tracts, both meduUated and unmedullated, all crowded into a very small space. The dorsal part of the isthmus is more than usually crowded in these types, not only by these great gustatory centers, but also by the enlarged cerebellum and valvula and their associated tracts.

The broad line of fusion of the valvula cerebelli with the torus semicircularis (colliculus) is occupied by a very dense mass of cells and meduUated and unmeduUated tracts whose relations are only imperfectly shown in my preparations. The whole complex, which lies mesially of the nucleus lateralis mesencephali (Edinger), or colliculus, I term provisionally the nucleus lateralis valvulae. It is, apparently, the most cephalic

Herrick, Gustatory Paths in Fishes.


part of the "Uebergangsganglion" of Stieda. This group of cells fuses behind with the dorsal and cephalic border of the secondary gustatory nucleus, but is separated from it more or less definitely by the meduUated vertical cerebellar tracts which encapsule the secondary gustatory nucleus. Sections stained by the method of Pal show large tracts of fine fibers with ex

c^re bellum


/obus inf.

Fig. 2j. Portion of sagittal section of the brain of a young carp, 5 cm. long. GoLGi method. X 4°.

The ventral side of the preparation is below and the cephalic end at the left. The section includes the superior secondary gustatory nucleus ("Rindenknoten," Mayser) and shows the origin and course of the tertiary path to the inferior lobe. A single neurone of the latter type is impregnated completely (a) save for the terminal arborization. Mingled with the neurites of these cells are those of the tract between the cerebellum and the inferior lobe (tractus lobo-cerebellaris, Edinger). Among the terminals of the tertiary tract are the cells of origin of the tractus lobo-bulbaris et spinalis (tr.l.b.); l.m., lemniscus ; tr.t.l., tractus tectolobaris ; tr.t.s., tractus tecto-spinalis.

ceedingly delicate medullary sheaths passing dorso-ventrally between the n. lateralis valvulae and the cerebellum and valvula (Fig. 24), and Golgi sections show diffuse unmedullated tracts taking the same courses (Fig. 20). Whether the medullated and the unmedullated fibers spring from neurones of the same type, I have not been able to determine. The neurones shown

41 8 Journal of Comparative Neurology and Psychology.

Fig. 24. Transection through the brain of a large carp just cephalad of the superior secondary gustatory nucleus. Stained by the method of Weigert-Pal. X 20.

The section passes through the most cephalic portion of the commissure of the secondary gustatory nuclei, and illustrates the relations of the nucleus lateralis valvulae to this nucleus and to the cerebellum. The trochlearis (n.IV.) decussates just cephalad of the gustatory commissure and in passing to its super

Hekrick, Gustatory Paths in Fishes. 419

by the Golgi sections are like the cells of the i^ranular layer of the cerebellum and probably are of that type, receiving afferent cerebellar impulses and transmitting them to the cortex cerebelli. Whether they receive their impulses from the adjacent gustatory nucleus or from other sources, my preparations do not show with certainty, but probabl}' partly from the former.

There is also a broad unmedullated conneciion between the ventral part of the secondary gustatory nucleus and the cephalic part of the n. lateralis valvulae. running cephalad through the vertical cerebellar tracts (designated by b in Fig. 20). Horizontal sections of the brain of Catastomus show numerous delicately medullated fibers running between the whole cephalic face of the gustatory nucleus and the n. lateralis valvulae. These connections lend further support to the belief that the n. lateralis valvulae is in part a gust Uory intermediary station for the cerebellum. This nucleus extends cephalad to the point where the valvula severs connection with the Hoor of the mesencephalon and here a large tract passes between its cephalic end and the nuclei of the third nerve and of the fasciculus longitudinalis which lie mesially of it. It is no doubt this connection which led B. Haller to state ('98. p. 522) that the "Uebergangsganglion of Fritsch and Mayser is an "upper motor oculomotorius nucleus." The nucleus lateralis valvulae

ficial origin become embedded in the most cephalic part of the secondary gustatory nucleus, the latter being represented in the figure by the neuropil surrounding n.IV. Sections immediately caudad show the gustatory nucleus occupying the entire space from near the median line to the lateral surface of the brain and from the level of the commissural fibers and tr. lobo-bulbaris {tr.l.b.) up to the valvula cerebelli.

