# Journal of Comparative Neurology 19 (1909)

Embryology - 2 Aug 2021        Expand to Translate
Google Translate - select your language from the list shown below (this will open a new external page)

العربية | català | 中文 | 中國傳統的 | français | Deutsche | עִברִית | हिंदी | bahasa Indonesia | italiano | 日本語 | 한국어 | မြန်မာ | Pilipino | Polskie | português | ਪੰਜਾਬੀ ਦੇ | Română | русский | Español | Swahili | Svensk | ไทย | Türkçe | اردو | ייִדיש | Tiếng Việt    These external translations are automated and may not be accurate. (More? About Translations)

Links: Historic Embryology Papers | Journal Homepage | PubMed J Comp. Neurol.
 J Comp. Neurol. : 1 - 1891 | 2 - 1892 | 3 - 1893 | 4 - 1894 | 5 - 1895 | 6 - 1896 | 7 - 1897 | 8 - 1898 | 9 - 1890 | 10 - 1900 | 11 - 1901 | 12 - 1902 | 13 - 1903 | 14 - 1904 | 15 - 1905 | 16 - 1906 | 17 - 1907 | 18 - 1908 | 19 - 1909 | 20 - 1910 | 22 - 1912 | 23 - 1913 | 24 - 1914 | 25 - 1915 | 26 - 1916 | 27 - 1916-17 | 28 - 1917 | 29 - 1918 | 30 - 1918-19 | 31 - 1919-20 | 32 - 1920-21 | 33 - 1921 | 34 - 1922

Internet Archive

THE JOURNAL OF COMPARATIVE NEUROLOGY AND PSYCHOLOGY

EDITORIAL BOARD

Henry H. Donaldson The Wistar Institute

C. JuDsoN Herrick University of Chicago

Herbert S. Jennings Johns Hopkins University

J. B. Johnston University of Minnesota

Adolph Meter Pathological Institute, New York

Oliver S. Strong Columbia University

John B. Watson Johns Hopkins University

Robert M. Yerkes Harvard University

VOLUME XIX 1909

PHILADELPHIA, PA. THE WISTAR INSTITUTE OF ANATOMY AND BIOLOGY

The Journal of Comparative Neurology and Psychology

CONTENTS OF VOL. XIX, 1909.

## Number 1, April, 1909

Some Experiments Bearing upon Color Vision in Monkeys. By John B. Watson. (From the Psychological Laboratory of the University of Chicago. ) With Five Figures 1

The Expressions of Emotion in the Pigeons. I. Tlie Blond Ring-dove. (Turtur Risorius). By Waixace Craig. {From the Department of Zoology of the University of Chicago.) With One Plate 29

The Reaction to Tactile Stimuli and the Development of the Swimming Movement in Embryos of Diemyctylus Torosus, Eschscholtz. By G. E. CoGHiLL. (Studies from tlie Neurological Lahoratory of Denison University. No. XXII.) With Six Figures 83

Sensations Following Nerve Division. By Shepherd Ivory Franz. (From the Lahoratory of the G-overnment Hospital for the Insane, Washington, D. C.) I. Pressure-like Sensations. With Five Figures 107

Alterations in the Spinal Ganglion Cells Following Neurotomy. By S. Walter Ranson. {Fi'om the Anatomical Lahoratory of the University of Chicago.) Introdution. With Six Figures 125

## Number 2, May, 1909

On the Relation of the Body Length to the Body Weight and to the Weight of the Brain and of the Spinal Cord in the Albino Rat (Mus Norvegicus var. Albus). By Henry H. Donaldson. (Professor of Neurology at the M-istar Institute.) With Three Figures 15f)

Note on the Formulas Used for Calculating the Weight of the Brain in

the Albino Rats. By Shinkishi Hatal. (From the Wistar Institute.) 169

The Nervus Terminalis (Nerve of Pinkus) in the Frog. By C. Judson Herrick. (Fromi the Anatomical Lahoratory of the Unirersity of Chicato. ) With Ten Figures 175

The Nervus Terminalis in the Carp. By R. E. Sheijjon. (From the Anatomical Lahoratory of the University of Chicago.) With Seven Figures. 191

The Criteria of Homology in the Peripheral Nervous System. By C. JxidsoN HiaiRicK. (From the Anatomdoal Lahoratory of the University of Chicago. ) 203

Literary Notices 211

## Number 3, June, 1909

On Sensations Following Nerve Division. By Shepherd Ivory Franz. (From the Laboratories of the Govenvment Hospital for the Insane, Washington, D. C.) II. The Sensibility of tlie Hairs. With Seven Figures 215

Modifiability of Behavior in its Relations to the Age and Sex of the Dancing Mouse. By Robert M. Yerkes. (From the Harvard Psychological Laboratory. ) With Four Figures , 237

The Reactions of the Dogfish to Chemical Stimuli. By Ralph E. Sheldon. (Contribution from the Woods Hole Laboratory of the United States Bureau of Fisheries. ) With Three Figures 273

The Work of J. von Uexkuell on the Physiologj' of Movements and Behavior. By H. S. Jennings 313

## Number 4, July, 1909

Imitation in Monkeys. By M. E. Haggerty. (From the Harvard Psychological Laboratory. ) With Thirteen Figures 337

## Number 5, November, 1909

The Morphology of the Forebrain Vesicle in Vertebrates. By J. B. Johnston. (University of Minnesota.) With Forty-five Figures 457

Some Experiments upon the Behavior of Squirrels. By C. S. Yoakum. (From the Psychological Labor-atory of the University of Chicago.), With Five Figures . .v ^"^^

Tropic and Shock Reactions in Perichseta and Lumbricus. By E. H. Harper. (From the Zoological Laboratory of Northwestern University.) With Two Figures 569

Literary Notices ^9

## Number 6, December, 1909

The Radix Mesencephalica Trigemini. By J. B. Johnston. (University

of Minnesota.) With Thirty-two Figures 593

A New Association Fiber Tract in the Cerebrum with Remarks on the Fiber Tract Dissection Method of Studying the Brain. By E. J. Curran. (Harvard Medical School.) With Three Plates 645

Visual Discrimination in Raccoons. By L. W. Cole and F. M. Long. With

One Figure ^57

A Statistical Study of the Medulla! ed Nerve Fibers Innervating the Legs of the Leopard Frog (Rana Pipiens) after Unilateral section of the Ventral Roots. By Elizabeth H. Dunn. (From the Anatomical Laboratory of the University of Chicago.) With One Figure G85

Factors Determining the Reactions of the Larva of Tenebrio Molitor. By

Max Morse. With Two Figures 721

## Volume XIX April, 1909 Number 1

### Some Experiments Bearing Upon Color Vision In Monkeys

John B. Watson. , From the Psychological Lahoratory of the University of Chicago.

With Five Figures.

For over a year, the writer has been experimenting with apparatus for obtaining large bands of spectral light suitable for use as stimuli in testing the color vision of animals by means of the discrimination method. In the process of constructing the apparatus finally adopted, the writer has availed himself liberally of the assistance of Professors Gale and Milliken of the physics department of the University of Chicago, and of Professor R. W. Wood of the physics department of the Johns Hopkins University. He is also under deep obligation to Professors Angell and Carr for many valuable suggestions.

The actual accumulation of the data bearing upon color vision in monkeys began March 12th and ended August 20, 1908. Two rhesns monkeys (J. and B.) and one cebus (H.), all gentle and accustomed to experimentation, were the subjects used in the investigation. The report is given in its present incomplete form, because of the fact that the writer's work could not be continued at the University of Chicago. The apparatus used there, however, has been duplicated in the Hopkins laboratory and the investigation will be continued in the latter place both upon two of the monkeys,

The Journal op Comparative Neurology and Psychology. — Vol. XIX, No. 1.

2 'Journal of Comparative Neurology an J Psychology.

J. and B., which served as subjects in the present work and upon other monkeys. In addition to continuing the tests upon color vision, steps are being taken thoroughly to test the delicacy of the white-light vision of the monkeys and their sensitivity to differences in the size and form of visual stimuli.

In view of these further tests, which are concerned with the nature and delicacy of the visual reactions as a whole of this animal, it would seem to be premature in this preliminary report to enter into any general discussion of color vision in animals or to take up the related structural facts which bear upon color vision. The color vision of several different species of animals is being tested in the various laboratories and for this additional reason the writer will confine his statements on the historical side strictly to the work of Kinnaman, which, so far as the writer is aware, is the only study of the color vision of this animal which can lay any claim to scientific accuracy.

Kinnaman's^ work was certainly as careful and as exact as his method would permit. His stimuli were obtained by means of light reflected from pigmented papers. The method was roughly as follows : A board, 1 inch by 7 inches, and 5 feet long contained six holes pierced at regular distances. Each hole was large enough to admit the bottom of a cylindrical glass. The convex surface of each glass was covered with colored papers or with different grays. Food was kept with one of the colored glasses, the position of which could be varied at will after each test. The discriminations were made rapidly. "In order to determine whether brightness or color was the basis for discrimination four control tests were made. In three of these, I attempted to determine whether the monkeys could discriminate greys and colors varying by the same degree of brightness equally well. If blue and red, for example, with a difference in brightness of 15° (determined by the flicker method) were differentiated perfectly, and two grays differing by 15° very imperfectly, then color very probably was the basis of the discrimination in our first series of tests." Both original tests and control tests were

^American Journal of Psychology, Vol. 13, p. 98.

Watson, Color Vision in Monkeys. 3

carefully made aud so satisfactory were the results obtained from them that we find the author expressing himself somewhat extravagantly as follows:

"1. There can be no doubt that monkeys perceive colors.

"2. Two grays having a given degree of difference in brightness are not discriminated as well as two colors having an equal diiference in brightness.

3. For accurate discrimination of diiference in brightness a difference of about 35 degrees or 9 per, cent of the white constituent of the gray is necessary.

4. The monkeys are able to distinguish colors from grays though the brightnesses are the same.

5. The male appears to have a preference for bright colors, but blue seems to be discriminated against.^

"6. In two instances there w^ere indications of at least a low form of general notion."

For detailed reasons, which can not be entered into here, the assertion is ventured that the use of colored papers can never give us a satisfactory test of color vision in animals.^ Certainly, the writer has room for "doubt" both in Kinnaman's work and in his own. The fact that there is a high jjercentage of correct choices of the positive color in both investigations cannot be doubted, but that the correct choices were made , upon basis of 'hue' and not that of 'brightness' cannot be decided so easily as Kinnaman supposes. It would not be at all ridiculous for one, skeptical of the possibility of testing the color vision of animals, to assert that every discrimination which has been made between two colored papers by an animal could have been made if it had been totally color-blind.

As a reason for such scepticism, a few of the defects of this method as a whole may be mentioned noAv:

1. The surfaces of the papers differ greatly, owing to accidents

^Blue certainly was not discriminated against by tlie animals used by the writer. See Table VI.

^Tbe committee appointed by the American Psj'cbological Association at the 1907 meeting to report upon methods and apparatus for testing vision in animals will take up in its report the defects and advantages of the various ■ methods of testing color vision.

4 'Journal of Comparative Neurology and Psychology.

in manufacture, dyeing, ironing, etc. In addition, it is extremely difficult to bend colored papers around glasses or to paste them upon doors so accurately that slight differences in form, size and depth do not apjDcar. To sum uj) these defects under 1, we may say that colored papers afford numerous secondary criteria.

2. They do not reflect monochromatic bands, but overlapping bands. This is especially true of those reflecting the shorter wave lengths.

3. The range of intensity ohtainahle in them is so timited, that if any given region of the spectrum should ojfer to the a)iinial a different order of intensity from that ivhich the same region offers to our own eyes, the slight change ivhich ive could introduce in the hrightness of a given stimulus, hy substituting a paper of the same color only lighter or darher {to our own eyes) might not at all reverse for the animal the intensity relation originally existing between the colors.

Yerkes,* in his work on the dancing mouse, mentions most of these objections and shows by experiment that for that animal the red end of the spectrum is probably extremely weak in intensity. Whether or not the different parts of the spectrum possess the same intensity for the eye of the monkey, as they do for our own, is a question which we have at present no data for deciding, but certainly in the present state of our knowledge we cannot assume such to be the case.

Believinw^ that many problems in the study of color vision in animals cannot be solved without the aid of a continuous spectrum, and being more or less disJieartened by the failure of colored papers and filters (as they have heretofore been used) to furnish suitable stimuli for testing even the more elementary questions at issue, the writer began work upon a spectral light apparatus which it is hoped will make possible an experimental treatment of the following problems :

1. Has the animal the power to discriminate between any given color and any other selected color with equal ease, when the relative

Yerkes, R. M. The Dancing Mouse. The Macmillan Company, 1908.

Watson, Color Vision iti Monkeys. 5

intensity of either color, and the absolute intensity of both may be altered at will ? In investigating this problem, we ought to be able to find totally color blind animals, red-green blind animals, animals with normal color vision, if such differences in sensitivity exist. After the problem has been solved, color theories based upon the phylogenetic development of a photo-chemical molecule will or will not have bases in fact,

2. How nearly identical in wave length may any two colors be and still afford a basis for, the animal to discriminate between them (the qualitative 'difference linien', D.L.) ?^

3. How nearly identical in intensity may two bands of the same wave length be and still afford a basis for discrimination (the D.L. for intensity) ?

4. Do the different parts of the spectrum possess different threshold values (stimulus limen, R.L.) ?

5. Is the spectrum of a given animal wider or narrower than the average width of the spectrum of man ?

Description of Appaeatus.

As Figs. 1 and 2 show, the light apparatus is made upon the principle of a spectrometer. Fig. 1 shows, in order, the arc, A, the condensing lens, Lj, the slit, S^, the collimating lens, L.2, the prism, P, the mirror, ilf^, and the lens, L3, (all are enclosed in a system of dark boxes).

The arc is an ordinary hand-feed arc. The direct 220 v. current supplying the arc is furnished by the university power-house. This curxent is very steady and uniform. The positive (cored) carbon is placed horizontally and in the axis of the optical system. The arc is so arranged that it can be adjusted by the experimenter in the adjoining room at K (Fig. 2). Two long rods, BE, extending from the are into the experimental room permit this. AC in Fig. 1 shows the gearing system by means of which the long rods are brought into connection with the short feeding rods of the arc. After practice the experimenter can control the arc at K through

^In the apparatus to be described, a double image prism would afford the conditions for this test.

6 'Journal of Comparative Neurology and Psychology.

long periods of time without allowing sensible alteration in its intensity. Since the arc is not in the dark-room where the animals react, the noise made by it is not a source of disturbance. When burning well, this arc rarely makes a noise which is noticeable even near at hand, after the heavy wooden box (metal lined) has been closed.

The lens used as a condenser is an ordinary 4" biconvex reading glass with 8" focus. This lens gives a clearly defined image of the crater of the positive carbon at its focus.

The slit, Si, which is placed at the focus of Li, is a common optical slit with knife edge opening. Its width was adjusted once for all to give a spectrum of high intensity and was never thereafter changed.

The collimating lens, L^, is a heavy 31/^" achromatic lens with a focal distance of 18". The slit, *S'i, is in the focus of this lens, consequently the face of the 65 mm. heavy flint glass prism, P, is filled with parallel light admitted by it. The now refracted beam falls upon the mirror, M^ (the use of this mirror is necessitated by the narrowness of the room) which is silvered upon its anterior surface. This mirror reflects the beam through the achromatic lens, Lg, (similar in all respects to L^ except that Lg's focal distance is 24").

The lens, L^, brings the refracted beam to a focus upon the double slit, 82, (not shown in Figs. 1 and 2, but shown separately in Fig. 3) in a series of colored images of 8^^. The solar lines are plainly visible and serve as a guide in the selection of particular wave lengths. The apparatus can be so arranged (by revolving Mj upon its axis) that these lines shall coincide with definite mm. divisions of the slit.^

In order that this double slit may be more easily understood, the following description is given: Fig. 3 shows the slit horizontally. The sharply outlined spectrum falls upon the polished surface of the slit between on the 8C scale, and on the 8C^ scale. Two selected portions of the spectrum pass out between the knife edge

'For example, the Na line can be made to fall upon 3 of the scale >S'C etc.

Watson, Color Fision in Monkeys. 7

openings, J-J^, and J 2'-^ i- The Avidtli of these openings is controlled by means of the micrometer screw system Col, ES, K, B, the mechanism of which is well known (i. e., turning Cal, e. g., forces the nnt, K, backwards or forwards as the case may be and consequently the edge of the jaw at J advances or recedes from J^).

The two small jaws, J^ and J 2, must be moved by hand. They are held in place (i. e., in the grooved track of the slit) by means of a small bowed spring. Since they are cut accurately to fit the track in which / and J3 slide, they are held firmly in vertical position by means of the spring. If it is desired to have the opening, J- J I, admit some other part of the spectrum, the apparatus easily permits it. Suppose we desire to have the opening at 6 instead of at 3.65, as it stands in the diagram (see relation of scale SC to index /). We turn the screw, Cal, until / falls at 6. The small jaw, Ji, is then pushed up flush against J. J is then pushed forward, carrying Jj with it for whatever wndth of slit is desired ; / is then backed again to 6. This leaves the opening, J-Ji, in its new position optically perfect. By means of the micrometer screw head this distance is made accurate to 1/1000 mm. This would leave a much wider space at than before. This variable opening at is closed with a strip of black cardboard. Four tiny points, p, Pj. P2, P:^, on the jaws, J^ and J 2, facilitate this.'^

The position of this slit in the system can be inferred from an examination of Fig. 2 : L.,, of Fig. 1 is sho^^^l on the right in Fig. 2 ; S2 is at the focus of this lens, 24" distant. The two small vertical mirrors, il/^ and .1/;., placed at an angle of 4.5° so as to form a horizontal V (with apex directed towards ^'2 and midway between the two openings) serve to catch the two selected beams (e. g., red and green) issuing from the openings, J-Ji and J-yT:^ of S-. and to reflect them to the mirrors, ilf^ and M- resjiectively. These latter two mirrors in turn reflect the two beams in a parallel way doA\Ti the room to the screen. The width between these two mirrors can be adjusted to any desired distance. Since the rays as they issue from the openings of /S'o are diverging, we have a broad, diffuse,

'The slots, SI and »S7,. serve to admit bolts for attaching slit as a whole permanently in its vertical position.

8 Journal of Co?nparative Neurology and Psychology.

Fig. 1.

COLOR VISION IN MONKEYS

I'Hlniil.ll I llnilli,iilillll milin,liii,l, lii^iiii|(S)

i]»mQ

The .Tournal of Comparatitb NBUROLOfiv AND Psychology. — Vol. XIX, No. 1.

Watson, Color Vision in Monkeys.

Fig. 5.

10 'Journal of Co^nparative Neurology and Psychology.

faint band of red and of green light on the screen. In order to intensify and sharply define the hands on the screen, two small achromatic lenses, L^ and Lr„ are interposed in the pathways of the two beams, B and G respectively. These two lenses are of short focal length (6"). They project sharply defined and enlarged images of the two openings of So npon the screen (marked red and GKEEN in the diagram). These images are abont 7" in height and 1%" in width. .

Four reversing mirrors, BM, are used in order to reverse the right and left positions of the two beams. These mirrors revolve in a vertical plane. They are mounted in bearings in such a way ■that the small weights, Wgt, pull them back to the 45° position whenever the cords, CBM, are slackened, as in the diagram. These cords are jointly fastened to a rod at X. A single forward pull upon this rod brings all four mirrors to the 180° position, in which position they no longer intercept the two beams.

A glance at the apparatus will show that when, e. g., the red is on the left, the reversing mirrors have to intercept the beam ; wdien it is on the right, they no longer intercept it and the beam is reflected directly from the mirror, M^, to the screen. It is clear from this that the absolute intensity of the two bands is slightly less in the "reversed position" than in the normal. But this reduction occurs in both bands equally.^ In order to compensate for this reduction, a vertical sliding bar, SB, is placed in the pathway of the beams. 1" x 2" Avindows are cut in this bar at the points where the beams impinge upon it. One half of each window is

^\11 the mirroi-s in the sj-stem are silvered on the anterior surface. They are kept highly polished by the use of "jewelers' rouge." When not in use. they are kept covered with silk handkerchiefs. The absorption of the light consequently is kept constant and at a minimum.

[Since the mirrors used in the apparatus are a source of a great deal of trouble and care and since their use causes a certain variation in the absolute intensity of the light, effort was made to find a substitute for them. After some experimentation it was found that total reflection prisms could be made to separate the beams, to space them properly, and finally to reverse them. In addition the dark room at Hopkins is large enough to accommodate the apparatus without the use of the mirror. M„ behind the large prism. The whole apparatus is now "self-maintaining" and completely constant so far as the absorption of the light is concerned.]

Watson, Color Vision i?i Monkeys. II

left open, the other half is covered with a sufficient number of strips of plate glass to compensate for the absorption of the reversing mirrors. To one end of this bar a spring is attached. When this spring operates alone, the bar is held in snch position that the beams of light have to pass through the open halves of the windows. A cord, C8B, runs from the opposite end of the bar around to the rod at X, which controls the miri'ors. Pulling upon this cord brings the sliding bar forward to such a position that the beams have to fall upon the halves of the openings which are covered by the plate glass. This bar is made to work synchronously with the mirrors in such a way that when the reversing mirrors are "out" the plate glass windows are "in" and vice versa. A simple forward pull or a release of the rod at X adjusts both mirrors and bar. This compensatory device was used in certain of the control tests described below (all of the red-green), but since many trials showed that the reactions of the animals were not altered by its insertion or removal, its use was discontinued.^

Behind the opaque screen, there is a vertically placed 12" x 24" pane of acid ground glass (acid ground is less granular than common ground glass ; milk glass would have been used, could it have been obtained). The moment the screen is raised by pulling upon H, the two bright-colored bands (surfaces) appear. Immediately behind each of these bands (they are 8" apart) ^^^ a hole is cut in the platform to admit the food-boxes shown in Fig. 4. A glass partition, GP, set in a low wooden base serves to keep the animal from opening both food-boxes at once. It also serves to force the animal to go clearly to the right or to the left and to keep a position habit from forming. ^^

•However, if one were workinc: with two colors approximately equal in intensity to the animal, it mi.slit very well happen that this change in absolute intensity would, owing to possible onset of the Purkinje phenomenon, alter the intensity relation for the animal.

"This distance depends upon the distance of M^ from M^.

"Before tliis partition was at hand, one monkey, whose records are not given, went always to the right, then down the screen of ground glass to the left until he came to the box which contained the food.

12 'Journal of Co7nparative Neurology and Psychology.

Method of Controlling the Intensity of the Two Lights. The diagrams (Figs. 1 and 2) show that the absohite intensity of the entire spectrum may be changed by opening or closing the slit /Sj. Since the present tests are not concerned with faint or weak spectra (as such), the width of this slit, as has been stated, was kept constant and as wide as was possible still to permit a sharply defined spectrum at S^. The intensity of the two selected bands can be altered separately in two ways: By attaching an iris diaphragm (or better, possibly, an Aubert) to each of the two projecting lenses, L^ and L^,, or by the use of an episcotister.

In the experiments here reported, the intensity of the red and of the green was controlled by the use of the iris diaphragm, while that of the blue and of the yellow was controlled by the episcotister. The chief objection to the diaphragm lies in the fact that the width of the band changes slightl}^ when the diaphragm is opened or closed. By an oversight, the spectroscopic reading was taken only when the diaphragm was completely open. It was my intention to return to the red-gr,een discrimination during the summer months when it was possible to obtain sunlight, and to use both the diaphragm and the episcotister, but I found that the time at my disposal did not permit this. Accordingly, all the red-green tests reported below were made with the arc as a source ; and all changes in intensity of the two bands were made by means of the diaphragm (in table of constants, e. g., "red maximum" was obtained by a wide open diaphragm, "red minimum" by a fixed pinhole opening in the diaphragm). On the other hand, all changes in intensity in the blue-yellow tests were made by means of the episcotister. In the present state of color photometry, it is desirable to have some check upon photometric readings. The episcotister furnishes such a check by allowing us to increase or decrease in a constant way the angular opening through which the beam is allowed to pass. In the present Avork, the episcotister proved eminently satisfactory. In the table of constants given below, the maximum yellow, e. g., was the nonnal intensity of the beam as it came from ^2 ; the minimum yellow was the intensity of the beam after it had been interrupted by the episcotister, set with a 30° opening (15° on each side). The angle

Watson, Color Vision in Monkeys. 13

of 30° was cboseii as the miuimum after several preliminary trials. It was desirable to keep the minimum intensity of any beam always well over the human threshold. Any smaller angle did not permit this, wath sunlight as the source. That this minimum w^as also well above the animal's (reaction) threshold was tested in the following way: The episcotister, set at the minimum (30 '), was allowed to interrupt the blue ; the yellow was cut out at ^'o and the animal tested at X in the ordinary way. When the screen was raised, only the blue band appeared. As a result, it was found in every case that the animal followed the light regardless of its right or left position. The blue was then cut out at >S'2 and the yellow interrupted, with similar results. There is the possibility, however, that the minimum intensity was over the 'brightness limen' but not over the 'color limen.' This objection cannot be met until extended threshold tests have been made.

In conducting such experiments in the future, the following procedure will be adopted: first, during the formation of the association, an episcotister opened to the maximum (320°) will interrupt each beam continuously during all tests ;^^ second, after the association has been established, the control tests will be made with the two episcotisters set at any desired angle ; third, the iris diaphragm will be used as an additional control. ^^

The episcotister was run at a very high rate of speed. As a test as to whether the beams were uniform for the animal, the screen

"In duplicating the apparatus at Hopkins, the motor and the two episcotisters were mounted upon a small revolving table. This table is so arranged that a pull on a cord (at X, Fig. 2, where the experimenter sits) will interchange the positions of the two episcotisters, thus making it possible to have the animal on the one trial react, e. g., to "minimum red," "maximum gi'een," and on the next to "maximum red," "minimum green."

"It will be remembered that there are three common ways in which intensity in the physiological sense can be altered: (1) by decreasing the amplitude of the ether waves of the beam which falls upon a given retinal area (increasing distance of source) ; (2) by lessening density of beam (use of diaphragms, etc.) ; (3) by interrupting beam (episcotister). In order to test whether the physiological effect, e. </., of distancing the source of the beam is the same as interrupting it, it is desirable to make changes in intensity by employing all three methods. The desirability of the use of the episcotister and diaphragm has been assumed in the present work.

14 'Journal of Cornparative Neurology and Psychology.

was sometimes raised before the episcotister had gained full speed. The flickering light never failed to frighten the animals. They would never leave my shoulder to make a choice until the lights appeared perfectly steady to my oirn eyes.

Method of Determining Intensity of Ielumination of

MONOCIIKOMATIC BaNDS.

After vainly trying to obtain a photometric reading of the minimal intensity of the monochromatic bands,i^ with a photometer based upon the Joly principle, I finally, on the advice of Professor Milliken, abandoned the photometer and had resort to the simple apparatus, the groimd plan of which is shown in Fig. 5. In the diagram, V is a band of monochromatic light visible upon the ground glass surface. C is a white surface (bristol-board; plasterparis is preferable) equal in width to Y , which reflects light to the prism, P, from a source the intensity, and distance from C of which, is known. P is a 90° prism silvered on the two surfaces which reflect the inuiges of V and C iuto the eye at E. The distances from V to P and from C to P are equal, 15.5 cm. The distance from P to -E' is 20 cm. The total distance from E to V approximately equals the distance of the color from the eye of the monkey when he reaches the partition GP in Fig. 2 and makes his choice as evidenced by his going to the right or to the left of the partition (the monkey B. often stoj^ped at this j)oint and turned his head first to the right, then to the left, etc., before finally making his choice).

The photometric determinations were made in a dark-room under conditions as nearly as possible like those under which the animal reacts. The eye was dark-adapted (15 to 30 minutes). A comfortable position was taken with the eye at E so that a clear reflected image of V appeared. An assistant then lighted a standard electric light which was screened from the observer's eye by the opaque screen, 08, and mounted it upon the board, B, which was graduated in cm. The distance of this light was varied until the observer at E judged the two lights to be equal. The judgments made under

"They are not really sources iu the technical sense, but surfaces.

Watson, Color Vision in Monkeys.

15

these conditions arc introspcctivelv similar to those made with an ordinary photometer,. The images of T^ and C appear side by side, with no dividing surface between. Six judgments for each change in the intensity of the band were made, three ascending and three descending, and the results averaged.

The units in which the results below are given are "hefnermeters." In other words, the apparatus measures the "intensity of illumination" of the variable surface, Y, in terms of C, the "intensity of illumination" of which can be calculated from the formula :