cb., ii>^ . , cl>^ . , c/>^ . , cerebellar tracts; co»i.s.g.n., commissure of the secondary gustatory nuclei ; y./.w., fasciculus longitudinalis medialis ; g.III, fibers arising from the granule cells of the nucleus lateralis valvulae and passing dorsally into the lateral lobe of the valvula cerebelli ; ^.J., tertiary gustatory tract for the inferior lobe appearing as oval bundles among the strands of the cerebellar tract cb^.; Im., lemniscus (fasciculus lateralis); n.IV., root of the fourth nerve; n.l.m., the extreme caudal end of the nucleus lateralis mesencephali (torus semicircularis, colliculus); H.I.V., nucleus lateralis valvulae ; -f., a large blood sinus between the lateral lobes of the valvula ; tcrt.opt., tectum opticum ; tr.l.b., tractus lobo-bulbaris ; tr.t.h.c, tractus tecto-bulbaris ot spinalis cruciatus; tr.t.b.r., tractus tecto-bulbaris et spinalis rectus.

420 Joiuiial of Comparative Neurology and Psychology.

has other connections (with the substantia reticularis of the oblongata, optic thalamus, etc.) which I have not yet fully worked out. Its homologies are discussed briefly beyond. Another part of the "Uebergangsganglion" which serves as a gustatory reflex center is the nucleus already designated as substantia reticularis grisea trigemini. There remain several considerable collections of cells in the "Uebergangsganglion" which may have gustatory functions, but their discussion would involve a more extensive analysis of the mesencephalon than we can here attempt.

A typical illustration of the chief tertiary gustatory neurones in the cortex of the secondary gustatory center of the carp is seen in Fig. 25. The dendrite plunges into the axial portion of the nucleus ("Rindenknoton," Mayser) and there spreads out widely among the termini of the secondary gustatory fibers (cf Figs 22, 23). The neurite passes downward to enter at once the tertiary gustatory tract for the inferior lobe of the same side, where it takes up a dense medullary sheath and so is rarely completely impregnated in Golgi preparations. These meduUated fibers mingle immediately with similar ones of the cerebellar tracts (lobo-cerebellaris and others) which encapsule the secondary nucleus and which likewise pass into the inferior lobe, so that I have found it impossible in Weigert sections to effect the separation of the two classes of fibers except at the origin of the tertiary gustatory tract from its nucleus. I have, fortunately, secured a sufficient number of complete Golgi impregnations to show the origin of the tract without uncertainty. Its entire extent save the terminal arborization is shown in Fig. 23.

The combined tract passes ventrad and somewhat cephalad to enter the caudal border of the inferior lobe and spread throughout the whole of the lateral lobule of this organ. Whether any of these fibers reach the median lobule my preparations do not definitely determine, but apparently not. (The terms lateral and median lobules are used in the sense defined by David ('92). The lateral lobules are the hypoaria and the median lobules the mammillary bodies of C. L. Herrick ('92) ).

Herrick, Gustatory Paths in Fishes. 421

The terminal arborizations of this tract in the inferior lobe of Ameiurus shown in Fig. 37 probably belong to the gustatory


The tractus lobo-hdbaris arises from the whole of the lateral lobule of the inferior lobe and its fibers mingle with those of the tertiary gustatory tract, the two tracts, however, being well separated where they leave the lobe (Fig. 20, 23). The tractus lobo-bulbaris is the main path for descending impulses from the inferior lobe (gustatory and other reflexes) and was termed in Mayser's descriptions ('82, p. 319) -'Nervenfaserbahn ^. "

sec. gustatory nuc


tertiary gustatoru tr

Fig. 2j. One of the chief tertiary neurones of the superior secondary gustatory nucleus of a young carp. GoLGl method. X 187.

For the relations of this neurone to the surrounding structures, cf. Fig. 22, which is drawn from the section next caudad. The dendrites spread out in the mesial part of the secondary gustatory nucleus. The neurite at once enters the tertiary gustatory tract for the inferior lobe and either passes out of the plane of the section or becomes medullated and so fails to impregnate.