Cos (9 ■ r*

(($>= 45^ in all cases; r is read directly from the scale of B). ^^ The following table of constants gives the wave lengths of the four bands used in the present work and their "intensity of illumination" under the various conditions. Table of Constants. Designation of Variable Light. Standard White Light. Average Distance from C. Hefner-Meters. 1. Max. Sun. Yellow 16 C. p. 49.8 ± 2.8 cm. 56.33 2. Mill. " " Hefner 29.4 ± 1.8 " 8.18 3. Max. " Blue 16 c. p. 84.6 ± 3.4 " 19.51 4. Mill. " Hefner 46.8 ± 2.5 " 3.23 5. Max. Arc Yellow 5 c. p. 81.0 ± 4.0 " 6.65 6. " " Blue 5 c p. 128.0 ± 2.8 " 2.665 7. " " Green 5 c. p. 86.0 ± 5.2 " 5.90 8. " " Red 5 c. p. 72.0 ± 4.0 " 8.42 9. Mill. " Green Hefner 130.0 ± 3.6 " .418 10. Mill. " Red " 69.0 ± 2.8 " 1.485 Width of Monochromatic Bands. Red = A 6485 — 5790. Yellow = A 5750 — 5600. Green = A 5250 — 4825. Blue = A 4800 — 4650. "The angle 6 refers to the angle at which the light from the source falls upon the cardboard C. "These are broad spectral bands, mutually exclusive, but not "monochromatic" in the sense in which the physicists use that term. l6 Journal of Comparative Neurology and Psychology. Method of Presenting the Stimuli. The animals were carried from their living room to the darkroom, and allowed to remain there for two or three minutes. A malaga grape was then placed in the food-box at the base of the colored band selected as the positive color. The animal was next permitted to climb to mj shoulder. The screen was then raised and the animal allowed to walk towards the lights (for a distance of three feet) and make his choice, i. e., open either of the two boxes. If the red box were opened, the grape could be obtained; if the green, the animal was pulled back to my shoulder. After the choice had been made, the screen was again dropped over the lights and the animal allowed to finish eating the grape. At first, the stimulations were given at intervals of three minutes, but the animals became so eager and restless during the long wait, that the tests were finally given as rapidly as the apparatus could be arranged. The animal was always j^ermitted to finish eating the food obtained at the previous trial before the next test was given. Differences between the olfactory values of the two boxes were eliminated in a variety of ways: The boxes were frequently interchanged ; other similar, boxes were used occasionally ; food was placed in both boxes, but the conditions were so arranged that the animals could get food only in the one box. Fig. 4 shows one of the boxes. A horizontal wire partition runs across the box, 1" from the top. When the grape was placed below this, the monkey never made the slightest effort to get it. The most important control test was made after the association had been firmly established by allowing one or two grapes to lie very near the box kept with the green, or with the yellow as the case might be (the color to which the animal was not reacting positively). In no case were these grapes ever disturbed — even if the animal made an occasional error and opened the wrong box, he never seemed to notice that the grapes were nearby. In the hundreds of cases where he had obtained the grapes, he had found them in the top compartment of the box after he had pulled open the lid. While these tests do not appear in the tabulated report, they were made exhaustive. In addition to controlling the factor of smell, such tests tend to show conclusively that possible Watson, Color Fision in Monkeys. 17 small differences in tlie visual characteristics of the boxes did not serve as secondary criteria to which the animal might have reacted. In presenting the lights, care was taken in the successive tests to vary the height to which the screen was raised. In order to keep this factor variable, I purposely left off a stop which would allow the screen to go a certain height and no higher. The screen could be raised from 0"-7", the full height of the band. The animals would, e. g., on one test have to react with the two bands only I/2" high, while on the next the bands would be 7" high. This procedure tended to eliminate the possibility of their reacting to slight differences in form. A few of the tests where the vertical and horizontal forms of the two bands were changed independently are included in the report. In the early part of the work, only six trials per day were given. Gradually, as the animal became more accustomed to his work, fifteen to twenty were given. The animals were always eager for the grapes, and so long as they were unafraid, the tests could be repeated again and again until the experimenter became fatigued. The positive color was presented very irregularly as regards left and right positions. In the early stages of the association, the position of the two bands was interchanged regularly. After the first three or four days, however, irregularity was introduced. The positive color would be presented first on the right on a given day, while on the succeeding day it would be presented first on the left. In order to give an idea of the variations introduced in the order of the presentation of the stimulus, I offer the following notes taken from J's reactions to red-green: April 23. Red on left to begin : Red and green alternated for the first three trials. Red given three times on right. Red given three times on left. April 24. Red on right to begin : Red and o-i-een alternated for three trials. 1 8 'Journal of Comparative Neurology and Psychology. Red given three times on right. Red given three times on left. Red given three times on right Red given three times on left. On April 22, in the nine trials given, the position of the red and green was alternated. The Results of the Experiments. The tables given below show, separately for each animal, the daily set of conditions to which he was reacting, and the results of the test. The date of the experiments is given first in the tables, then, in order, the number of choices of the positive color (i. e., color with food), the number of choices of the negative color, the percentage of correct choices, the source from which the spectrum was obtained, the relative and absolute intensity of the two bands (refer for photometric statements to table of constants, p. 15), and the remarks. A reference to the tallies shows that they are divided into two parts: A, with stimulus constant, and B, with stimulus variable. The first part of division A, shows in all cases the gradual rise of discrimination, Avhile the second part gives the animal's maximum of steadiness under an unchanging set of conditions. When a maximum of steadiness had apparently been obtained, changes in the relative brightness of the two bands and changes in their form and surfaces were introduced. Division B of each table shows clearly just what changes were made and the effect of such changes upon the i^ercentage of correct choices. Discussion of Results. 1. The surprising result of the early part of the test on the redgreen is seen to be the failure of the animals to read to the red. This was noticeable in all of the animals tested, see especially Tables I and II. The early records of monkey B. are not given in Table III. His failure to react to red on the early trials was as pronounced Watson, Color Vision in Monkeys. 19 as in J's case. The same is true of the early records of another monkey (S). My notes are full of such comments as the following: "Animal apparently is not stimulated by red," "apparently can not see in red light," etc. While I can give no explanation of the results, it is worth while to mention that they are suggestive either of a 'preference' for green, or possibly, in the light of Yerkes' experiments on the dancei', of the low stimulating effect of red, I am inclined to favor the first of the two possibilities, if any, for the reason that after the habit of reacting to the red was perfected, the discrimination could still be made when the absolute intensity of the two bands was enormously lessened (I had abundant incidental opportunity to make many such tests on occasions when the arc momentarily, for one reason or another, failed to give an intense spectrum) , 2. In the case of all three animals the blue-yeMow discrimination arose more rapidly than the red-green. Indeed, in the case of B,, the habit of reacting to blue was formed with extraordinary rapidity. The question arises as to whether blue was not a 'preferred' color. Again, however, there are no unequivocal data at hand to aid us in reaching a decision for two reasons : 1st, the general question raised above, but not answered, as to whether the wdiole red end of the spectrum does not have a low stimulating effect upon the photochemical substances in the retina of these animals. If the blue were more intense, the animal might from the beginning tend to react to the blue and to neglect the yellow ; 2d, the effect of practice. It must not be forgotten that when the animals began upon the blueyellow test, they had previously had training in the red-green. For this latter reason, no conclusion can be drawn from the present work bearing upon the views sometimes expressed that the yellow-blue photo-chemical process is phylogenetically older than the red-green. 3a. The onset of position habits in the case of both H. and B. in the red-green test, and in that of J. in the blue-yellow and the constant struggle of the experimenter to keep such position habits from forming are significant Although these animals had been obtaining food for several days with a given color stimulus, we find instead of a steady and rapid increase in the ability to associate 20 'Journal of Comparative Neurology and Psychology. the color with the food, a total break in the process, and a resort, on their part, to the use of a sensory process genetically lower than the visual, namely, the kinsesthetic. Such behavior is suggestive of the relatively unimportant role which color vision plavs in the life of til is animal. 3b. The onset of the position error in such a crucial place in J's series (see p. 26) is especially unfortunate. His percentage of correct choices increased normally during the formation of the association. This percentage remained high for the several days, during which the various changes were being made in the negative color, then all at once we find the habit disintegrating. On the whole, in ■this case, it seems best to reserve judgment as to whether the animal was reacting to blue on the basis of its possibly greater intensity, and to await further tests upon him. 4. If 0716 were to draio the general conclusion that the wave length of a given monochromatic light stimulus is, or might he. under suitable conditions, a factor in the adjustment of the animal to that stimidus, one apparently would find abundant support for the position in the above tables. The writer for the present, however, prefers to allow the experimental data to stand as such without drawing any conclusions from them. The reason for this position becomes apparent when we consider the following points : Did the changes which were made in the relative intensities of any two bands really reverse the intensity relation for the animal (a glance at the table of constants will show the enormous changes which were made) ? The answer to this question must come from experiments. In order to answer it we would need to know for each species of animal, 1st, the relative stimulating eifect upon it of the different parts of the spectrum, that is, we would need to have a curve for the animal corresponding to the curves which have been constructed for similar reasons for the human eye in its normal and abnormal states. The first step in constructing such a curve might come through obtaining the animal's reaction thresholds (stimulus limen, R. L.) for the separate spectral bands, e. g., the red, yellow, green and blue ; 2d, beginning with these values (expressed in photometric and in radiometric terms) as the lowest points in the Watson, Color Vision in Monkeys. 21 intensity scale of each of the above named monochromatic light stimnli, we might next lav off a series of steps, the reaction D. L's. (diiference limen) along the scale of each of the four bands. In order to illustrate the foregoing method, let us suppose that we are testing the ability of an animal to make a discrimination between red and green. We have found previously a reaction threshold value (R. L.) for green, say X, and a similar threshold for red, say, 20 X (that is, red is lower in its stimulating effect than the green). The next step in the experiment would be to confront the animal in the usual way with these two stimuli at the intensity of 20 X for red and X for green. If the discrimination arose, we would have to assume that the animal was discriminating between the two colors by reason of the difference in their wave length. Absence of discrimination at this level of intensity, however,, would bo no proof of the lack of 'color vision' {stating the situation in conscious terms for fJie sake of convenience) for the reason that the values X and 20 X might represent the 'brightness thresholds' at these places in the spectrum and not the 'color thresholds.' In the absence of color discrimination at the level of the thresholds (R. L.), we should have to carry our experiment further and test the possibility of the discrimination arising when the intensities of the two stimuli are raised respectively to the level of their previously determined first D. L, (with red, e. g., at the intensity 20 X 4" c, where c represents whatever constant the Weber-Fechner Law requires, and green at X + c^. If discrimination failed also at this ])oint, at intermediate points and at points high up on the intensity scale, we would have just grounds for denying color vision in the animal; or if discrimination were possible, for affirming it.^'^ That many difficulties are in the way of the successful carrying out of this experiment, the writer is painfull} aware. The chief difficulty in the way of such an investigation will be found to lie in our present crude methods of color photometry. With conditions as they are, therefore, I felt that the safest method to use was the one adopted in the present paper, namely, to alter the relative intensity of the two bands by enormous steps, hoping that when tests on stimulus and "Provided, as is not the case, tliere were no otlier factors to consider. 22 'Journal of Comparative Neurology and Psychology. difference limens shall have been made, the results will show that the relative differences in intensity with which I worked were far in excess of those needed to reverse for the animal the intensity relation of any given pair of colors. The secoind question to be raised is concerned with the differences in the energy of the different parts of the spectrum. Might not the animal after all (apart from the intensity relations) be reacting to secondary energy criteria of one form or another ? The writer is not able at present to enter profitably into a discussion of this phase of the subject. That there are problems lying here Avhich must be solved, but which can be coped with in no easy manner, no one acquainted M'ith the facts can doubt. With such questions raised (and they are not raised here for the first time) is it any wonder that we find it impossible to accept the uncritical results which have been obtained by the use of filters, colored papers, etc., as evidence for the presence of color vision in animals ? Watson, Color Vision in Monkeys. 23 TABLE I. J's Reactions to Red-Green. A — With Stimulus Constant. Date. Red. Green. Per Cent. Correct. Source. Intensity. Remarks. Red. Green. 3-12 3 Arc Max. Max. 3-13 1 5 16.6 3-14 6 3-16 1 5 16.6 3-16 5 3-17 1 4 20 3-18 2 4 33.3 3-19 5 3-20 5 1 83 3-21 3 4 43 3-22 2 4 33.3 3-23 6 100 3-24 3 3 50 3-25 4 2 66.6 3-26 5 1 83 3-27 6 100 3-28 6 100 3-29 6 100 3-30 6 100 3-31 5 1 83 4- 1 5 100 4- 2 5 1 83 4- 5 5 1 83 4- 6 6 100 4- 7 5 1 83 B — With Stimulus Variable. 4- 8 6 100 Arc Min. Max. 4- 9 9 3 75 Max. " 4-10 12 3 80 Min. '* 4-11 8 1 89 " " 4-12 14 2 87.5 " 4-13 12 4 75 " Animal very hungry. 4-14 9 1 90 Max. Min. 4-16 8 4 66.6 " " Fed too much. 4-17 9 1 90 '* '* 4-19 5 1 83 "I Change in intensity made in middle 4-19 6 100 Min. Max. ) of series. 4-20 14 100 E* " 4-22 9 1 90 " 4-23 7 2 77 Max. Min. 4-26 13 100 E Max. 4-27 12 1 92 Max. Min. Only lower half of red exposed. E designates subjective quality. The above tests were continued from April 27th to May 28th. An average of from 85-90 per cent of correct choices was maintained throughout the whole period. In these last tests, all variations in the presentation of the stimuli which could be thought of were introduced, such as presenting the red on the right and left alternately, red twice on right, once on left, then three times on right and three on left, etc. On account of the position error eiftering into B's reactions, he was fully one month behind J. Since it was desirable to keep J in practice, he was put through all the control tests with B. 24 Journal of Comparative Neurology and Psychology. TABLE II. H's Reactions to Red-Grekn A — With Stimulus Constant. Date. Red. Green. Per Cent. Correct. Source Intensity. Remarks. Red. Green. 3-18 4 Arc Max. Max. 3-19 1 3 25 3-20 2 4 33.3 3-21 6 3-22 4 2 66.6 3-23 1 5 16.6 3-24 4 2 66.6 3-25 3 3 50 3-26 4 2 66.6 3-27 4 2 66.6 3-28 3 3 50 3-29 5 1 83 3-30 3 3 50 3-31 6 100 Disturbance in general physical con dition of animal for several days. Position error became noticeable when tests were again started with animal, 5 days later, animal going always to left. This error persisted until May 4. In the interim 6-10 trials per day were given. The in troduction of a partition broke up error. 5- 4 8 1 89 5- 5 8 1 89 B— With Stimulus Variable. 5- 6 5- 8 5- 9 5-10 5-11 5-12 5-13 5-14 5-15 5-17 5-18 5-18 5-19 5-20 5-21 5-22 5-23 5-24 5-25 5-26 5-27 5-28 6 8 2 8 2 8 2 10 1 8 2 10 2 9 1 9 1 13 1 6 1 7 11 7 5 11 1 14 2 9 1 9 1 16 15 2 12 4 12 100 80 80 80 90 80 83 90 90 93 86 100 100 58.3 91 87 90 90 100 88 75 100 Arc Max. Min. " E. Max. J •• Min. " " Max. Min. •' E. Min. '• Change in intensity made in middle of series. Animal too hungry: all 5 errors made in succession. Only lower half of red exposed. Red i as wide as green. Watson, Color Vision in Monkeys. 25 TABLE III. B's Reactions to Red-Green. B was given test for test with J (see J's record) with similar averages up to March 25th. Noticeable position error began to appear which grew steadily worse. This error, as in H's case, was finally eliminated by the introduction of the glass partition. The error persisted for fifty-one days. Ten to twelve trials were given per day during this entire period. A — With Stimulus Constant. Date. Red. Green. 5-15 — 5-20 Held av'age Per Cent. Correct. 90 Source Arc Intensity. Red Max. Green. Max. Remarks. This average will serve as basis for comparison with those in B. B — ^With Stimulus Variable. 5-21 13 1 93 Arc E. Max. 5-22 14 2 87 5-23 13 1 93 Min. 5-25 15 1 94 Max. Min. 5-26 15 2 88 Only lower half of red exposed. 5-27 12 2 85 Half vertical strip of red shown. Animal frightened. 26 Journal of Comparative Neurology and Psychology. TABLE IV. J's Reactions to Blue- Yellow. A — With Stimulus Constant. Date. Blue. Yellow Per Cent. Correct. Source Intensity. Remarks. Blue. Yellow 7-14 4 4 50 Arc Max. Max. 7-15 6 4 60 " 7-16 8 3 72.7 " 7-17 9 4 70 " 7-18 12 5 71 •' 7-19 10 5 66.6 " 7-20 17 8 68 " 7-21 12 1 92 " 7-22 12 2 86 " 7-23 13 2 86 " 7-24 11 100 " 7-26 16 6 73 " Disturbing noise. 7-27 IS 5 78 SunUght. 7-28 14 1 93 Arc " 7-29 15 4 79 Sunlight. 7-.30 17 4 81 Arc 7-31 14 3 82 Sunlight. " 8- 1 15 3 83 " 8- 4 13 1 93 " 8- 5 13 3 81 " 8- 6 13 1 93 " 8- 7 18 2 90 B — With Stimulus Variable. 8- 8 12 1 92 Sunlight. Max. Min. 8-10 18 2 90 " ' 8-12 17 6 74 Arc " Max. 8-13 15 1 94 Sunlight. 8-15 8 4 66.6 Monkey growing very careless. I pulled him back vigorously so as to punish when errors were made, hoping thereby to obtain a more careful choice. 8-16 Min. blue was thrown in for the first time. Monkey was entirely confused. After jerking him back vigorously for several trials, he began to go always over to left, out of range of beam; then, after remaining still for a moment, he would suddenly thrust out his paw to open left-hand box regardless of color exposed there. I tried in many ways to overcome this position error, even to the extent of allowing him to react as in the beginning, to max. yellow and blue with no changes, but the error was not overcome in the time at my disposal. Watson, Color Vision in Monkeys. 27 TABLE V. H's Reactions to Blue-Yellow. A — ^With Stimulus Constant. Date. Blue. Yellow Per Cent. Correct. Source Intensity. Remarks. Blue. Yellow 7-14 3 2 60 Arc Max. Max. 7-15 4 5 44 7-16 4 8 33.3 7-17 10 10 50 7-18 8 3 73 " 7-19 4 4 50 7-20 7 5 58 7-21 8 6 57 7-22 11 5 69 7-23 n 6 65 7-24 12 3 80 7-26 12 3 80 7-27 18 5 78 Sunlight. 7-28 9 6 60 Arc ,, 7-29 15 4 79 Sunlight. 7-30 14 2 87.5 Arc 7-31 14 4 78 Sunlight. First 3 choices wrong. 8- 1 12 1 92 '* 8- 2 17 2 90 " 8- 3 11 5 69 Made very angry by being jerked back on errors. 8- 5 13 2 86 " 8- 6 17 3 85 '* 8- 7 10 100 B — Witl^ Stimulus Variable. 8- 8 8-10 8-12 8-13 8-14 8-16 8-16 8-16 8-17 8-17 8-18 8-18 8-19 8-20 11 14 10 15 10 1 21 5* 8 8 1 12 2 1 13 2 i 8 ! 8 10 15 100 100 100 100 90 81 100 89 86 86 100 100 100 100 Sun Max. Min. light. . .. ^^ Arc " Max. Sun " ' light. Min. " 1 Max. Min. 1 Max. J " I " Min. " i " Max. Changes in intensity made in midst of series. Changes in intensity made in midst of series. Half vertical strip of blue shown. Lower half of blue shown. Surface value of colors altered by pasting tissue paper over glass^ Two of the five errors were made when the sun was overcast. 28 'Journal of Comparative Neurology and Psychology. TABLE VI. B's Reactions to Blue-Yellov A — With Stimulus Constant. Date. Blue. Yellow Per Cent Correct. Source Intensity. Remarks. Blue. Yellow 7-14 2 3 40 Arc Max. Max. 7-15 6 4 60 " " 7-16 7 3 70 ' " 7-17 8 4 66.6 " 7-18 16 2 89 " " 7-19 12 100 " " 7-20 18 2 90 " " 7-21 14 1 93 " " 7-^2 12 1 92 7-23 12 1 92 " " 7-24 13 100 " 7-26 12 2 87 (( " 7-27 18 2 90 Sun " 1 7-28 7-29 17 15 5 2 77 88 light. Arc Sun " Change to fainter spectrum seemed to be noticed. 7-30 14 100 light. Arc .. <. 7-31 14 100 Sun " 8- 1 9 100 light. .. .1 8- 4 13 1 93 " " " 8- 7 17 100 B — With Stimulus Variable. 8-10 8-12 8-13 8-15 8-15 -16 -16 -16 -16 -17 -17 -17 -18 -18 -19 -20 14 16 1 ' 13 1 15 10 13 3 17 1 8 5 8 17 17 14 8 8 10 15 100 94 100 100 100 81 94 100 100 100 100 100 100 100 100 100 100 Sunhght. Max. Min. Arc Sun1 ght. Min. Max. Max. " 1 Min. Max. Min. Max. Min. J Max. 1 Min. 1 Max. J By mistake on this day the episcotister, set in the tests on the other animals usually with a 30° opening, was clo-sed to 10°. The animal dashed to the yellow on his first two trials, then became steady and made only one more error in the series. Filmy clouds were passing over the sun and at times the blue was barely over my own threshold. Changes in intensity made in midst of series. Changes in intensity made in midst of series. Half vertical strip of yellow shown. Half vertical strip of blue shown. Surface value altered by pasting tissue paper over surface of ground glass. TirE EXPKESSIONS OF EMOTION IN THE PIGEONS. I. THE BLOND RIXG-DOVE (THrtur risorius). BY WALLACE CRAIG. Fvom the DcpartDiciit of Zoology of the University of Chicago. With One Platk. CONTENTS. PAGE lutrodiiftioii — I'urposes of this sludy — Aekiiowledgeineiits .30 Description of sounds made by the blond ring-dove, and act-ompanylng movements ;^1 Prefatory remarlvs yi 1. Silence 33 2. Fear 34 3. Alarm SH 4. Cry intermediate between the alarm and the kali 37 5. The kah 38 ( I ) The ordinary-kah 3!) ( II ) The kah-of-excltement 4U (III ) The copulation note 42 G. The charge 42 7. Parent's call when ready to feed the young 43 8. Cry intermediate between the kah and the cot) 44 9. The coo 44 ( I ) The perch-coo, or the song 47 (II ) The bowing-coo 48 (III) The nest-call 51 Life-history of the blond ring-dove ,53 A.' Beginning of the life cycle 54 Development of the young 54 First appearance of the cries of the adult : the alarm-note and the kah 5!) The change of voice (JU First appearance of the coo G2 Influence of old birds 64 The charge 65 Summary of development 66 B.' Beginning of the annual cycle 66 Wooing and mating 67 C Beginning of the brood cycle 69 Copulation, nest-building, and egg-laying 69 D. The daily cycle 70 Changing places on the nest 72 C." The brood cycle, continued 74 Incubation, brooding, and preparation for a new brood 75 The; Journal of Comparative Neurology and Psychology. — Vol. XIX, No. 1. 30 "Journal of Cojiiparative Neurology and Psychology. B." The seasonal cycle, continued 78 The lapse of broocUng and loss of voice at the end of the breeding season IS A" The life cycle, continued 7S The prime of life, and old age 78 Sunnnary of the life-history 79 Intkoductiox. This study of the behavior of i^igeons was undertaken seven years ago with intention^^'rs^, to describe the various sounds produced by one type species of pigeon, and the bodily movements which invariably accompany the utterance of the sounds ; second, to compare these with the sounds and movements of all the species in Professor AVhitman's large collection of living pigeons; third, to throw light upon any problems which seemed naturally to connect with the study as it progressed. The present paper is drawn up to fulfill the first of these intentions. It is a descriptive account of the vocal and bodily expressions of emotion in one species chosen as a type. It may make this first j^aper more valuable if I indicate briefly the nature of the work by which it is to be followed. The following is a brief outline of the whole. 1. Description of the vocal and bodily expressions of emotion iii the blond ring-dove. Life-history of this species, in so far as it c(mcerns the use of voice and accompanying gesture (present paper). 2. Comparison of the sounds and gestures of different species ; showing specific characteristics, homologies, and the possibility of voice and gesture throwing light on problems of phylogeny. 3. Irdieritance. The forms of expression in jiigeons are strictly hereditary. They are not learned by imitation (coj^ying). In hybrids the voice is intermediate, except when it is im])erfectly develo])ed. 4. Variation. Comparative study shows that the vocal utterances v^ary from group to group in a manner indicating detenninate or orthogenetic variation. 5. Selection. Pigeons are subject to sexual selection of a kind more or less like that described by Hacker.^ But the theory of ^Hacker, Valentin. Der Gesang der Vogel. .Tena, 19<)0. Hacker's statement is an improvement upon that of Groos, in his "Die Spiele der Thiere," Jena, 189G. Craig, Expressions of Efuotioji lu Pigeons. 31 sexual selection takes account of only a fragment of the great utility of the voice. Voice and gesture are of prime importance through all the cycles of the life-history. This is shown, but not at all fully explained, in the present paper. It is somewhat further shown in the following (N^o. 6). 6. Sociology. A preliminary account of the sociologic interpretation of pigeon behavior has already been published.^ 7. Psychology. The psychologic conclusions are so numerous and so intimately connected with the details of description, that it is impracticable to summarize them in this place. My indebtedness to Professor, Whitman is so evident from beginning to end of the paper that there is no need to speak of its details. 1 wish, however, to acknowledge in gratitude the two chief debts I owe to him. In the first place, Professor Whitman knows the emotions, the voices, and the gestures of the pigeons very much better than I do; he has told me a great many facts about the birds which my more limited experience has not afforded; and he has always given helpful answers to my questions as to what a bird is thinking about when it does a certain act. In the second place, more important than the facts, I owe much to Professor Whitman for the influence of his spirit of research. Enthusiasm and steadiness of labor, sympathetic insight into the animal mind, patience with details, yet a constant reference to general problems, I hope I have learned to some degree. I wish to express grateful obligations also to the University of Chicago and to the Marine Biological Laboratory at Woods Hole, especially for that freedom which allows a student to develop his own ideas. DESCRIPTION OF SOUNDS AND ACCOIVIPANYING MOVEMENTS. Prefatory Remarks. There is among scientists a widespread impression that bird-songs are not susceptible of accurate description. But this impression is ^The voices of pigeons regarded as a means of social control. The American Journal of Soeiologj-, Vol. 14, 1908, pp. 86-100. 32 Journal of Comparative Neurology and Psychology. certainly erroneous, at least so far as the utterances of pigeons are concerned. It may be granted that the qualities (timbres) and the intensities of sounds cannot be accurately determined outside of the physical laVx»rator}' ; but in this respect the qualities and intensities of sounds are not very different from shades of color, feeling to the fingers, and many such vague impressions which are used in so-called accurate description. Those features of sounds which have to do with pitch and with time, on the other hand, are as susceptible of accurate description as are the forms and dimensions of visible organs. It must be remembered, too, that for the purpose of comparative study, a description need not go minutely into every detail. This study is to include a comparison of each utterance of the ring-dove with other utterances of the same bird, and with corresponding utterances of the opposite sex, of the young, and of different species. For all of these comparisons it suffices to have a general knowledge of each utterance as regards timbre and intensity" and an accurate knowledge of each as regards pitch and time. For the study of expression, on the contrary, it is desirable that intensity and timbre, in addition to the other two sets of characteristics, be measured with extreme accuracy. Such a work of measurement, for one species of pigeon alone, would involve years of labor ; indeed, such work is just beginning to be done, and its methods are just beginning to be developed, even for the human voice.^ Hence we must be content for the present to describe the changes of expression in the dove's voice by means of non-quantitative musical signs and popular language ; and though these means of description are broad and indefinite, they are full of meaning, and may convey a good idea of expression. There is only one point in which I have found it necessary to depart in any way from the regular musical notation. That point concerns the glide, or portamento. Pigeons' notes very commonly glide with absolute continuity from one pitch to another. I have not been able to find any convenient musical sign which indicates such a glide, as distinct from a mere legato; hence, I have adopted the 'E. W. Scripture. The Elements of Experimental Phonetics. New York, 19(J2, pp. xvi -h 027, PI. xxvi. Craig, Expressions of Emotion in Pigeons. double slur, thus. 33 ^^ ^ which must always be understood to mean a perfect glide, or portamento. 1. Silence. Many birds, especially among the Oscines, are uttering some sound continually, being silent only when they are asleep. For example, the various species of American blackbird (as, Quiscalus, Agelaius) repeat their chuck" so frequently, both while flying and while perching, that the presence of a flock is always made known to the ear at a considerable distance. The Fringillida3, similarly, are ever repeating a short "chip" or "chirrup." But the ring-dove has no such incessantly repeated note. The dove's notes are voiced only when prompted by some form of excitement. When engaged in any nonsocial occupation, such as eating, drinking, preening its feathers, or merely resting, the ring-dove is silent. And when on the wing, even in the midst of excitement, the blond ring-dove never utters any sound, except on rare occasions (only one occasion within my experience) an apparently involuntary grunt. Some other forms of birds even prostitute their most useful notes to purposes of play. The blue jay (Cyanocitta cristata), for example, often gives alarming cries when no danger is near, and seems to enjoy, so far as the limits of avian intelligence wall allow, the consternation which it can thus produce among its feathered neighbors. But the pigeons, perhaps on account of their lower grade of intelligence, are incapable of carrying play to such a point; they never use the alann-note except when really alarmed. Certain of the dove's calls are given at times in a manner which might be styled half-serious, half -playful ; but of the utterances of the adult ring-dove, there is only one (the song, p. 47) which ,ever appears to be given and enjoyed purely for its own sake. 34 'Journal of Comparative Neurology and Psychology. 2. Fear. In the case of any object threatening or frightening a ring-dove, the bird being, for one reason or another, disinclined to turn tail and flee, it exhibits attitudes and movements of terror and anger which we may call, for the sake of brevity, the expression of fear. The reasons why the bird may be disinclined to turn tail are numerous : it may be young and unable to fly ; sick or wounded and hence unable to effect a speedy retreat ; it may be defending its nest or its mate ; or it may be simply quarreling with a neighbor on equal temis. In all such cases the dove shows the expression of fear, which is now t.j be described; these cases are to be distinguished from those in which the dove uses its energies merely to escape and fly away, for then it shows a very different expression which we call alarm and which will be described later. The expression of fear in the ring-dove is not at all peculiar to the species ; it is essentially the same as the expression of fear in all birds, and very similar to the expression of fear in reptiles and mammals. It consists chiefly in the erection of appendages — bristling the feathers, spreading the tail, lifting the wings — and in the emission of threatening sounds. It should be noticed that in the expression of this emotion all the feathers are raised to the utmost degree ; in sudden fright the tail also is wudely spread. The wing nearest to the feared object is raised and is used to strike with, dealing blows of great power and of such swiftness that, if a man allows his hand to be struck at, the hand feels the blow before the eye can see it. In some cases the near wing alone is raised, but in many cases the two wings are raised symmetrically. The head is drawn in close to the body, but is always turned toward the object of fear., ready to deal blows with the beak. The eye assumes a ferocious glare, utterly different from its ordinary mild look. This great change in the eye is caused largely by the following conditions. The eyelids are drawn back so as to open the eye to the widest limit ; such wide-openness gives a staring appearance to the eye of any creature, and gives to the eye of the dove an especial glare, by exposing the maximum of the fiery red iris. The black pupil of the eye also remains large, not contracting to a pin-point as in the case of some of the other emo Craig, Expressions of Ef?iotioti in Pigeons. 35 tions. Since the feathers of the head are slightly raised, they swell out around the eye and give somewhat the effect of a frown. Whether these changes are sufficient to account for the total change in expression of the eye, I do not know. The important point, in any case, is, that the change in expression of this one important feature is just as marked as the change in appearance of the bird as a whole. The sounds emitted under the influence of fear are various. Those ordinarily given in these circumstances are a hiss and a snapping of the bill. Both of these, however, are so feeble that they appear to be but impotent relics of a once powerful snap and hiss. They are so feeble, indeed, that an observer recognizes them by seeing a pufiing movement and seeing the bill close, rather than by hearing either the hiss or the snap. A dove in extreme terror (especially when acting with tense exertion, as when a wild one is struggling in the hand, may utter powerful grunts or even a sort of scream. But in ordinary cases the frightened dove expresses all its emotion in the attitude and the movements of defense, not in any effort to make sounds. It appeals strongly to the eye but not at all to the ear. 3. Alarm. Fear is the emotion shown by doves toward an enemy at close quarters ; alarm is the emotion shown toward an enemy, or a possible enemy, in the distance. The alarm-note is heard many times every day, for the doves are always on the lookout for dangerous-appearing objects, especially for the appearance of a hawk in the sky. The bodily changes in the expression of alarm (Plate I, Fig. 1) evidently serve two purposes : — first, to prevent the enemy, so far as possible, from seeing the dove; second, to allow the dove, on the other hand, to see the enemy. The first purpose is served by changeFs which reduce the apparent size of the body to a minimum : the contour feathers are all appressed until they lie as close as possible, the tail and wings are closed, and the wangs pressed tight against the sides. The second pur])ose is served by the bird standing high on the legs and stretching the neck in a manner which shows that a great strain accompanies the concentration of attention upon the alarming object. 36 'Journal of Coinparative Neurology and Psychology. This stretching to full length, combined with the reduction in girth due to appression of the feathers, makes the bird look ghastly thin. The stretching is usually upward and forward, but not always so; for when intervening objects partly obstruct the view, the dove may stretch its neck backward or somewhat to one side. This goes to show that the emotion of alarm depends, not upon the assumption of one specific position, but upon stretching in general. The expression of alarm includes the utterance of a very distinct cry, a cry of great utility, because it communicates the alarm to all pigeons within hearing. This cry of alarm is a single, short, emphatic note. Its chief characteristic is its emphasis, and it becomes more and more emphatic with greater degrees of the emotion. Its emphasis depends upon the fact that it is evidently made with effort. If the birds are much startled, and not sitting on the nest (which would have the effect of making them more quiet), they, give a loud sound which reaches a hi2;h key, thus: ^ ye - 1 Short and vehement ; abrupt rise and abrupt fall. Timber : chest-tone. ha I Rate: 4 crotchets per second. A clarinet tone. Beginning abrupt, loud, explosive. Fall in pitch abrupt. With a less-alarming stimulus, the note is less loud, and lower in pitch, with less rise and fall, thus : Craig, Exprrssiotis of Emotion in Pigeons. 37 When the birds are on the nest they are scchisive in their actions, and, accordingly, disinclined to make a noise. Hence, the alarmnote sounded by a bird on the nest is always less loud, even though the emotion be extreme, as shown by the tension displayed in the bird's attitude and movements. But notwithstanding the lower intensity and pitch, the quality of effort is shown in this alarm as clearly as in the louder ones. Just what it is that gives evidence of effort, it is not always easy to say. But, for one thing, the note usually begins and ends abruptly; and such abrupt beginning and ending is an equivalent, so far as expressiveness is concerned, of the abrupt rise and al)rnpt fall of the notes represented above. For another thing, this note, when intense, has a hoarse sound, a sort of stage w^hisper" effect, like that produced when one contracts the chest strong-lv but obstructs the breath in the vocal organs ; and this is no doubt what occurs in the bird, for one can see the hard breathing movement of the body wdien the note is sounded. This subdued alarm is given by either the male or the female, when sitting on the nest, or when near the nest containing the eggs or young. No.4. t8 VE li Timbre : approaching that of an electric buzzer. Intensity: So low that sometimes, I judge, it was inaudible at 2 yards distance. NO. 5. ii^ A pure, resonant chest-tone. Not loud. 4. Cry Intermediate between the Alarm and the Kah. The specific cry of the ring-dove to be described next, which I have named the kah, resembles the alarm-note in tone but differs > 8 Journal of Comparative Neurology and Psychology. from it in inflection and in being divided into a series of separate sonnds. The kab, being generally a social call, is widely, different from the alarm-note in meaning. But on rare occasions, on one occasion in particular, I have beard a ring-dove give a note which seemed to be an intermediate between the kab and the alarm. On the particular occasion referred to, when the bird heard other doves fighting in a neighboring cage (being unable to see them), it appeared to be moved both by alarm and by a social attraction toward the other doves, and it gave, several times, a sound which was intermediate between the alarm and the kah ; thus : The second note may be tlistinet. or may be, as it were, merely the declining end of the first note. The fact that the dove can give an intermediate between two sounds which usually are so distinct, shows that the bird has more freedom in the use of its voice than might at first be supposed. 5. The Kaii. The kah is a note which, in speaking familiarly about the doves, we often designate as the laugh," because it resembles the laugh of a young child so closely as to suggest that sound to anyone who hears the kah for the first time. But I have avoided calling it the laugh" in this paper, because, though the cry sounds like a laugh, I must gaiard against leading the reader to think of it as a laugh in any other sense. The kah consists of from three to ten notes, like kah kah kah kah kah," in an uninternipted series, all the notes being, as a rule, closely alike. The timbre is a chest-tone, with a sort of nasal twang, suggestive of the harder notes of a clarinet. The various forms of the kah may be imitated with much accuracy by the human voice. This cry is an exceedingly variable one, for it differs greatly as Craig, Expressions of Emotion in Pigeons. 39 given by difforoiit individuals, and in the cas(? of each individual it varies according to the circumstances under which it is given, I shall describe three types: (I) the ordinary-kah ; (TI) the kah-ofexcitement; (III) the copulation-note. The first two of these, however, are not two specific cries ; the}^ are but two types chosen to represent a perplexing multiformity of sounds which might, perhaps, be as well represented by a greater number of types. (I) The ordinary-hall. — In order to give a general statement which will include all the uses of this utterance, one may say that it is a greeting. If translated into English it would be hallo," with varying intonation. It is given by a dove upon rejoining the flock, or upon alighting on a perch where another ring-dove is sitting, or upon seeing a friend after a short period of separation, or upon going to the nest where the mate is sitting, especially when the intention is to exchange places with that bird upon the nest, or upon going to the mate with intent to caress it, as by preening its head. Conversely, a dove may give this call when it is approached and caressed by the mate ; when, sitting upon the nest, it is approached by the mate Avith intent of taking its turn in sitting ; and so with the reciprocals of the other situations mentioned. These uses are all obviously social. The cry is given also at certain conjunctures which do'not necessarily involve other birds : thus, the kah is often given by a dove when it alights on a perch, even if there be no other birds near; or it is given when the dove goes to its nest and eggs, even if the mate be not near. But the use of the kali at these conjunctures is evidently an outgrowth from the social usage ; and, moreover, even in any one of these situations the probability of its use is much greater when other birds are present. The dove sometimes voices this sound when, seeing food put into its cup, it comes to eat; in this case the cry prol)ably has some social reference to the man who has brought the food. For this cry of greeting is very often used, toward the dove's human acquaintances ; and a dove that has been long isolated from its kind will give a kah of greeting to any human being who comes near. The sound of the ordinary-kah is distinguished from that of the kah-of-excitement by being light, free, and careless in expression. 40 'Journal of Comparative Neurology and Psychology. The ordinary-kali is less loud than the other. It is shorter and quicker, more brisk, consisting of only 4 to 6 notes, rarely only 3, and these notes typically staccato. The pitch is in some cases sustained without alteration through all the notes ; when there is any change in pitch it is not of the wailing, chromatic type which characterizes the kah-of-excitement, but it is of a bright and cheery sort. Here are some typical examples of the music of the ordinary-kah. No. 7. ^^ No.e. 9. i^m a ^ {II) The hah-of -excitement. — The ordinary-kah passes by all possible gradations and by numerous variations into that general type which I have named the kah-of-excitement. The kah-of-excitement, or an approach to it, is given in sundry situations of which I shall try to give a general description in three groups (a), (b), (c). (a) The kah-of-excitement is given at the same conjunctures as is the ordinary-kah, if only there be more excitement than common, due either to outward circumstances or to the bird's o^vn inward state. Aw instance of outward circumstances occasioning excitement, may be found in some cases of a bird going to relieve its mate of duty on the nest. At this conjuncture, in ordinary cases, the bird gives the ordinary-kah ; but if it finds the mate unwilling to leave the nest and stubbornly opposing the change, it may give a more excited kah. As to causes of excitement within the bird itself, in general it may l)e said that the ordinary-kah is given more by the female, the kahof-excitement by the male ; the ordinary utterance in winter, the excited utterance in the breeding-season ; and, within the breedingseason, the ordinary fonn on days when the birds are quietly incubating or brooding, the excited form on days when they are pairing, or preparing for a new brood, or when, though incubating or brooding, they are disturbed by the presence of other birds, (b) The kahof-excitement is used under the same circumstances with the charge, which is to be described presently. The cry may be given before. Craig, Expressions of Emotion in Pigeons. 4^ chiring, or after tlie charge, or it may be given independently of the movement. It is sounded both in charging an enemy and in charging the female, (c) The kah-of-excitement often serves as a prelude to the boAving-coo. In some cases, this use is not to be distinguished from (b), the bowing-coo being preceded not only by the kah, but also by the charge. But in other cases, when the bowing-coo is not connected wdth a fight or a chase, but arises from a sudden inward impulse to express the feelings, it may be preceded by a kah-of-excitement unaccompanied by the charge, the apparent use of the kah being merely to introduce the bowing-coo. The sound of the kah-of-excitement is far different from the gentle tone of the ordinary-kah ; it is a strain expressive of high emotion and tense effort. The excited utterance differs from the ordinary in much the same way that the wdiining tone of an angry child differs from the child's ordinary speech. The details of the change causing this heightened expression are not always the same, but I shall descril)e what seems to Ix^ most characteristic in the cases of the individuals that I have studied. The kah-of-excitement is nearly ahvays of longer duration than the ordinary sound; it consists commonly of 5 or 6 notes, sometimes as many as 10, and the notes are, in some individuals, long drawn out. The effect is usually legato, in contrast with the staccato of the ordinar>'-kah. The sound is louder than the ordinary-kah, and the pitch higher, the loudness and pitch rising with each rise in excitement. The strain always descends in pitch toward the close ; it generally rises at the beginning, culminates somewhat before the middle, and then descends ; but in some cases it begins with the highest pitch, maintains it for two or three notes, and then descends. The changes in pitch are usually chromatic, the interval from one note to the next being a semitone or even a fraction of a semitone ; this is one of the chief causes of the emotional, wailing expression. N0.9. Q. ' i\^^ rii^ NO.tO. 4^p[?[»|tq»l*|[y»^p ii w^ 4: ^: JS P 2l: 42 'Journal of Comparative Neurology and Psychology. NO. 1. -(f ^^J ^^ = — ti NO.I2 a ff -f v ^ ^[>pia>p p ^p^^^ Time : 3 crotchets per second. (Ill) The copulation-note. — In all species of pigeons which I have studied, the act of copulation is immediately followed, on the part of both male and female, by the assumption of a singular attitude and the emission of a copulation-note. In some pigeons this note is perfectly distinct from any other utterance of the species. In the blond ring-dove the copulation-note very much resembles the kah uttered on other occasions. Comparing it with the ordinarykah and the kah-of-excitement, I should say that it more closely resembles the former. i NO. to. NO.I4. ^^^^i^ t^ p Time: ZV-j, crotcliets per second. G. The Charge. It has already been stated that the kah-of-excitement, whether uttered to an enemy or to the female, is often accompanied by the charge. In charging (Plate I, Fig. 2), the male raises his body high above the ground, by extending the legs and standing on tip-toe. The body, thus elevated, is directed horizontally, with the head pointing straight before, and the tail straight behind, as if to cleave the air with thea least resistance. The feathers of the rump and the lower back are raised and held stiffly erect, and the feathers of the wing (both quills and coverts) are slightly spread out. The feathers on the fore part of the body, in contrast, especially those on the head, are smoothly appressed. The reason for this contrast, I think we can explains as follows. The bristling feathers on the back and wings have the same effect as in the case of the expression of fear; they make the apparent size greater and the aspect more terrible. On the other hand, the smooth outline of the head and neck gives a more pointed appearance to the oncoming charges; gives more free play to the neck and head in pecking or in avoiding the pecks of the Craig, Expressions of Eniofioii ui Pigeons. 43 adversary; and also allows the flashing eye to produce its most powerful effect. For the bright red iris, during the charge, expands to a niaxiniuni, reducing the black pupil to a pin-point ; and the eye, thus transformed to a fiery red orb, is a focal point in the appearance of the charging cock-bird. Having assumed this striking aspect, the dove charges at his utmost speed upon the bird which has roused his passion, uttering at intervals the long, loud kah-of-excitement. If confined within a cage and thus separated from the bird which has excited him, he charges up and down the cage, back and forth in all directions, now stopping to stand and glare, for a few moments, at the other bird, now starting again on his mad career. Though on all ordinary occasions the dove's gait is a walk, during the charge he often progresses by long leaps ; the leaping might be accounted for as being a necessity at the great speed at which the bird charges, but the great bounds are useful also in that this uncommon mode of locomotion contributes to the expression of irresistible energy and reckless determination. The charge, as the reader may have gathered from what has already been said, is an activity of the male bird especially. It is used in attacking or driving away rivals or enemies. It is used also in driving the female; sometimes, as an expression of jealousy, in driving the female away from other males ; but in other cases in driving the female to no apparent purpose except to express the male's inherent quality of maleness and his mastery over the female. Though this behavior is rarely seen in the female, she may be observed to charge in some cases, especially in either of the two following circumstances. First, when the safety of her nest and eggs (or young) is jeopardized by the approach of strange birds; secondly, when a female has been kept long in isolation, in which case her behavior comes to resemble that of the male not only in this l)ut in many other aspects (cf. p. 46). 7. Parent's Call When Ready to Feed the Young. Each species of pigeon has a call by which the parent signifies to the young, his (or her) readiness to feed them. This call is always more or less distinct from the other calls of the species. The signal 44 'Journal of Comparative Neurology and Psychology. is of use to birds kept free or in a large pen, where parents and young are liable to be at a distance from one another, when the former are fed by their master; but the same signal is not often observed in caged birds, on account of the young being always at the parents' side and ready to feed, not needing to be called. On this account- 1 have had but few opportunities of studying the call to the young to feed ; but I have heard the note in a few cases, in each of which I noticed that it bore a general reseiublance to the kah, and in one case I was able to study it in some detail. In this case the call was given in the same tone of voice as the kah, but was usually a single note, long-drawn-out and very plaintive ; it was given by the father bird, and was repeated by him even in the intervals between retchings in feeding the young. Each time the father began to give this note. one of the young, being more energetic than the other, ran at once to the parent and received the lion's share of food. 8. Cry Inter]mediate Between the Kaii and the Coo. Occasionally the male gives a cry which has precisely the character of the kah except for the timbre, which is a head-tone precisely like the tone of the coo. I should name this cry, were it not for the strange sound of the appellation, the coo-toncd kah. It is heard not uncommonly Avhen a ma^e gives first the kah-of-excitement and then the bowing-coo, this intermediate sound forming a transition from the former to the latter. It is heard also, though more rarely, unconnected with the bowing-coo. This intermediate cry, like the intermediate between the alarm and the kah, is of interest as showing that the vocal reactions of the ring-dove are not so definite and invariable as one might suppose. 9. The Coo. The coo is, in several respects, more musical than any of the sounds previously described. For, first, the coo is more deliberate, the other notes being more hurried ; second'y, the coo is more formal, more fixed, more definite in pitch and in pitch-intervals ; thirdly, the coo is in a head-tone, which is more musical than the chest-tone of the alarm-note and the kah. Craig, Expressions of Etnotion in Pigeons. 45 Each coo consists of three syllables, which may be represented as cook coorr roo (or in German, which has the advantage of a definite pronunciation, kuhk kuhrr riih). The first and the last syllable are the emphatic syllables, the middle one sounding like a connective between the other two. The emphasis of the first and the last syllable is usually accompanied by a heightened pitch, there being a fall in pitch, usually somewhat abrupt, from the first syllable to the second, and a rise, usually a gradual one, from the second syllable to the third. The sound represented by the letter r has nothing in common with the r as pronounced in most parts of the United States ; it is a distinctly rolling sound; yet it is not even like the r as rolled by the tongue, but like the rolling sound produced by the uvula. The rolling produced by the dove seems to be merely a rapid repetition of the k sound which is heard singly at three points in the coo (thus, kooJc A'oorr roo) and heard singly also in the kah. If this is true, then the single k sound and the rolling must be produced in one and the same organ ; what organ that is it would not be easy to say, though it is probably the syrinx. In imitating the ring-dove's cooing, then, it is most precise to make the rolling sound with the uvula ; persons who cannot roll the uvula will produce the next approximation by rolling the r with the tongnie. The ring-dove's cooing may be imitated very closely by the human voice, in the soprano register. Any reader of this paper who can read music can produce for himself a sufficiently accurate imitation of the dove's cooing by singing the syllables to the music given, remembering the one peculiarity of notation already mentioned (p. 32). One who cannot read music can have a sufficiently accurate imitation produced for him by any musician, capable of singing soprano, who will read the syllables and the notation given on the following pages. The three clear syllables constituting the coo proper are followed generally, though not invariably, by two guttural syllables which seem to have something to do with drawing in or regulating the breath, though what is their precise function I have never been able to observe. Gutturals similarly closing the strain are heard in the case of many other birds which are accustomed to pour forth the song wath one continuous, tense effort ; for example, such a sound is 46 'Journal of Comparative Neurology and Psychology. heard after the coo of the common pigeon, after the crow of the domestic cock, after the cry of the whij^poor-will (Antrostomus vociferus), and sometimes at the close of the song of the meadow-lark (Sturnella magna and S. neglecta). The guttural sounds following the coo of the ring-dove are usually a distinct enunciation of the syllables "go 0." These gutturals, however, show extreme individual differences in quality, intensity and duration, in some cases the two syllables being reduced even to a single sound. This extreme indi\idiial difference, conforming to no law or system, is one of the facts which indicate that the gutturals merely have to do with regulating the breath, and should be regarded as involuntary after-effects of the coo proper. The coo of the female is always less powerful than that of the male. !Not only is the intensity less and the pitch in many cases lower, but the notes are much shorter, thus destroying the richness of the strain. Often, too, the inflection is lacking, or is hurried and slurred in such a manner as to produce a travesty of the coo of the male. If a female be kept long in isolation, with no chance to satisfy her sexual and social desires, she becomes so self-assertive, bold, and boisterous, that she is scarcely to be distinguished from a male, either by her cooing or by any other form of behavior (cf. p. 43). But when a female is quietly pursuing the normal activities of the breeding-season, her voice is so different from that of the male that her coo alone is usually sufficient to determine beyond a doubt the fact of her sex. Aside from these variations in the case of the female, the coo is, as has already been mentioned (p. 15), a pretty constant sound, much less variable than the kah. In different individuals the coo is slightly different in pitch, inflection, and duration. But the melody, in its main outlines, and the syllabication are, so far as I know, invariable. The general description just given applies to all the coos of this species, but these coos are divisible into three types which are kept perfectly distinct, there being no gradations between them. The chief means of distinguishing these three types of coo is the difference between the bodily attitudes and movements accompanying each. Craig, Expressions of Emotion in Pigeons. 