The fact that it is larger (as mentioned by Mayser) in cyprinoids and the peculiarities of its course in the oblongata make it probable that it is the chief motor return path for the higher gustatory reflexes. This of course is not its only function, for it clearly serves as the chief path for all descending impulses from the inferior lobe to the oblongata and cord, and therefore carries olfactory and other reflex impulses in addition to gustatory.

42 2 Journal of Comparative Neurology and Psychology.

The course of the tractus lobo-bulbaris as it leaves the inferior lobe is clearly shown in some of my Golgi preparations. In Weigert sections this tract is so confused at its origin with the fibers of the tractus lobo-cerebellaris and tertiary gustatory path that it cannot be clearly distinguished at first. Its fibers throughout their course are very fine and feebly medullated. The tract passes up from ^he inferior lobe and arches around the ansulate commissure to turn caudad along the inner border of the lemniscus. Upon reaching the level of the secondary gustatory nucleus it turns laterad over the dorsal surface of the lemniscus (Fig. 24) and takes up a position ventrally of the secondary gustatory nucleus (Fig. 20, tr.l.b.). Farther back it becomes partially enveloped by the ascending fibers of the secondary gustatory tract. The latter tract in cross section is crescent-shaped and the concavity of the crescent, which faces inward, is occupied by the descending fibers of the tractus lobobulbaris. This tract is thus placed along the dorso-lateral border of the substantia reticularis, which, as we have seen, is the chief avenue of communication with the motor nuclei of the oblongata. The tractus lobo-bulbaris diminishes caudad, breaking up in the substantia reticularis and so puts the gustatory center in the inferior lobe into relation with the peripheral musculature. It doubtless extends into the spinal cord, but it is impossible to distinguish its fibers from the others in the substantia reticularis alba farther back than the vagal lobes.

Section IV. The Central Gustatory System of Siluroid


As we have seen above, the anatomy and physiology of the peripheral gustatory system of the cat fishes, or horned pouts, is now well known. The cutaneous portion of the system is here hypertrophied, while the buccal portion is about as in other fishes. This is the explanation of the fact that the cat fishes possess enlarged facial lobes, while the vagal lobes are of the typical teleostean form. We shall now proceed to a description of the internal anatomy and secondary connections of these structures in Ameiurus and closely related species. The mater

Her RICK, Gil Stat 0)')' Paths in Fishes. 423

ial consists of serial sections of ten adult brains cut in the three conventional planes and stained with Delafield's hematoxylin and by the methods of Weigert and Weigert-Pal, together with serial sections of about lOO younger brains prepared by the method of Golgi and cut in various planes.

The vagal lobe of the cat fish is essentially similar to that of most other teleosts. The general relations as seen in transections stained by Delafield's haematoxylin are shown in Fig, 26. The lobe is crowded with numerous small nerve cells,

dors- sec. neurones

intrinsic neurones

superfic. root of X

lat. sec. neurones

deep root of X

nuc intermedius X spinal V tract

desc. sec. Vil tract asc. sec. gust, tract motor root of X

ventral cornu

Fig. 26. Transverse section through the vagal lobes of a large adult of Ictaluriis punctatus (Raf.). Drawn from a haematoxylin preparation. X i6.

y./.w., fasciculus longitudinalis medialis ; n.a?nb., nucleus ambiguus ; j'.r.^., substantia reticularis grisea.

which take up a very pale stain in haematoxylin. Most of these intrinsic cells are neurones of Golgi's type II, whose processes do not extend beyond the lobe. Along the dorsal and lateral borders of the lobe is a layer containing larger cells whose nuclei stain very little in haematoxylin and whose cytoplasm practically not at all. The axones of these cells constitute the chief sensory conduction paths from the vagal lobe to other parts of the brain, and I shall term them the dorsal

424 Journal of Comparative Neurology and Psychology.

and lateral chief secondary gustatory neurones. Deeply embedded in the substance of the medulla oblongata ventrally of the main body of the vagal lobe is an area filled with cells somewhat larger than the intrinsic neurones but much smaller than the chief secondary neurones, which may be termed intermedi

Fig. sy . Three intrinsic neurones of type II from the left va