47 Yet the sounds themselves are sufficiently different for a person familiar with the birds to tell, usually, by the ear alone, which of the three coos is being given. Professor Whitman has named the three types respectively, the perch-coo (or song), the bowing-coo, and the nest-calL (I) TJie perch-cvo, or song. The perch-coo of the ring-dove is its song, properly so-called. While singing, in almost every instance, the bird sits on the perch, hence the name perch-coo. Less commonly the bird sings on the ground, and rarely in the nest. 'No special concurrence of outward circumstances is needed to move the bird to sing, for the perch-coo seems to express, in many cases, simply a feeling of good spirits ; in this, the song differs somewhat from the other two forms of coo, which are usually reserved for special occasions or particular situations. The perch-coo is the only utterance of the adult ring-dove that ever appears to be given and enjoyed purely for its own sake (cf. p. 33) ; in this sense it may properly be regarded as "play." Of course the feeling of good spirits which is expressed in this coo is favored by outward circumstances which are comfortable or stimulating. Singing is most frequent in spring time, when the birds are in the fullest vigor and vitality; it diminishes during the summer as vitality is lowered by the exhausting labors of the breeding-season ; it is least indulged in during the autumn, at which time the molt renders the bird somewhat unwell. But in the spring time, the season of lusty singing, the perch-coo may be repeated at intervals from dawn till sunset, irrespective of what passes during the day. The only circumstance which, to my knowledge, acts as an immediate excitant of the perch-coo, is the sound of another bird singing in the distance. There are occasions when the birds answer and re-answer one another for considerable periods, and the perchcoo is invariably used in such cases of simply calling-and-answering. It is thus seen that even this case, the only case in which the perchcoo is directly related to the environment, partakes of the nature of play. No highly specialized attitude or movement accompanies the utterance of the perch-coo, no movement save what facilitates vocalization. The bird simply stands in the normal perching position (Plate I, 48 Journal of Comparative Neurology and Psychology. Fig. 3), braces its muscles for the effort, and draws a deep breath; then, during the expulsion of the breath, the expelled air fills the crop, causing a great round swelling of the fore neck, oxtt^nding to the sides of the neck, the throat, and the upper breast. In the interval between coos, the swelled crop subsides somewhat, and during each coo it is again blown out to the fullest extent. The music of the perch-coo is as described for the coos in general. The perch-coo may be taken as a mean, from which the bowing-coo departs in one direction, the nest-call in another. The perch-coo is often given singly, but usually in series. The number of coos in a series, recorded throughout the early morning (if a certain summer day, varied from ?> to 9, averaging 6. During the evening of a similar day, the number in a series varied from 2 to 5, averaging 4. The following are typical examples of the perch-coo of the male. NO.15 NO.I6. i ^'^ n^'prir^ - i N iiln'^fria cook coorr roo goo ^ - _ _ E" cook coorr roo qbo The perch-coo of the female differs from that of the male, as has already been stated for the coos of the female in general (p. 46), in being less loud, and poorer in both quality and expression. With regard to the perch-coo in particular, it must be added that the female often omits the guttural go o" at the end, and even when she does give the guttural it is usually reduced to one'syllab'e, sounding like "wah." N0.I7. NO. 18. =^ E ^ m <a p q . - ' - r -. _ "t^ cook coorr roo cook coorr roo wah cook coorr roo wah (II) The bowing-coo. The bowing-coo is given on the same occasions with the kah-of-excitement and the charge. These three Craig, Expressions of Emotion in Pigeons. 49 modes of expression commonly follow one another, or alternate with one another, in quick succession. Indeed the bowing-coo, instead of being named thus, might well be called the coo-of-excitement. It differs from the perch-coo, in regard to its use, in much the same way as the kah-of-excitement differs from the ordinary-kah. The perchcoo expresses moderate emotion ; the bowing-coo, excessive emotion. The perch-coo, though heard more frequently from the male, is given commonly by both sexes ; the bowing-coo, under normal circumstances, is given almost exclusively by the male. The perch-coo is never directly aimed at another bird ; the bowing-coo is always so aimed. The other bird at which the bowing-coo is aimed may be another male, whom the cooing bird wishes to fight or drive away ; or it may be a female, in which case the cooing bird may be wooing, or expressing affection, or driving the female away from other males, or (apparently) merely asserting his mastery over the female. It is thus seen that the bowing-coo, like the other expressions of excitement, does not attach to any one emotion nor to any one type of the situation, but is used in case of great excitement due to any cause whatever. As the name implies, the bowing-coo is accompanied by a bowing movement ; the bird bending, at the beginning of each repetition of the coo, to a perfectly prone position (Plate I, Fig. 6), and rising, at the end of each, to an extremely erect position (Plate I, Fig. 5). To speak more precisely, the downward movement is made with suddenness at the beginning of the first syllable of the "cook coor too, goo oj" the prone position is maintained inflexilily during the first two syllables ; the upward movement is made, somewhat less abruptly than the downward niDvement, at the beginning of the last syllable, "roo," the erect position is stiffly maintained during the last syllable of the coo and the guttural "goo 0." While in the erect posture, the dove lifts its feet, right and left alternately, high above the ground, as if marking time. Its crop, during the whole performance, is swelled out as in the perch-coo, or even more so. The feathers on the head, neck, and breast, are smoothly appressed ; but the feathers on the back are stiffly erected. This arrangement of the feathers reminds one of the feather arrangement of the charge. The sound of the bowing-coo differs from that of the perch-coo in 50 'Journal of Comparative Neurology and Psychology. such a way as to convey a feeling of intense excitement. The expression of excitement is dne to the entire character of the coo and cannot be completely analyzed. However, it may be said that the sound is, for one thing, louder than that of the perch-coo. It is usually higher in pitch, with the changes in pitch greater. Haste, as if the bird were in a great hurry to express all its excitement, invariably characterizes this coo, notwithstanding the fact that the total duration of the coo may be no less than that of the perch-coo. 1'lie series of bowing-coos are somewhat longer than those of perchcoos; on a certain summer's morning, the series varied from 4 to 10 coos each, averaging 6 and a fraction ; and, whereas the perch-coo may be given singly, or only two or three times in succession, I have no record of the bowing-coo being given less than four times in succession save in case of the bird being disturbed when in f\dl opeiation. The following are good examples of the bowing-coo. The guttural "^0 o" at the end is conspicuous, and in extreme cases it is quite loud, this being highly characteristic of the l)Owing-coo. NO. 19. ^7^ ^. ! ^ )-^ i rr i i r •m'^ ^ cook coorr roo go o Time : 3 eroteliets per second. N0.20, 1^^ Ss s J _ 'J l^X cook coorr roo qoo Time: about 4 crotchets per second. (The whole lasts 2 seconds -(-). Accompanied by a more perfect bow than the preceding. These two coos were given by a bird with an unusually clear, resounding voice, a bird who cooed deliberately and musically. But in many individuals the voice, instead of sounding such a deep rich Craig, Es press ions of Enioiioti /;; Pigeons. 51 tone, breaks into a liig-licr, more shrill sonnd. In some cases only part of the notes break into the hig-lier pitch. In other cases, the whole strain is in a high, shrill voice, as in the following example. N0.21 cook coorr koo qo Time : G crotchets per second. Timbre : Falsetto. The female characteristically does not give the bowing-coo. When an adult female has been kept long in isolation, and has in consequence acquired almost identically the masculine bearing and behavior (p. 16), then she gives the bowing-coo like a male. But under any other circumstances, the bowing-coo is heard from her very, very rarely, and when heard it is of a comparativeh^ feeble and perfunctory type. (Ill) The nest-caJJ. The nest-call, as the name implies, is a coo which is given typically in the nest, by either male or female, and serves to call the mate to the nest. Before the nest has been built, when the pair are hunting a nesting-site, the nest-call is used by either bird which has found a likely site, to call the other bird to the sj^ot. On some other occasions, this call may be given by a bird which is not in the nest. But in all cases the calling bird places itself in a corner of the cage, or in the corner formed where a perch joins the side of the cage, or in some such partly inclosed space ; one male that Professor Whitman made very tame would crouch in the hollow hands of his master and nest-call lustily ; it is evident that such hollow places have, for the ring-dove, somewhat the same suggestive power as a. nest. The nest-call is characterized by a distinct attitude (Plate I, Fig. 4), which the bird invariably assumes in giving this sound. The body is tilted forward until in many instances the tail points almost vertically upward, and the head is as low as the feet, or lower; if the bird is standing on the floor, the tip of the bill and also the swelling crop touch the floor ; if on a perch, the head may be held so 52 yournal of Comparative Neurology ami Psychology. low that the eye gazes at one from under the perch ; if in the nest, the position usually is less perpendicular, the tendency being to sink not only the head but the whole body into the nest. When the nestcall is given in a corner, the bird's face is turned towards the wall. The eye is partly closed, and may be entirely closed in moments of ecstasy. The feathers of the whole body are comfortably appressed. During the time this attitude is held, there is always a gentle flipping of .the wings. This gentle wing-flip, considered merely as a movement, is usually quite different from the wing flutter of the young bird begging for food ; for the strokes of the wings are made singly, at as slow a rate sometimes as two per second, and the movement is chiefly confined to the ti]) of the wing. But, considered psychologically, the nest-calling wing-flip bears an unmistakable resemblance to the begging expression, for it is given wath reference to another bird, with a supplicatory significance ; this is ' more evident in the female than in the male, and is seen especially when she is anxious for sexual union ; in such case her wing movement may become a true flutter, and she may flutter separately that wing which is next to her mate, thus exhibiting clearly the similarity between her wungflip and that of the begging young. But, whereas the fluttering -of the hungry youngster is notable for its violence, the wing flutter of the amorous adult is notable always for its gentleness. The nest-call is always given singly. Though the bird may continue nest-calling for many minutes together, there is invariably a considerable interval between each call and the next. These intervals of silence give opportunity for the vocalist to look about and watch the bird to which it is calling, and also to express its feelings by the wing-flip just described; for, during the effort of uttering the coo, the dove must hold its body rigid, the head square to the front, and the wings tight against the sides ; only in the intervals between coos can it give the delicate flipping movement of the wings. The sound of the nest-call differs in several respects from that of the other coos. The nest call is less loud than the others; the guttural at the close is usuall}' omitted altogether. But this is the most protracted of all the coos, its individual notes, especially the last two out of the three, being of long duration ; this protraction Craig, Expressions of Effiotioti in Pigeons. 53 seems to be connected natnrallj with the fact that the coo is given always singly. Further, the notes of the nest-call show greater and more precise changes of pitch than do the notes of other types of coo ; they rise and fall between tones which are distinctly marked and sustained ; they do not, typically, fade into those vanishing tones which make the other types of coo somewhat less musical than this. No. 2 N0.23 -i-zs _ _ _ ZTzr . _ _ cook cooKK roo goo avE cook coor roo qo 6 Sve A break (yodel) between the last two notes. Sometimes tbe musical intervals exact, even tbe "go o." NO. 24 JJ s & ^ P l Y lk^:^^^r¥^f- r7 cook COOKK YOO cook coorr roo Time: about 2 seconds for tbe 3 syllables. Tbere is a "go o" at tbe end, but it is almost inaudible. The guttural go a is usually omitted in the nest-call. LIFE-HISTORY. The life-history of the ring-dove falls in cycles of four orders, one cyc'e within another. These are: (A) the life cycle; (B) the annual cycle; (C) the brood cycle; (D) the daily cycle. (A) The most comprehensive of the sycles, since it comprehends all the others, is the life cycle ; which consists of a period of immaturity, beginning with the ego: and extending through several months of growth and development, and a period of maturity extending through several years. (B) The period of maturity is divided into armual cycles, each consisting of a winter period during which the birds show little or no reproductive activity ; and a summer period, or breeding season, during which the birds are continuously active in the processes connected with reproduction. (C) The breeding 54 Journal of Comparative Neurology and Psychology. season, since it is occupied with the rearing of several successive broods, is divided into as many brood cycles, each consisting of several days of love-making and nest-building, fourteen days of sitting on the eggs, and at least as many days of caring for the young. (D) At the time of brooding, the birds' activities fall into a very definite round, repeated daily, constituting the daily cycle. Of course the dove's whole life is divided into daily cycles of a sort, because the birds invariably roost through the night and do all their work in the day time. But the most specialized daily program, the one which will therefore be described in this paper as a type, is the daily program at the time of brooding. These four cycles will best be treated in the order in which they come and go in the dove's life. Accordingly, the order of the treatment Avill be as follows: — A'. Beeinnina' of the life cvcle. Co o B'. Beginning of the annual cycle. C. Beginning of the brood cycle. D. The daily cycle. C/'. The brood cycle, continued. B". The annual cycle, continued. A". The life cycle, continued. A'. Beginni2\^g of the Life Cycle. The day of hatching. — When hatched, the pigeon is a little blind and naked body, able to slightly raise a shaking and swaying head and open its bill to receive the food regurgitated by the parent, and just able to drag itself slowly, by using the feet, from one position to another in the nest. When a parent wishes to feed a young one, she (or he) puts her lull down to the young one and gently touches its head, or takes hold of its bill. Then the young one raises its head in its shaky and indefinite way and after a number of random movements gets its bill into that of the parent and is fed. It seems as if sometimes the mere movement of the mother in raising her body from the young is sufiicient to cause it to raise its head for food. If the young be touched by the hand at this time their response is either Craig, Expressions of E mot ion ni Pigeons. 55 to raise the head for food or to give a few little jerks of the head which have no apparent significance. They show no sign of fear. To speak of the converse relation, of yonng to parent — a movement of the yonng nnder the mother seems to be a gentle stimulus to her v.-hich may cause her to feed it. Within half a day from the time of hatching, the young may be heard to give a very faint "j)eep," so faint, in fact, that it is inaudible at a distance of three or four feet ; a very brief note, also, for it lasts but a small fraction of a second. By the second day the voice is a little stronger ; it may sometimes be heard at a couple of yards' distance. It is a rather musical, sibilant sound, easily imitated by whistling the letter S very gently through the teeth. There is a gentle rise in pitch at the beginning and a sudden slight fall at the end; the pitch of the sustained part of the note is about b," but is variable. It may be asked if the voice is of any use at this time, or if it is merely undergoing development for the latter period of strength and usefulness. In answer, it may be said that even the young on the day of hatching squeaks when it is hungry and is silent when satisfied, and that even its microphonic voice is useful as a stimulus to the parent birds, including them to feed. The small nestling. — The little bird grows with wonderful rapidity, and its development keeps pace with its growth. On the third day the eyelids begin to become detached from one another, but the eyes are kept closed almost all the time, and the young bird is several days old before the eyes seem to be of real utility. AMien the eyes begin to function, and not before, the young bird begins to show signs of fear. At first, the body is depressed and the head lowered to a hardly appreciable extent. Then, as these signs become more marked, the little pin-feathers begin to be raised slightly from the skin, and the bill is just barely opened and closed again, this movement being the first beginning of the puff or hiss. Each of these signs becomes gradually more nuirked, till finally they develop in the fledgling into a most extravagant expression of fear. The voice alters very little for several days after hatching, though growing in some cases slightly stronger. The large nestling. — About the ninth day the parents begin to 56 'Journal of Comparative Neurology and Psychology. leave the young uncovered part of the time ; at first they leave them uncovered only when they themselves come off to feed, then leaving them for a greater and greater part of the day till finally they cease to cover them at all. This last occurs on the tenth or eleventh or twelfth day. It is not surprising that the parent leaves the nest at tills time, for the young hird has grown so large that the parent ciinnot qwiUi cover it, not to speak of covering two such big youngsters. When there, are two young in the nest, the parents leave sooner, I think, than when there happens to he only one. The nestlings have been as rapid in development as they have been in growth. The feathers on the wings and back have grown sufficiently to hide the skin. The eyes are wide open, and the birds are looking about most of the time; tliey are more active altogether, their activity being evidently in preparation for activity after leaving flu; nest. For days before venturing from the nest the young may stand and stretch themselves in the fasliion of the adult, first i-aising both wings above tlic back, and tlicti stretching each wing in tnr-n together with the corresponding legs. The day before leaving, too, J have seen a yotmg one stand in the nest and flap its wings as if practicing for flight. The expression of fear has at this time reached a maximum. If a hand be brought near, the young bird scpiats down in the hollow of the nest, making itself remarkably fhit, erects all the feathers, draws its head down while at the same time pointing its bill toward the intniding lunid, and repeatedly shuls tlic bill with a sounding snap. The flatlciiiiig of ihc 1)ody is very characteristic of the nestling, for it is absent, in its typical form, in adults, beginning to decline just before the young leaves the nest. The expression of fear thus reaches a maximum development bfd'ore any other, indeed it gradually disappears, to a great extent, as the young becomes older; the snap of the bill, especially, beccnnes weaker after the l)ird has learned to fly, and degenerates in the adult into that rudiment of a snap which, though it can be seen, can hardly be heard. The age at which the young bird learns to fly is, ])retty definitely, the age at wiiich it gives up the expression of fear. When able to fly, the bird shows a'arm instead of fear, in the presence of dangerous objects; it is imjielled to flee rather than to remain and make a show of resist Crak;, Lxprrssions of Emotion in Pigrons. 57 aiiee. In tlio adult, the attitiulo of fear is not assiinircl nnk^ss some special circumstances prevent flight, as when the bird is sick, or brooding. The voice in these large nestlings is still weak and seldom heard, and generally undifferentiated (excejit that the alarm-note may be given a day before the tirst trij) from tlio nest). Begging for food is jnst beginning (o bo accompanied by a slight shaking of the wings, although the bird's po\v(M' of coordination is still so undeveloped that it cannot direct its hill with any detinileness toward the mouth of the parent. Tlie fledgling. — The time of leaving the nest is very variable. The first trip on the floor of the cage may take place at any, age from nine to fourteen days. When the young birds first leave the nest, they always return to it to spend the night, ami in many cases they return to it long before night comes. After a few days, however, they fly up on a perch and settle there for the night, huddling as close as possible to each other, or, if there be oidy one young, huddling thus close to the fatlier. This iirst night o\it of the nest comes ordinarily at the age of fifteen to sevcniteen days; it is more definitely fixed, I should say, than the first trip from the nest, probably because it is detenuiiud bv the crowth of the win<>; feathers and the resulting ability to fly. As soon as the young birds leave the nest they begin to pick up bits of gravel or other small objects from the floor. Tt nuikes generally more tlian one day for them to learn to eat out of the seed cup, and when they do first accomplish this they eat very slowly, as if they had to think about each seed taken. When the young one is well abl(^ to f e( (1 itself, the parents feed it gradually less and less; the amount (vf their tVcMling (h^pends largely, it would seem, on the (damorousness of the young, whi(di, in turn, \-ari(^s inversely to the young bird's ability to obtain food by its ewu efl'orts ; in this way the amount of food given by the feeders is adjusted to the necessities of the fed. As may be sui'niis( d. the time of weaning is verv indefinite. T had one youijo' one which learned to eat very quickly and received only a small part of its nourishment from the parents after the fifteenth or sixteenth day (although the father continued to feed it occasion 58 'Journal of Comparative Neurology and Psychology. ally until the twenty -fourth day). Another young one, in contrast, was only learning to eat on the twenty-third day, begging very hard from its parents, who were by this time unwilling to feed it, and it w^as still fed in a spurious fashion by the father on the thirtieth day. ISTow^, as to the sound w^hich forms an important part of the begging expression. About the time the young bird first leaves the nest tlie voice makes a sudden growth ; the little peep which has been made in begging for food grows much stronger and becomes somewhat squeaky in tone. It continues to grow louder almost as long as the begging note is used, that is, almost until the parents quit feeding. It is at that time a sibilant squeak beginning soft and low, becoming rapidly higher and louder, and then ending abruptly. It may be imitated by whistling the letter S through the teeth, loudly, with the inflection just described. This (K'o. 25 of the musical notations) is the emphatic note, given when the bird is begging hard from the parent. But when its enthusiasm dies down, its note becomes lower, softer, and shorter, until it may, become a gentle "st" like ^o. 28. The pitch, as can be seen, varies through a great range. While it ranges very much low^er than that of the young nestling, I think it may also extend as high as the highest notes of the nestling. A slight shaking of the wings when begging for food is noticed, as was mentioned above, before the young have left the nest. This shaking of the wings becomes more ample as the voice becomes stronger, until, in a pair of three-week-old birds, it makes a lively scene. Each young one stands in front of its father (or mother), sticking its bill into his face and trying to push it into his mouth, squeaking without intermission, wath its wungs htilf spread and flapping strongly, following the father wherever he goes, and running around him if he attempts to turn away, every movement being made to an exaggerated degree. If parent and young happen to stand side by side, as they are compelled to do when on a perch, for example. Craig, Expressions of Eiuotion in Pigeons. 59 then the wing farthest from the parent is shaken very little or not at all, but the nearer wing is shaken strongly, in many cases being spread across the parent's back, slapping him vigorously. N^ow the question arises : Wliy should the three-weeks-yonng need to beg so hard for food, whik> the new-hatched get it without begging? The answer, I think, is not so far to seek. The parents' mouths have become sore from the fre(]uent distension and friction caused by the insertion of the young ones' bulky beaks. I^esides. the parents give a great deal more food to the large young than to the newly -hatched, and they work harder to bring it up from the crop. They are tired of feeding, and will quit, if not imp(n'tuut'd by the young. The less a young one begs at this stage, the less it will receive. First appearance of the crws of the adult: the alarm-note and the Inh. — The first utterance to become differentiated from the begging squeak is the alarm-note. This is first given on the twelfth to the fourteenth day. At that time it has the same pitch and timbre as the squeak of hunger, but differs from the latter in being very short, abrupt, emphatic. It has a quick fall in pitch at the end, and in some cases it seems to have a slight rise at the beginning, though in other cases it appears to be at its highest from the very beginning; its inflection is thus exactly like that of the adults' alarm-note, although its tone is that of the bal)y voice. As the inflection is precisely like that of the adult, so is the attitude struck during alarm, the little fellow standing with neck stretched out, looking at the object, whatever it may be, that has excited the emotion. It must be said, however, that at first the attitude struck is only a slight one, the head being only very moderately raised ; and tlie alarm-note as first given is not nearly so emphatic as in the adult, not so loud in proportion, so to speak. The pitch of this sound, as given in my notes, varies from d' to c." As regards the economy of alarm, parents and young are in agreement from a very early stage. The young give the signal upon hearing it from the adults, and the parents likewise may catch the infection from the young, and in all cases the alarm leads to preparation for avoiding danger either by flight or by hiding (squatting low in the nest, depressing the feathers, and keeping very still). 6o Journal of Comparative Neurology and Psychology. The next utterance to be differentiated is the kah, and its origin resembles that of the alarm-note in that at first it is given in the sibilant baby voice yet with the rhythm and inflection of the adult cry, under the same circumstances as the adult cry, and apparently with the same meaning. Two young birds, of different broods, began to give this call on the twenty-seventh day. Another brood of two birds began, apparently, at exactly the same age, for I find in my notes that on the twenty-seventh day they "have an intermittent call. I|- is in the same tone as the ordinary squeak of the young and hence resembles the contented chirnip of a chicken. It seems to be given when the birds are moving about and sociaVe." This call is given in nearly all the many circumstances in which the adult kah is given, but it is not so commonly uttered upon merely alighting on a percli. It is heard in general, as quoted above, "when the birds are moving about and sociable," and it is heard particularly when the bird charges upon another one, in which case the kah is often followed by the bowing-coo. Like all other utterances given in the baby voice, this kah may be imitated by whistling the letter S loudly through the teeth ; the following notations will funiish a guide in such imitation. N0.29 12* S ^ N0.30. ^111111 Time : 5 crotchets per second. The change of voice. — By the term "change of voice" is meant the change from the high-pitched sibilant voice of the young to the more grave and sonorous tones of the adult. It seems well to introduce this topic here because, while the change of voice affects not only the alarm and the kah but also the coo, which will not be treated until later, yet in the coo it is complicated by the simultaneous occurnmce of great changes in modulation, Avheroas in the alarm and tlic kah the change in pitch is the only change which occurs. The change in pitch does not occur by a gradual deepening of the baby voice; the voice "breaks," just as it does in a thirteen or four Craig, Exjyrrssioiis of Emotion in Pigeons. 6 1 toeii-year-old boy. The first change observed is that the sibilant notes of the young have become impure in tone. The impurity increases until a decided harshness is produced, due to the admixture of low tones with the high ones, making evident the analogy to the breaking of the voice in a youth. The low tones become more and more prominent and the high ones dwindle until, after many weeks, the high tones have disappeared altogether, leaving the voice with purely the adult sound. As to the age at which the change of voice occurs, there is an intimation of the change, perhaps, as early as the age of four weeks, for at that time the pure sibilant has changed to a squeaky tone, less pure than the first, and louder. But the earliest distinct break in the voice occurs at about six weeks. The following notation is to represent the combination of high and low sounds which characterizes the voice at this time. "" J^^ 'h.t^^ ksa ksa ksa ksa ksa Kali on alighting on perch. Pitch of the "s" not definitely determined. The "a" is a hard chest-tone, impure. Each note begins with the sibilant sound but drops suddenly into the lower pitch. As the bird grows older the sibilant is reduced more and more, but many weeks elapse before it has entirely disappeared. I have observed a slight trace of it in the alarm-note, for example, at the age of seventeen weeks. The change of voice is due, no doubt, merely to the development of the vocal organs, just as it is in the adolescent man. This puts it on a different plane from the other developmental changes in expression. The inception of the fear reaction, of the alarm-note, of the kah, or of the coo, and the changes in the form of the coo, must be due to the coming into play of fresh tracts and centers in the nervous system. But the deepening of the voice must be due to 62 Journal of Comparative Neurology and Psychology. changes in the syrinx. Observation of the birds leads me to believe that they have no control whatever over the breaking of the voice ; it is purely mechanical. First appearance of the coo. — The first attempts at cooing usually appear much later than the alarm and the kah. Only in one case, in a bird which showed other signs of precocity, did the song originate on' the same day as the kah, the twenty-seventh day. In the nestmate of this bird the coo was not heard until the fortieth day ; in another bird not until the forty-seventh day, and not decisively until the fiftieth day. The first coo is thus very variable in the time of its appearance. And it is equally variable in its character. The varial)le character of the early cooing is shoAvn in this quotation from my notes. The young bird '"takes few hasty steps toward mother on perch, head directed toward her, giving kah in squeaky voice. He repeats this about three times, then stands up straight and stiff (in attitude of male in the up phase of the bowing-coo), then he bows, down and up several times, making not a sound. He goes through much the same performance two or three times. Little later he gives coo in a purely squeaky voice (pitch a") without bowing." In this case, then, the perch-coo and the bowing-coo apparently developed at the same time ; Imt, while the perch-coo was audible, the bowing-coo, if sounding at all, had not passed the threshold of audibility. In many cases, however, I think that the bowing-coo precedes the perch-coo by a day at least. This was tnie of the nest-mate of the bird just referred to: "After giving the kah in its squeaky voice, bird went through bowing motion, roughly, making just a single short note now and then, pitch g', sometimes two notes, with timeinterval between, and second higher than first by one tone or less." In another case the first coo was a "perfectly nondescript sound. Much resembled its attempts at kah, but notes more irregular, some of them more prolonged." All these accounts go to show that the coo at its first appearance is not only variable but extremely imperfect. While the alarm-note springs into being perfectly formed, as it were, and tlie kah, at its inception, is almost as perfect, the coo, at first, is an insignificant fragment which does not in the least suggest the sound it is ultimately to assume. It seems that in those Craig, Expressions of Emotion m Pigeons. 63 individuals iu which the coo appears very hUe, it is correspondingly well developed when it does appear. Development of rlnjflnii and melody. — The modulation of the alarm-note and of the kah being practically of the adult type from the very beginning, these utterances exhibit no development in rhythm or inflection. The coos are the only utterances which go through a long course of development in modulation. By the third month, the coo has been extended considerably. There is a greater number of notes, and the notes are, on an average, of longer duration. But even at this time the rhythm is so imperfectly developed, so irregular, that it often bears little resemblance to that in the adult song. Moreover, the rhythm varies so much, even in a single series of coos, that one must conclude it is largely accidental. The following are examples. N0.3a, kukuku kuKu u uu ? GOtli day. Nest-call coo. Tiiuo: 3 crotchets per second. Attitude: uest-call. Tune very variable. In fact, no constant tnne at all. N0.33 A N0.33B. &: 3 ^s ^i i 9 l ll' ?a? ?^-M ^ ka u kur ku ku ku ka u N0.33C. i^^ s =^^ 1 T ^ ■' ■ 11 ka u kur ka u ku (^82nd day. 1'hough the coo is so formless at first, it very soon begins to show the general form of the adult coo. It comes to consist uniformly 64 'Journal of Comparative Neurology and Psychology. of three notes, and gradually this trisyllable comes to have exactly the accent, the tone-quality, and the melody of the adult coo. But even after attaining the trisyllabic form the early coo has three definite differences from the adult utterance, as follows: First. The different notes of the coo are separated more in the young than in the adult, often allowing a considerable rest between. This fact, together with the general character of the utterance, gives the impression that the young bird coos with difficulty and at the expense of considerable effort. Second. The rolling sound, represented by the letter r, is absent from the earliest coos, and develops rather slowly, for even when it does first appear it is a perfunctory performance. Third. The appendix to the coo, represented by the syllables "(/o 0," is not given until the age of three or four months, and when it is first given it is only a monosyllable. In addition to the three features enumerated, the juvenile coo is characterized by poverty in the quality of sound and a hurriedness and lack of all beauty in the inflection. As the bird grows older, the coo becomes loud, voluminous, and mellow, and acquires a graceful, gliding inflection, which, without changing the general form of the melody, gives it an entirely new and improved character. The perch-coo and the bowing-coo develop at an equal rate and become practically of the adult form at the age of about seventeen weeks. But at this age the nest-call coo is still decidedly imperfect (at least in the male). All through the development of voice the nest-call lags behind the other coos. This is perhaps because the nestcall is purely a sexual expression, whereas the other two forms of coo are used to express emotions which may be developed before sexual maturity, such as combativeness, or simple good-spirits. Influence of old birds. — Pigeons, young and old, are extremely sensitive to suggestion. The young ones often give a certain note when they hear the parents give it; this is noticed as soon as the first of the adult cries appears, i. e., the alarm-note. The more the young hear other birds, the more they call. Thus the calling of other birds may lead the young to give a certain sound earlier than they would give it if left alone. But the young do not imitate the adults, in Craig, Expressions of Emotion in Pigeons. 65 the sense of copying thein, or learning new sounds. The forms of utterance (herein the pigeons ditier from many other birds) are strictly hereditary. The charge. — The charge is associated in its development with the kah, as has already been stated (page 40), and thus appears at an early age, even at the age of twenty-seven days. The charging activity in the young, as in the adult, includes chasing another bird, pecking her (or him), assuming the peculiar horizontal attitude, progressing by leaps as well as by steps, and uttering the kah (kahof-excitement). But each of these acts is at first of a weak and gentle sort. The charging activity, like the vocal activities, passes through a prolonged and gradual development before it reaches the form seen in the adult. Development of certain other instincts. — Since the first attempts at cooing appear at an age of from twenty-seven to forty-seven days, it might seem that they are too early to have any sexual significance whatever. Yet some activities connected with sex begin at an equally early age, — sitting on eggs, for example. In one instance, on the twenty-first day, a fresh egg having been laid by the mother, the young one entered the nest, observed the egg intently, and carefully sat on it. At fifty-one days, a young one entered the nest, settled very carefully on the pair of eggs, sat for several minutes, and when the father tried to drive it off persisted for a considerable time in holding its position. This sitting on the eggs is not an accident, for the little fellow is very careful to have the eggs under him, and if there are two eggs he takes a great deal of pains in tidying to get them both under, finally settling down upon them with that sidewise rocking movement always seen in the case of the adult. A young female showed the courting propensities of her sex, practising the art upon her father, at a very early age. It is difiicult to say at just what age this began, because it is impossible to draw a sharp line between filial and amorous attentions. At fifty-six days this young female responded to the cooing of the father and some other pigeons by assuming the nest-call attitude, head down and wings shaking, and making an attempt at the nest-call coo. At seventy-four days she showed the typical courting behavior, for in the 66 'Journal of Comparative Neurology and Psychology. evening bj lamp-light she huddled close to the father and preened his breast and neck, sometimes preening her own feathers in that spasmodic manner which is a sign of eros, and sometimes interrupting these proceedings to give the nest-call coo. From this day forth she did not cease to show her readiness and anxiety to mate. SUMMARY OF DEVELOPMENT. ueveiopmenr anu aeciine oi iiie voice and habits of the nestling. and habits of adult. the The change of voice. 1st day. Voice just audible. Voice a stimulus to parent. Young and parent communicate also by touch. No fear. 3d day. Eyes begin to open. About 6th day. Slight expression of fear. loth to 12th day. Young are left uncovered by parents all day. Expression of fear reaches a maximum. 0th to 14th day. Young first leave the nest. Begin to pick up food. Begging begins to be accompanied by shaking of wings. Expression of fear begins to decline. 12th to 14th day. pression of a Alarm-note. Exarm. 15th to 17th day. First night out of nest. 15th to 24th day and later. Weaning. Maximum development of baby voice and of begging behavior. 21st day and later. Sit on eggs. 27th day. The kah. 27th to 38th day. charge. The 27th to 47th day. coo. The 4 weeks. First impurity in baby voice (?). 56th to 119th day. The nest-call coo. 74th day. Female shows courting behavior. 4 to 6 months. Coos all differentiated and perfected. 6th week. Distinct break in baby voice. 3 months. Alarm and kah nearly as in adult. 17 weeks. Still f\ trace, in some cases, of the sibilant. 4 months and Begin to breed. ater. B'. Beginning of the Annuae Cycle. The age at which a ring-dove begins to breed depends upon the season, for the tendency is in all cases to begin breeding in the spring. Birds maturing in the autumn are delayed by the tendency Craig, E.xprrssions of Emotion in Pigeons. 67 to sexual inactivity in winter; and birds maturing in the spring are accelerated, in comparison, by the tendency to begin breeding in spring. The autumn and the early winter are marked not only by inactivity in breeding but also by disuse of the voice ; at least a disuse as compared with its copious use at other seasons. But as winter advances, long before warm weather has definitely set in, a change toward the musical life is noticeable. The voice is used more and more, and it gradually regains the volume of sound and perfection of form which characterize it in spring and summer. Whether the preliminary exercise of the voice aids at all in its develojjinent, it would be difficult to say. The fact is that the perfection of the voice and the tendency to use it arise gradually and coincidentally; and it seems probable that each aids the other. Yet there are reasons for believing that practice has very little effect in developing the voice of the dove. As the birds begin to coo, they naturally begin to coo to each other ; and while the whole pigf'onry bombards the ear with an abundance of sounds, each pen presents to the eye an abundant spectacle of bowing and charging, wooing and fighting, love and jealousy. This may continue a long time before each bird secures a mate. But, to notice in detail the formation of a union between two birds, it is more convenient to st\idy the case of two ring-doves isolated in cages. If a cage containing an unmated male ring-dove be suddenly brought alongside another cage containing another ring-dove, of unkno^\Ti sex, the male becomes highly excited at ouce, and gives vent to his excitement in all possible ways. First he bows and coos with all his might, and he continues to do so for a long time. Then he charges about the cage, assuming the attitude peculiar to the charge, and frequently repeating the loud kah-of-excitement. At intervals he stops to glare at the strange bird and sometimes to peck at it through the bars, but soon he starts again to bow-and-coo and charge. After more or less of this display of aggressive impulse, he begins to show eros, by a certain spasmodic preening of the inside of the wing (a movement which invariably accompanies erotic activity), and \)\ assuming' the nest-callinc; attitude and soundino; the nest-call. 68 Journal of Comparative Neurology and Psychology. If left beside the stranger's cage for some hours, the male must sometimes rest and be silent ; but even the intervals of rest and silence are broken frequently by series of perch-coos. This behavior on the part of the male is useful in that it stimulates the strange bird to respond, and, in responding, to reveal its sex. If the strange bird be a male, it shows similar excitement and aggressiveness. And the two males are sure to fight if they can reach one another. But if the strange bird be a female, she acts far otherwise. She is at first very indifferent, unles she is particularly anxious to mate. And after some days, when she begins to show an interest in the male, she does not give the bowing-coo, nor charge up and dowm the cage, nor show other signs of pugnacity and aggressiveness. So far from tending to aggress upon the male, her conduct is rather an expression of submission to him. She shows a certain excitement; for instance when she utters the kah it is a kah expressive of gentle excitement. But she spends the greater part of her time in alluring the male by means of the nest-calling performance — the nest-calling attitude, seductive cooing, and gentle flip of the wings. She often tries to get through the bars of her cage to the male ; and, failing to do so, she sometimes lies down with one side pressed against the bars. She shows eros by the usual method of preening inside the wing; she may even take the copulation position while the male is cooing and bowing to her. When the male sees the strange bird behaving in this submissive and seductive manner, he loses the intensity of his pugnacity ; though he always continues to be masterful. He spends less time now in the bowing-coo and more time in nest-calling and in trying to get to the female. If the doors are now opened and the bird? allowed to come together, they become mated. The time it takes the doves to become mated varies greatly. In case of some old, experienced birds that are ready and anxious to mate, two or three days in contiguous cages may make them acquainted, and then as soon as the doors are opened and they come together, they are ready to copulate. In other cases, especially in cases of inexperienced birds, the male is so cruel to the female at first that it is not safe to leave her with him until after a Craig, Expressions of Emotion in Pigeons. 69 long period — even weeks — of acquaintanceship. But once the birds have had their attention concentrated on each other and have become affectionate, the business of breeding proceeds smoothly and rapidly. C. Beginning of the Brood Cycle, In the preparation for a brood of young, whether it be the first brood of the season or a later brood, there is always first a period such as has been described, in which the male by means of the kahof-excitement, the bowing-coo, charging upon, the female and even pecking her severely, gains a mastery over the female that draws her attention to himself to the exclusion of all other males which may come ill sight or which may be surviving in the female's memory. The female on her part submits herself to the male and draws his attention to her. And both birds become worked up to a state of tense sexual excitement. This period is ahvays followed by a second period in which the excitement, venting itself in copulation and in work upon the nest, becomes less violent, though perhaps not less powerful. The charge and the kah-of-excitement fall to a very low ebl>, and even the bowing-coo is used much less than at first ; but the perch-coo and the nest-call are in frequent requisition. Copulation is repeated a great number of times, there being many repetitions per day and continuance for a number of days. It is continued until near the time when the first Qs^g is laid ; and sometimes even after the first egg is laid. The number of days of copulation seems to be ordinarily four or five ; but there is at hand as I write, a pair of l)irds still continuing a series which they began fifteen days ago. The number of copulations or attem]:»ts at copulation in one daj^, I have never determined under normal conditions. In certain abnormal experimental conditions, devised for another purpose, I counted on several different days from twelve to fifteen attempts per day. I should think that even in normal cage conditions the number of attempts might be equally great. The first day of copulation is a day of high excitement, and the divers expressions of this excitement may be divided into two classes ; namely, those that occur through a great part of the day in general, yo 'Journal of Comparative Neurology and Psychology. and those that occur immediately before each act of copulation. Throughout the greater part of the day, the male frequently gives the bowing-coo, the nest-call, or the perch-coo, the female gives the nestcall, and both birds kah frequently and loudly. Preening of the feathers in a spasmodic manner, especially the preening of the wing on the inner side, and the preening of the head of the mate as the two birds sit side by side, are equally characteristic activities of the day. Immediately before copulation there is usually a special cooing and a special show of eros by preening inside the wing, and there is invariably the act of billing, the female putting her bill into the mouth of the male, and he disgorging a little of the contents of his crop for her to take. This is the signal, as it were, Avhich is immediately followed by copulation. The search for a nesting-site and the building of a nest, which have been going on at the same time with the operations already described, are accompanied by a great deal of vocal performance, especially nest-calling. Both birds engage in the search for a nesting-site. When either bird has found a likely place, it sits there and nest-calls by the hour. The mate, hearing this call, is drawn to the spot, and then both sit together and nest-call, gently flip their wings, and preen each other's heads for a long period. The construction of the nest is carried on by one bird, usually the female, sitting in the nest and building in the materials which are brought by its mate. Each time the male brings a straw the female receives it with the gentle flip of the wings and the nest-call. When the eggs have been laid, the male and the female take turns most regularly in sitting on them. This fact gives to the daily program during incubation a complexity and definiteness which are not equalled at any other time in the dove's life. The present point in the life-history is therefore a good place at whj<^h to insert a description of the daily cycle. D. The Daily Cycle. The daily cycle of activities reaches its maximum complexity and its greatest definiteness at the time of incubation. At any other time it is indeed noticeable that the birds follow a daily program : Craig, Expressions of Emotion in Pigeons. Ji they always wake with the sun ; they begin to coo before leaving the roost ; they are most active in the early morning, fighting and lovemaking, cooing and calling and running about; they rest, or even sleep, in the hottest part of the day; in the evening they are again active and musical; they go to roost as soon as the light has begun to fade; they frequently coo after going to roost; and, finally, while strictly diurnal in their habits, and helpless at night as are most diurnal birds, yet they may often be set cooing at night by the lighting of a lamp, by a bright moon, by the cooing of oth.er birds, or perhaps by their own inward inclination. In order to obtain a more complete record of the daily cycle during incubation, I have more than once watched the birds continuously for half a day at a time, noting every movement and every utterance, now from before dawn until noon, and again from noon until night. Thus on July 2d I entered the room at 3.50 A. M. The birds were still in their nocturnal positions, female on the nest, and male on the perch. At 3.54, although there was not yet light enough to read by, the male cooed four times ; the female in answer cooed four times ; he cooed once at the end of her song. Then he preened his feathers. Cooing and dressing the feathers always occupy the first part of the day. The coos given were mainly of the perch-coo type, varied at intervals by series of bowing-coos ; but at 5.09 A. M. the male turned on his perch so as to get his head in a coTner," and gave the nest-call, continuing until he had repeated it twenty-four times. Not until 5.36, or one hour and forty minutes after I first heard him coo, did the male come down from his perch to feed. The female greeted him with gently fluttering wings, and as he flew up on the perch and down again, she again gave this sign of feminine affection. Thus matters continued, with but little interruption, until at 8.19 A. M. the male took his place on the nest. In these four and a half hours from the time of waking to that o.f taking his place on the nest, the male repeated his coo 487 times ! He gave about 70 bowing-coos, 386 perch-coos, and 24 nest-calls. The bowing and perch-coos were given in 95 separate series, each series consisting of from one to ten coos. The female during all this time cooed only once ; this once was when, in answer to the male's song of four coos at da^vn she 72 'Journal of Comparative Neurology and Psychology. gave a similar series of four coos. After that she often fluttered her wings when the male happened near the nest, but she never cooed. The male's 487 coos were pretty evenly distributed over the whole time. But they began at the hour of dawn with a somewhat slower rate than the average, rose to a maximum just after the bird had left his roost and breakfasted, and then declined somewhat until the time of taking the nest. The male takes the nest at 8.30 A. M. and keeps it until 4.45 P. M., when he yields it again to the female, who sits steadily until 8.30 the next morning. Of course the birds are not punctual to a single minute, but their regularity during early incubation, if nothing occurs to disturb them, is remarkable. Towards the close of incubation, and after hatching, they are much less regular. And at any time, the presence of other birds or of alarming objects is likely to throw them out of the regular order. Thus on the 2d of July which I have been describing, the female left the nest at 8.15 A. M., being alarmed by the barking of a dog, and the male entered the empty nest at 8.19, which was probably a few minutes earlier than he would otherwise have done. The most potent disturbing factor, however, is the presence of other birds, which arouses the jealousy of the male. Changing places on the nest. — ^^Vhen the male comes, at his due time, to relieve the sitting female, or when the female comes similarly to relieve the sitting male, there is always a little communication or ceremony. There is little difference in behavior between the male and the female on this occasion. There is much variation in the ceremony, but the usual procedure is about as follows. The bird that is out, comes to the nest, giving the kah as it arrives ; it jumps on the edge of the nest-box, kahs again, flips its wings and tickles the head of its sitting mate. The sitting bird responds by fluttering its wings and showing evident satisfaction with its mate's attention. This exchange of greeting is usually sufficient; after a few caresses, and sometimes cooings, on the part of each bird, the sitting bird gently rises and steps forward, and the other steps in behind and settles upon the eggs. It sometimes happens that the sitting bird leaves before the other comes, as in the case mentioned above when the sound of a dog's bark caused the female to leave a little before Craig, Expressions of Einotion lu Pigeons. 73 her time. On the other hand, it is not uncommon for the sitting bird to be unwilling to leave, and for the bird that is due on the nest to paw the sitting bird's back, probe with its bill all around the sitting bird, feeling for the eggs, and finally enter the nest and squeeze the former occupant until at last, slowly and deliberately, it leaves the nest. This happens especially when the eggs have just hatched, for the feeling of the young birds under the breast apparently is a greater attraction than the feeling of mere eggs ; and so it often happens that both birds sit at once on the young, crowding each other, and each trying to cover as much as possible of the coveted nestlings. That the touch of the eggs or young and of the nest itself give pleasure and satisfaction to the sitting bird, is evident from many highly expressive acts ; such as the manner in which the bird arranges the eggs with its bill, touching them again and again, arranging and re-arranging many times ; from the complacency with w^hich it finally settles down upon them ; and from the absorbed interest it shows in arranging the straws and gently picking at anything about the nest (cf. p. 75.) When the male has taken the nest, all is quiet. The sitting bird always makes itself as inconspicuous as possible. Though this useful instinct has lapsed to some extent in the long-domesticated ring-dove, yet even the male of this house-bird rarely coos when on the nest. On this day on which the male cooed 487 times before taking the nest, he did not coo after that for three hours, and then he gave only one series of six coos. The female, after leaving the nest, goes first to breakfast at the seed-cup, after which she flies about the cage, preens her feathers, and busies herself with such small matters. Most of her activities have little interest for the present discussion, but it is worthy of note that she often alights on the edge of the nestbox, and on doing so she often sounds the kah. The female, it would seem, is always somewhat more attached to the nest than is the male. Although the female often uses the kah, she goes but little at any time ; during the four hours I watched this female after her leaving the nest, she sang only once, a series of four coos. In the middle of the day, no matter at what stage of the brood 74 'Journal of Comparative Neurology aud Psychology. cycle, the birds always sleep, or rest. The sitting male thus sleeps through the hottest hours. But the first and last hours of his brooding are spent in alert, though quiet, wakefulness. When the male is again free from the nest in the evening, he indulges in another period of cooing, though a much less noisy period than' that of the morning. When first relieved by the female, having had a long fast, he goes at once to feed ; then he usually performs an elaborate toilet ; and only gradually does he rise to the evening musical performance. This performance, indeed, as has been men; tioned, is much less than that of the moniing: for example, to compare with that morning's performance of July 2d which has already been described, I made a similarly complete observation on an afternoon just four days previous, Avith the result that, during the whole time between leaving the nest and going to roost, i. e. 31/^ hours (as against 41/2 hours for the morning), the male sang only 16 times (as against 95), making 65 coos (as against 487) ; moreover, while 70 of the morning's repetitions were accompanied by bowing and 24 were of the suppliant type known as the nest-call, the evening performance was entirely of that less emotional and less elaborate type known as the perch-coo. The cooing generally reaches a maximum just before the bird goes to roost. After taking the roost, the male usually coos a number of times, but his songs become rapidly less frequent till all is silent. And this silence ensues while daylight is still much brighter than that by which the bird first begins to coo in the morning; which again makes the songster's evening performance inferior to that of the morning. C". The Beood Cycle, Continued. After the laying of the eggs, pigeons in general spend their days in comparative quiet. This is not always very evident in the common ring-dove, as may be gathered from the foregoing pages ; but in some of the wild species the change is sudden and almost complete. The birds go about with a haunted look, with a perpetual expression of alarm, as it were. The male sings only at sunrise and at sunset, and when singing he goes away from the nest as far as possible. Craig, Expressions of Emotion in Pigeons. 75 Even when it is necessary to sound the ahirm-note while on the nest, the bird subdues its voice into something very like a whisper. The quietness of the brooding- time is thus a forced quiet, an active silence, caused by inhibition. In fact, in the tame ring-dove, which has so far lost its fear as to be at ease even during brooding, the inhibition is largely removed, and the birds are far more noisy during incubation than are any of the wild species. Strong attachment of the mates to one another is shown throughout the brood cycle by tame and wild species alike. The notion that the comparative quietness of the birds during brooding is due to lack of conjugal feeling, is a mistake. Quietness and retirement fonn only one phase of a great alteration of disposition during brooding; another phase is a sudden defensive bravery and irascibility. The sitting bird, whether male or female, defends the nest as valiantly as a brooding hen. And even when off duty from incubation each bird is now moTe bold than ever in attacking and driving away enemies. The eggs hatch in 14 days ; that is to say, on the 14th day after the laying of the second egg. The hatching of the eggs, the arrival of the young, gives again a stimulus of the same sort with the first appearance of tlie eggs, and makes the parents again still more quiet, more jealous, and more devoted to their parental duties. It has already been shown (page 73) that the movement of the young under the breast of the parent is a stimulus to the latter. Professor Whitman has found that when he needed a foster-parent for some valuable young pigeon, he could take a ring-dove whose eggs were not yet ready to hatch, and, by stroking her breast gently with his fingers in imitation of the movements of the young, he could induce her to connnence feeding. Thus we see that the feeling of the movements of the young is a stimulus not only to the feeding impulse but at the same time to the secretion of "pigeon's milk" in the crop. That the young are a greater attraction than are the eggs to the sitting bird, is evidenced by. the frequency with which the parents sit both together on the little birds, often crowding each other to get a larger share of the coveted nestlings. That the hatching-out of the young gives an additional stimulus to maternal jealousy, is shown by the 76 yoiirnal of Comparative Neurology ajul Psychology. fact that if, on the day of hatching, there are fledglings still in the cage from a former lirood, the motlier now ceases to tolerate the presence of those fledglings ; her eye begins to glare and her feathers to bristle, and soon she attacks her fledged offspring with snch fnry that the owner is obliged, for humanity's sake, to take them out of her cage. Within a few days after the hatching of the eggs, the birds begin to become irregular in their brooding hours. The young are still kept covered, to be sure, but the occasional desire of the parents to sit both at the same time, and the frequent necessity of their coming to feed their young, gradually breaks up the regularity of the sitting exchanges. Brooding ceases entirely, at least if the birds are kept indoors, in a period of 10 to 12 days. But while brooding has thus been gradually given over, the business of feeding has become rapidly more and more arduous, as a result of the rapid increase in size of the young and the enoTmous development of tlieir begging powers. There may even be added a new note to the parents' vocabulary at this time, a call to the young to feed. But as soon as the young have reached their maximum importunity they begin to pick some food from the ground, and the parents, tired and sore-mouthed from the feeding of youngsters almost as large as themselves, are ready to quit feeding; thus the young are gradually weaned, at an age ranging from about 15 to 2.5 days. The mother quits feeding before the father, foT she is always more devoted to the next pair of eggs and young, while the father feeds the fledglings in the day-time and roosts beside them at night. A succeeding pair of eggs and young has already been mentioned. Preparation for such begins very early, in that the parents, while feeding young, commence another round of cooing and love-making and mating; the cycle of one brood is not finished before the cycle of the next is begim. Just how early the new cycle Avill be begun, depends upon the season and upon all the circumstances. As to season : In the spring and early summer the succession of broods is more rapid than at any other time of year. As to other circumstances : For example, the destruction of eggs or young at any stage sets the parents at once to cooing and love-making. Professor Craig, Expressions of Emotion in Pigeons. 77 i AYliitinan, oai removing the eggs from a nest, has observed the birds to begin fondling one another within half an hour afterwards. When only one egg hatches, so that the labor of feeding is only half what it usually is, the birds have more energy and come more quickly to the preparation for a new brood. The shortest interval I have observed between hatching and laying is, when only one bird is reared, 13 days ; when two birds are reared, 14 days. In the normal course of events the inauguration of a new brood cycle is gradual, being a repetition, perhaps somewhat abbreviated, of the performance by which the birds first become mated. There is first a period of bowing and cooing by the male and a gradual rise of excitement in both birds ; then a period of copulation, nest-calling, and nest-building, with a gradual decline in the excitement, followed by the laying of the eggs and the birds' devotion again to incubation. Thus (even before the old brood cycle is finished) is a new brood cycle begun. It has been said that after the birds have begun a new round of mating they still foster the young of the last brood, but there is a limit to tliis fostering of the old fledglings ; there comes a time when the parents not only refuse to feed them but cease to tolerate their presence. This desertion of the former brood happens much earlier with the female parent ; so soon as the mother has taken to sitting again, she begins to acquire a hostile attitude towards her nearly grown-up-children ; so far from feeding them, she pecks at them when they try to share the seed-cup with her; and so far from brooding them, she keeps them alwaj^s at a little distance from her body. Affairs generally continue in this smoldering condition for several days; but a day comes — and according to my observation it is almost invariably the day on which the new eggs hatch — when the fire of this maternal jealousy bursts fourth and the mother persecutes the fledglings with such fury that if they were not taken from the cage they would perhaps even be killed. The male, though not nearly so aggressive in this matter, has become more or less completely indifferent to the old fledglings, and shows no regret at their departure. Thus ends the brood cycle. y8 'Journal of Comparative Neurology and Psychology. B". The Seasojn'al Cycle^ Contixued. Renewed efforts, as shown in cooing, kahing, nest-building, and a host of other activities, are necessary to initiate each brood cycle. If the birds be disinclined to effort, from any cause, such disinclination will delay or prevent the commencement of another brood cycle. This is what happens in the molting season, beginning in the latter part of August. There may be a decrease of breeding power, especially in some pigeons, even before the molt; but most of the domestic ring-doves retain ample breeding powers up to the time when the molt begins to diminish their general vitality. The breeding powers lost at this time are not regained, by the wald species, until the following spring ; and though the domestic ring-dove may be bred all through the autumn and winter, yet the frequency of repetitions of the brood cycle is lessened, the health of the birds may suffer, and it is evident that this extension of the breeding season is unnatural. Coincident with the lapse of lu'eeding propensity, in all species of pigeons, is a loss of voice, a loss especially of the more emotional, more musical notes. The loss of voice is not so conspicuous in the domesticated and unnatural ring-dove. The loss of song is not complete even in most of the wild species, for their coos may be heard at irregular inter\^als through the months of September and October at least; but the coo at this time is in some cases notably different from that of the ]ireedino--season. And thouo-h the sone's may be given thus sporadically, their sum total is exceedingly small. The comparative silence which reigns in the pigeonry is gloomy; the hushing of the birds in August is an annual surprise, a change so sudden and so great that one does not l)ecome accustomed to it. A". The Life Cycle^ Continued. Though ring-doves begin to breed at a vcn- eai"ly age, even at four months, and thereafter continue to pass through the regular succession of brood cycles and annual cycles. Professor Whitman has found that they do not reach their maximum breeding powers until the age of about three years. After that age, the breeding powers remain at the maximum for some years. Craig, Expressions of Emotion in Pigeons. 79 Professor AVbitinaii has kept blond ring-doves till they were about ten years old. In one such case be knew pretty definitely that the bird's death was due to causes other than old age. Yet he thinks that he has observed somewhat of a decline in t-be breeding powers in a bird about ten years old. Summary of the Life-History. A'. Beginning of the life cycle. The voice of the young ring-dove is heard the first day, and is useful to induce the parents to commence feeding (page 55). The voice of the growing young is useful to cause the parents to give a sufficient amount of food, and to continue feeding until the young one is able to feed itself (page 57 ). Fear begins to be shown as soon as the young have the full use of their eyes (page 55). Alarm develops somewhat later, 12th to 14th day (page 59). The kah appears often on the 27th day (page 60). The charge appears on the 27th day or later (page 65). The coo appears from the 27th to the 47th day (page 62). It is at first very imperfect, and develops very slowly to the adult form. Development of the voice is of two sorts which may be referred to two causes ; namely, development of the syrinx or vocal apparatus, and development of the nervous system (page 01). The young often give cries at the suggestion of the parents, but they do not imitate the cries of the parents (page 64). B'. Beginning of the annual cycle. The elaborate cooing and other performances of the spring season serve to proclaim the sex of each bird (there being no markings distinctive of sex), to bring the birds together in pairs, and to unite each pair by a firm bond (pages 66-69). C Beginning of the brood cycle. The male and the female, by mutual stimulation and self-stimulation, work up a pitch of excitement sufficient to start them on the arduous, month-long labors of the brood cycle (page 69). 8o 'Journal of Comparative Neurology and Psychology. After days of copulation and nest-building, all of which are controlled by cooing and other ceremonies, two eggs are laid and the birds enter upon fourteen days of brooding. D. Tlie daily cycle. The male and the female take turns very regularly in sittiug on the eggs. Each time when one bird relieves the other, there is a ceremonial communication between them. C". TJie hroocling cycle, continued. After the eggs are laid, the birds are guardedly quiet when near the nest, but there is no diminution of conjugal affection (page 74). The hatching of the eggs, and the movements of the young under the breast, are strong stimuli to the parents (page 73 ). The parents may even add at this time a new call to their vocabularly, a call to the young to feed (pages 43, 44, 76). The parents, while still feeding the young, gradually work up, by cooing and other performances, to that pitch of excitement which is needed to start a new brood cycle (page 76). B". The annucd cycle, continued. At the end of the summer, especially when the molt begins, the birds have not sufficient energy to work up to the beginning of a ne^v brood cycle. Thus brooding stops. A". The life cycle, continued. Pigeons do not reach their maximum breeding powers until an age of about three years. The lengih of life is not definitely known. EXPLANATION OF PLATE. Fig. 1. — The alarm (page 35). Both are in the attitude of alarm; but tlie more extreme in attitude is the adult bird, distinguished by the black half-ring on the uecli. Fig. 2. — The charge (page 42). Male. Fig. 3. The perch coo (page 47). The bird that is cooing is an adult male. The other bird is a young one. Fig. 4. — The nest-call (page 51). The bird with its head down in the nest is a male, nest-calling to the other bird, which is a female. Figs. 5 and 6. — The bowing-coo (l^age 48). Fig. 5 and Fig. 6 both show the same bird (an adult male) in a phase of the bowing-coo. Fig. 5 showing the up phase, and Fig. G the down phase. The bird is bowing-and-cooing to his own image in a large mirror which was placed close against the end of the cage for the purpose of getting these photographs. i:\I'l!KSSI<)NS OF KMOTION IN PIGEONS. WALLACE CRAU;. 1 UK Jol KSAL OF COMPARATIVE NELKOLOGY AND I'S YCU()I,(>(! V. VOL. XIX, No. 1. THE EEACTION^ TO TACTILE STIMULI AND THE DEVELOPMENT OF THE SWIMMING MOVEMENT IN EMBRYOS OF DIEMYCTYLUS TOROSUS, ESCHSCHOLTZ. BY G. E. COGHILL. Studies from the Xeurolosjivul Laboratory of Denison University, No. XXII. With Six Figures. In 190G I began a series of experiments upon embryos of Rana and Amblystoma with a view to determining whether there is any regularity in the earliest neuro-muscular responses to tactile stimuli in the amphibian embryo. During the season of 1907 these experiments were continued upon embryos of Diemyctylus torosus, Eschscholtz (Triton torosus). Although the work of the first year gave interesting results and convinced me that the field of investigation was a fruitful one, it was less exhaustive and critical in its methods than the later work has been, and there is no occasion to give an account of it in this connection. It will, therefore, receive no further treatment here and all the data and discussions of this paper will relate exclusively to Diemyctylus torosus. These experiments were originally planned for correlated anatomical and physiological studies. As an introduction to such work upon Amphibia they form the basis for the anatomical part, since they reveal distinct phases in the development of neuro-muscular response to the most primitive system of cutaneous receptors. But, apart from this significance to pure anatomy and physiology, they are, of themselves, an interesting contribution to the science of animal behavior, for they deal with a most important phase of behavior, namely, its very beginning in the embryo. If, for instance, there is any such thing as a ^'simple reflex," such as Sherring The Journal of Comparative Neurology and Psychology. — Vol. XIX, Xo. 1. 84 'Journal of Coinpnrnfive Neurology aud Psychology. ton suggests/ it must be found in the earliest reflexes of the embryo as observed in these experiments, and if it is possible to trace the development of a "simple reflex" into a form of acknowledged instinctive behavior, this link in the development of behavior would seem to appear in the development of the swimming movement as described in the following pages. In view of this bearing of the experiments upon the subject of animal behavior certain results of the experimental part of my investigations are here made known before the anatomical phase of the work has been completed. Methods. The embryos were removed from the egg membranes at various stages in development, ordinarily before they showed any sign of irritability to tactile stimuli. They were then placed in shallow Petri dishes, a single specimen in a dish, and tested from time to time for reactions. Usually an experiment continued until the animal began to swim. The stimulus employed was a touch with the end of a rather fine human hair, mounted in such a way as to render the touch very gentle. The extreme sensitiveness of some very young eml)ryos is remarkable. Even the touch of a fine piece of lint will at tiuies evoke a vigorous response, as if it were a violent irritant. Without critical consideration the tactile nature of this mode of stimulation might be held in doubt. The touch of a hair such as was used in these investigations might easily cause a considerable pressure, so that there might be a question whether the responses were to a strictly tactile stimulus or to a mechanical stimulus upon the muscles or central nervous system. Indeed, in the very early phase of development, when the irritability was for some reason unusually low, some of the reactions, I believe, may have been to direct pressure upon the muscles or central nervous system. But such instances, if they occurred at all, in these investigations, were, 'Sherrington, Charles S. "The Integrative Action of tlie Nervous System," p. 8. CoGHil.i., The Reaction to Tactile St/i/iiil/. 85 I believe, relatively rare. K(ir iiislaiiee, wlieii the stimulus is applied to the under side of the head as the animal lies on its side, and the response is a moxcment of the head awaj frnni the side touched, it is inconceivable that this response is to a direct pressure upon the muscles effectino- jho moveuKint, and it seems altoo-ether ipiprobable that such a stimnlus could be brought to bear upon the central nervous system directly in such a mauncu- as to give rise to a constant form of response. Or, in case the stimulus is applied to the margin of the dorsal or ventral caudal fm and a movement of the head only results, as regularly occurs in certain phases of development, it is absolutely impossible for such a reaction to be given in response to pressure either upon the acting muscles or upon the central nervous system. As reactions of this sort occur here and there throughout nearly every one of my experiments, it seems to me certain that the stimulus employed was, with possibly rare exceptions, purely tactile, and that, so far as the mode of stimulation is concerned, my conclusions are valid. Ordinarily the stimulus was applied to the upper side of the specimen as it lay on its side on the bottom of the dish„ Frequently, however, it was applied to the under side of the specimen from beneath, in order to determine whether contact with the dish had any influence on the mode of reaction, but it was impossible to detect any factor of this kind in the responses. Some embryos, also, were suspended in an upright position and tested for the same purpose, and with the same result. An individual record in detail was kept of each eml)ryo froui the time it was removed from the egg membranes till the end of the experiment. In the record of each trial, or a])plication of the stimulus, the following factors were noted partieuhn-ly : the region and side touched, the form of the response and the time of the trial. Tabulated schemes for rapid recording were tried in my first experiments of 1906, but it soon became apparent that such forms could not be adhered to, for they were necessarily based upon presumptions of some sort and were, therefore, a hindrance rather than a help to alert observation. These methods were wholly abandoned and have no part in the records from which this paper is written. 86 'Journal of Comparative Neurology and Psychology. Reaction to Tactile Stimuli. A. Response to Stimulalion on the Head. According to their reaction to a touch on the side of the head, in the region innervated by n. trigeminus or n. vagus, embryos of Diemyctylus torosus may be grouped according to three types, as follows : Type I. Embryos which from the beginning and during a considerable period, respond regularly or almost regularly with a movement of the head directed away from the side touched. Type II. Embryos which for a relatively short period at first respond irngularly with movements of the head toward or away from the side touched, and then enter upon a relatively long period of response like that of Type I. Type III. Embryos which are at first asymmetrical in response, that is to say, they move their head in one direction only, regardless of the side touched, and then enter ujxm a short period of irregularity like the first period of Type II, and finally upon a relatively long period of response like that of Type I. Or individuals of this type may pass directly from the period of asymmetry to the regular form of Type T. The accompanying charts illustrate the behavior of typical specimens from each of these three types. The first column on the left in these charts records the serial number of the trials made, and the record of each trial is represented in the corresponding horizontal line to the right. The figures in the second column from the left record the time in hours and minutes that elapsed since the last ])reccding trial in each case. The diagrams in the third column from the left represent the form of reaction in the various trials. ^^Tlere there is more than one diagram in a space these are to be read from left to right, and each represents a distinct phase in a series of movements. The arrow occasionally placed in these spaces indicates that a cephalo^caudal progression of the movement w\as distinctly observed. Where an ^'S" occurs the specimen swam, and the following diagram in the same space indicates the composition of the swimming movement. It should be noted that these diagrams of the movements are simply free-hand representa CoGHli.L, The Reaction to Tactile Stimuli. 87 tions of the reaction according to written descriptions made at the time of trial. They can not be considered as absolutely accurate in every detail, but they do represent truthfully the general order of the development of trunk movements in these animals. The curves of the charts represent the side touched and the direction of the initial movement in the reaction relative to the side touched. The solid line records the direction of the movement of the head; divergence to the left from the vertical records a movement toward the side touched ; divergence towards the right, away from the side touched ; conincidence with the vertical, undetermined. The broken line records the side touched; divergence to the left signifies a touch on the left side of the head ; divergence to the right, a touch on the right side ; a blank, no record. Obviously, where the two cur\'es are parallel the movement recorded was to the left; where they diverge or converge the recorded movement was towards the right. The apparent incompleteness in the serial numbers of the trials in the first column of some charts is due to the fact that in these experiments alternate or occasional trials were being made with reference to touch on the tail bud. The charts represent perfect series of trials with reference to touch on the side of the head. The charts presented here are selected from a series which, with descriptioins, has been deposited with the Wistar Institute of Anatomy and Biology, for the advantage of students who may be interested in a more exhaustive report of my experiments than this paper affords. The accompanying table presents schematically some of the data upon which this classification into three types is based. It is the tabulation of the records of 36 specimens which have been selected solely uj)on the basis of completeness of the record and duration of the experiment. Owing to the difficulties in the manipulation of the work and unavoidable hindrances many experiments were not carried continuously through the entire period which is here under consideration, and, although contributing materially to the evidence on the problem as a whole, can not, on that account, be included in a comparative study of this kind. 'Journal of Coynparative Neurology and Psychology. ■■ "■ ~ ~ ~ ~ ~" ~" ~ "" ~ ~ ~ ] p T ~ r T r T r p p p p p n _ _ _ ~ ~" ~ _ _ / s / / _ ~ _ _ _ _ _ _ _ _ I I z I I I ~ " ~ ~' ~ ~ ~ ~ ~ " 7 / ~ _ ~ _ _ _ _ _ _ _ _ / ._ ~ ~" ~ ~ _ _ . , / _ _ _ _ / ~ / / ■"" ~ ~ ~ ~ ~ ~ ' ~ ^ " 7 7 ^ ■■ ~ _ / " ~ ~ _ _ ~ ~ ~ _^ _ _ _ 7 _ ~ / " ~" ~ ~ ~ / ~" ~ ~ ~ _ _ _ _ _ _ _ _ _ _ _ / ~ ~ ■~ ~ ~ ~" ~ ~" ~ .. / _ _ _ _ 7 ,_ ~ -1 u -J _ J _ ^ _ _ _ _ _ _^ _ J l_ _ _ _ ~ ~ ~ ~ ~ ~ r -^ —J — 7 L _ ^ U _ r _ _ 1 — 1 , ^ y _ _ _ _ _ _ ._ . L _ _ _ _ z\ _ / _ ^ _ / 1 h _ _ 1 , / _ _ _ _ r 1 n / / { 1 / ■— ~~ \ / i ■■ ~ ~ ~ ~ -^ / ~ / "T .._ ^ "■ ~" — ' / ~ ~ -h I— _ / ^ ~ ~ ~ ~ _ -^ { ~ _ _ _ _ / _ . / / h _ / ~ _ _ / 7 1 / ■■ ~ '~ l~ ■" ~ — r~ / h ~ ~ ~ "~ ~ ~ ~ _ / / ~ -r ~ _ ^ ^ -\ / ■~ ~ ~ ~ " ~ " "~r / 1 _ Z ~ ~ ■~ ~ ~~ ^ ^ n~ / / " ' ~ 1 >— — r— / ~ ~ ~ ~ r r ^-p _ Z __ _ _ _ _ " ~ ~ ~ ] / / " 1 / ^ ~ ~ ~ — j ! \ 1 '■ ~ ~ N , N '• \^ \ / , ' ^■ / V y ^ , , \ ' , 1 -. _ _ _ ^ ' \ \ ^ L Ly V V « \^ _> t -> ^ fl 6 L 8 Vj 3 4 8 u s Ky -, Z 1-1 « \j lo 5 s 5 9 ^ ? ^ \y 8 9 s v^ ^ \J c) d ^ k^ ^ r^ h cri r^ z ■^ ^ -> ^ '^ 5 ? I ^ S ' ? '~ ? z. '• i 2 ^ ^ ^ "; S ^ " ^ =; " ^ -■ ' "■■ i, ^ ^ -^ <r\ ^ ■" ^ ^ ^ ^ n ^ CJ ■^ '^ " ■^ " 2 1 1 ^ 3 ■^ ^ ^ ^ 1 -i 5 S s ^ s s; % \ •* 3 f " i s J " ? ?,1 3 3 •^ 2 2 S '^ Rl ^ ^ !Q ^ ^ 00 S 2 i ^ ?i. t 5 ro S; s 5 O o r z s 1 O aj 2 "^ " g CO o 2 '-' r- S ' ^^s 7 ^-ZSZ Z i -,2 -4- 2 ^2 it _,? it i 2 t 7 Z Z ^ y 1 / / ^ 7 / / / 7 "T 7 7 T^" -.^"-^ /- ^r' "„ := I /" P' \ :^ ^ IL it ^ N/ 1^ ^ ^ ^ \/^ " ^^ "b 'it "^'-It Z ' t z 7 X Z ^ -1 7 7 Z^^SZ 7 IT ^^ ^^ 7 2S^ z :"^„ ^/ If: \7 .' . ^ ?S-;?.SSss?;s«|S?!,^STE ^«^^sg5a:>)5T-%-5.kss5;sp?3^^|?2l|^5? ■^c\ji^';i'ri^0No>ff'5^*^<!5*^'?-^*^'Mryfy:MMfMM'v"\Jf\J'*3'*^f^ '•^"^p '^j'^ '^'^'^'^•^'^■^t'^ '^'^ ':t "' "^^ "~ ~ _ "" ~ ~ ~ ~ ~ "" ~ ~ ~ ~ ~ ■" " ~ ~ " 7 N / S / 7 / / ■^ / ~ ~ 7 < 1 7 1 / -^ _ J _ I_ _ ^ _ _ _ _ _ _ _ _ _ _ h _ _, _ _ J L Y iS, / -, -^ H P |— ^ ^ ^ \y c ^ -^ '— 7 ■^ -^ V ^ ^ ~ ^ ~ \ /' , / J / / ^ \ / 7 7 \ , ' \ ^' / -^ N / / , \ , ' 7 / / /" / /* / ■. / / / >* / 7 \ / ^ 7 \ / \ 7 \ / / ^ / V / r < / ^ / ■v / < / \ 7 S / ^ n ^ f\ (\ ti C p 8 8 16 KJ % «J j » ^ a » J r\ i 8 5 >J « ft ^ 8 8 ? <2 8 8 w, a ^ » r n "' O 5 SI ^. s ?. en 5 ":; rn ?■ ":: S 5 5 2 "i fe$

in

?

S

f^

^

't

^

ij

"

k

<^

•^

"

?■

^

^

^

"

■t

^

iu

cy

t

^t

In

\o

>

1

£

^

i

^

lS

^

5

?

\

^

5

5

Si

^

^

^

S

tb

M

§

n

s

3

s

•r,

£

^

00

m

t

i

^

T

jt

^

^ _2

^

J

■c <<-.

i

^

I

V

■o

^

T

S

L

_

.

3 G

Sol

►^ ~ ii

•^ -a j2

CO Ci r—

■i:^

>•.

0)

0)

.^

^

a

<r>

0)

g

3

s

o

a

p>>

><

r/J

H

cj

l>5

d) t*

f^^

&

00 "Journal of Comparative Neurology and Psychology.

1

z::zz-z—z : igr:

n

n

^

-1

-1

q

n

-pl

«,?

H.

J

3/?

1 Z^

4

Li^

S- /^ ^

~

9

5 / s

6

/o

. ^^

\ ,

so

5= /

s

8

9

tT) V

/

~

3

54

■7 />

\

{<>

/*

? y

~

^

/^

\

"

13

^9

5 /

\

'—

/4

3

' , y

\

~

16

4.-J

? X

\

~

!■'

22

J y

\

'

/S

/'

if'

s

-

/9

/.'

^^^

/

20

■.4.t>

S ^i

^

~

21

^

~

^2

a

? ^^

^

//

1—

24

y

N. . . ..

/

~

25

9

. ^v

/

26

■*:'6

N .

/

l~

2?

^■/t?

" / ....

^^

~

ZS

^.

,

"~

29

/■*

^.

Z

^~

30

L" ^v '

~~

31

■*/

? N, .^

32

it>

S .N'.

~

■n

/■s

.^v,^

~

34

7

■» / s^

_

35

5U

J \ . s

3?

38

<^= N

S

^y

/i

r , - ^

~

39

/3

^

"

40

II

i^ "■

^

'~

41

13.

Cc

A, S

'~

42

2

c

-i Si

^

4?

k

<:<=

^

'~

44

1

?«

\ c

r~

45

•d

5=

h " 5^

■"

1

J

^, s^

4

z

-^ ., ..

It :i S

411

1

>' ^ ■

N S

[

4^

10

f-""

^ s

£<?

1

•=>r

i 5

'~

pc

\ S

'~

52

1

In S

fj

1

•a

^ ^

54

Z

Co

s

a

2

?^

'

^

s

57

2

^

V

s

,.,

~~

5li

?=.

\

\

^;

s

'~

60

SO

c

V

6/

c?

\

^

62

34

p/

\

\

63

/o

Me

s

2.

c

\

[~

63

1

e

^

i~

2

e

^

s

67

/

c

s

68

2

\

s

69

/

c

^

\

70

/

fts

<

s

-pi

7

cat

s

7^

=> ■

s

F3

2-.Z>

I

/

74

SI

/J

75

£

y.

\

7(>

2

C5

\

7?

^

JCl

78

51

,

P,9

T5

,

/

80

1

•15C

^

8/

l5Ci

/

I

C

?3

1

8-T

^-l

fc

/

86

2Z

e

/

\

S?

n^

/

K

ff^

f^

u

}

f^°\

'^3S

f

V

90

» TJ

^i

^

/

31

4SZ

( i

'■

N.

^?

1

?

9-)

2

?i

J

94

/

/

V

95

^ 1

^

J

96

lt\

JT

/

\^

9ff

62 C

/

s

99

1

<?

/

S

«~'

s

X :

_

_

_

L_

Fig. 4.— Experiment 37, illustrating Type III, in a case wliere asymmetry in response passes abruptly into regularity, and the asymmetry is preceded by two movements at variance witli it.

"■

■■ ■

"

~

■■

"■

~

~

~

""

"

■"

-.

~

"

~

"

"

"■

"

"

■"

■"

w

^

"

^

"

"1

■■

■■"

~

~

/

^,

7

'

/

•'^

/

~

\

l/

'

\

^

,

.■

■^

■7

^.

^

-3 —

J

-'

-;—

^^

■"

"7"~l

/

"

N

,

/

~

,'

,

/

~

_

/

\

\

/

r

^

s

/

/

S

/

/

\

/

\

/

\

/

\

/

/

V

/

■~

/

s

^

/

N

/

'

/

\

/

s

'

/

s

/

i

-:^

-^

/

\

/

\

/

\

/

^

_

/

/

S

/

s

/

"

S

/

\

\

/

\

/

s

-^<

/

s

/

\

/

h

^>

/L.

_

_

_

/

s

/

s

/

— _

_

_

_

_

^

/

s

/

s

/

- —

'

-r^>

V

r^

c

t

KJ

1/

\,

C

r

»"

^

r

n

r

r\

r

r\

r\

\!

ri

«

Lk

yj

t\

1)

r

n

i)

hS

r\

rt

5T

%

«

"

M,

V

8

h

r.

1-,

U

^

~"

~

CM.

1*

tM

"

^

^

^

^

"

"

~

"

1

"^

^

i

~

■0 <»

^

"

"

Ir,

^

'^

^

"

'*

^

■^

"t

-^

«

'^

i

V.

s

>

1 *;'

«

5S

^

j

Si

^

"^

t^

'.\

XA

A

tl^

i^

«

S

\

a

d

'^

t^

^s;

Si

•T

U

■n

S

^

f,

,7

Hi

f^

i

"N

p

^

J

P

■^'

^

m

^

S

^

'

■^

~

"^

N

/'

S

r*

k

/

/

■^

7

/

s

/

i~

/

/

1

/

/■

^

/

"

/

\

/

"

1

/

s

/

^

\

/

/

/

/

<^

^

/

\

}

\

/

,,

N

/

\

/

^

^'••

\

/

/

s

/

S

^

/■

s

/

s

^

\

/

S

•^S

/

s

/

S

/

s

\

^

S

\

/

S

/

s

^

/

\

.

\

^

\

,■'

\

\

\

\

/

^,

\

^

/

^

■,

^

',

,

"^

\

/

-r

^r

^

- V

?

,

K.

^

>

L

<

\

\

5

a

r>

o

.

(_

8

^

I.

t

5

e

c

ft

\j

V-

I

\.'

,

O

c

4

^

\^

yj

%

B

L^

A

V.

8

S

B

v>

5

\

\j

V,

V.

r^

V-/

»

r\

r

K/

»

y

^

^ '^

"C

^

^

^

"•

'«1

'

■^

S

^

t!

'^

"

•^

<

1

^

1

^

■*

■^

^

^

'^

'5

^

'^

^

■4

•^

^

w

^

^

-^

^

^t

^

?

•o

^

'^

"

^

^

s

2

"

k

"

±_

M rv

^

" ^

■?

"j

"

"

E

?

-^

n

5;

5

n

2

3

3

^

1

%

^

^^

?

s

"i

^

1

\

^

^

%

?■

^

^

"?.

li

^

^

f^

s

^

§

S

«

^

S^

=^

3>

^

^

^

^

S

S

Oi

Fig. 5. (Upper figure). — Experiment 48, illustrating Type III. in a case where tlie period of irregularity is influenced, apparently, by the preceding asymmetry. It should be observed that the reactions were taken rapidly during the period of asymmetry and irregularity.

Fig. 6. (Lower figui'e). — Experiment 162. The embryo fx'om which this record was made was, on the whole, the most irregular specimen of my series. Still, after the period of asymmetry there is a marked general tendency to move the head away from the side touched.

For further data see Table, p. 02.

92 'Journal of Cojiiparative Neurology and Psychology.

A 45

B

7 ?

24:00 24:00 24:00 20:22 13:30

C

D

E ' F

G

64:36 52:07 72:00 96:14 73:34 67:52 96:43

H

31 53 40 49 40 35 57

I

96.7

94.3' 100.0

95.9

95.

97.1' 100.

32

a

146

><

150

145

144

156

1

Average. . .

21:10

i

74:43

305

97.4

151

20 20:00

5:45

5:18

?

3:53

20:00

1:30

24:00

? .

41^ 1:30

1:30

14:00

7

12:19 6:42 2:27 2:04 1:35 1:52 6:29 4:36

11:49

26 5:10

6:18

8:34

33:21

19

70:49 48:09 71:09 64:41 81:37 49:02 47:58 59:04 47:51 59:15 64:57 47:17 93:31 32:39 37:34

34 25 35 24 37 37 27 23 23 25 38 27 41 31 19

100. 96. 91.43

100. 97. 2<

100.

100.

100. 95. 6«  92. 94.76 92.5 87.8 93.5 94.7

143

11 9

1 17

14

19 00 14 17 29 12 27 39 23 42 24 10 11 01

147

148

149

34

I-H

142

H

155

Ph

158

>*

161

H

163 154

140

141

36

26:47

Average . . .

8:12

8:41

3:47

58:22

446

95.

136

7

5:20

2:54

5:25

24:00

7

■14

12:00

•40

1:28

9:00

20:00

12:00

24:00

14:51

7:12

8:26

5:26

32:26

38 7:31

12:35

23:48

10:04

16:38

5:38

12:18

4:13

22

27

64:51 58:04 69:13 66:10 32:36 50:22 72:53 45:08 34:49 68:12 62:25 98:48 58:36 115:58

33 22 24 32 53 26 30 46 20 39 33 60 37 30

96.97

100. 91.6 96.8 98.1

100. 96.6 78.2 95. 8 97.4 93.3 98.3 89.1 80. 9

152

10

15

15

2

22 22

14

58 22 43 32 52 39 27 29 45 44 09 41 16

159

153

37

5

39

^

157

H

162 164 139 137 126 160 38

9:37

5:38

22 07 18:48

43 3:44

2:32 11:59 2:26 6:26 4:01

41 4:26

Average . . .

9:45

11:33

5:34

4:30

9:17

64:08

489

93.

iThe first 30 responses, distributed ttirougli 23 hours and 13 minutes, were all directed away from the side touched.

^During one period of 41 hours and 30 minutes there were 30 consecutive responses away from the side touched.

^During one period of 46 hours and 13 minutes there were 22 consecutive responses directed away from the side touched.

^During one period of 57 hours and 36 minutes there were 25 successive responses directed away from tlie side touched.

^During one period of 36 hours and 50 minutes there were 18 successive reactions directed away from the side touched.

"During one period of 41 hours and 32 minutes there were 19 successive responses directed away from the side touched.

'During one period of 47 hours and 6 minutes there were 25 successive reactions directed away from the side touched.

^During one period of 22 hours and 50 minutes there were 17 successive movements directed away from the side touched.

"There were in all 52 responses, distributed through a period of 137 hours and 33 minutes. Of these responses, 40 were directed away from the side touched, a percentage of 76.9.

CoGHiLL, TJie Reaction to Tactile Stimuli. 93

The several eolunins in this talnilation have significanee as follows :

Column A. The number of the experiment, the data of which read to the right.

Column B. The time that elapsed between the last trial which gave no response and the first to which response occurred.

Column C. The time during which the embryo was asymmetrical in response.

Column D. The interval or time that elapsed between the last observed response that accorded w^ith asymmetry and the first response that accorded with irregularity.

Column E. This is the second phase in the development of embryos of Type III, and the first phase of embryos of Type II. It is described above as the period of irregularity in response.

Column F. The interval or time that elapsed between the last observed reaction that accorded with irregularity and the first that accorded with the regular form of response as described above for Type I.

Column G. The time during which the embryo is considered as moving its head regularly away from the side touched.

Column H. The number of responses given during the period represented by Column G.

Column I. The percentage of the responses indicated in Column H that were away from the side touched.

The time is recorded in each instance in hours and minutes, excepting in a few instances in C^olumn B where the time was not determined. Averages are given in the several columns for each of the three types, excepting in Colunni IT w^here the corresponding numbers represent totals.

With reference to the side touched in each trial my records are complete, but, inasmuch as the records in Column G clearly have no references to the side touched as determining factor, this element of the question is omitted from the table.

A comparison of the averages in Column B of the table might be interpreted to mean that the specimens of the second and third type came under observation relatively earlier in the period of development than did the specimens of the first type. But it should be

94 'Journal of Comparative Neurology and Psychology.

noted that the figures in Column B represent the maximum possible time of irritability before the observation of it began. On the other hand, a comparison of the averages in Cohunn G shows a clear distinction between Type I, on the one hand, and Types II and III, on the other. There is a difference of, say, 10 to 15 hours in the length of the period of regularity in moving the head away from the side touched. Furthermore, if the average of Column G for Type I be compared with the corresponding average for Type II plus the averages of Column E and F of this type, it will be seen that the embryos of Type I were longer in passing throngh the one period of regularity than were the embryos of Type II in passing through the periods of both regularity and irregularity, including the interval. It would seem, therefore, that a period of irregularity has not been passed over unobserved in Type I, and that the distinction between these two types is not based on the relative age of the individuals when they came under observations.

A similar comparison of the corresponding figures for Type III with those of Type I shows that the time represented by Columns E, E and G for Type III approximately equal that of Column G for Type I, But for the excessively long period of ISTo. 38 in Column G, the comparison would result about the same as that with Type II. But when the period of asymmetry and the following interval is taken into account it is clear that the specimens of Type III were a much longer time in passing through the periods represented by Columns C, D, E, F and G than were the specimens of Type I in passing through the period of Column G alone. This would seem to indicate that the condition of asymmetry is due to a precocious development of one side of the neuro-muscular svstem rather than to a retarded development of the other side. At any rate the sum of the averages in Columns C and D for Type III is greater than the average in Column B for Type II. It would seem altogether improbable, therefore, that a period of asymmetiy like that of Type III has been passed over unobserved in Type II.

^Miile I do not place any great dependence upon this comparison of the averages in the table, I believe they do tend to show that the difference between the different types of reaction as observed in these

CoGHiLL, The Reaction to Tactile Stimuli. 95

and numerous other embryos is not based upon relative age but upon the relative development, and probably the variable physiological condition, of the various constituent elements of the neuro-muscular system. When a period of asymmetry occurs, it appears before the period of irregularity or regularity, and never follows either of the latter, excepting in rare cases when one or two movements right at the beginning of the experiment are at variance with the asymmetry (Figs. 3, 4, 5, 6). The asymmetry clearly influences the irregular reaction in some cases, so that the movements towards the side touched appear to bo determined by a partial persistence of asymmetry (Fig. 5). But this is not always the case. The period of regularity persists, ordinarily, till near the time of swimming. The actual length of the period varies greatly in different specimens, but a comparative study of numerous specimens convinces me that the regularity in response is purest for a period of about 48 hours.

The structural basis for a regular asymmetry in response must be in the ascendency of the effector system of one side over that of the other, rather than in structural difference in the receptor systems of the two sides. Two facts particularly support this interpretation: (1) All spontaneous movements (somatic) that have been observed in embryos which conform to a given asymmetry are in accordance with the asymmetry in each case, towards the right in dextrally asymmetrical specimens and towards the left in sinistrally asymmetrical specimens. (2) In any given asymmetrical embryo the asymmetry is the same with reference to stimulation on the tail bud as it is with reference to stimulation on the head, and specimens that are asymmetrical in one respect are so also in the other.

The stnictural basis for a regular movement of the head away from the side touched must obviously lie in the ascendency of the descending tracts which decussate in the cephalic part of the central nervous system over the uncrossed long tracts which descend into the cord. In comparing this condition with the response to stimulation on the tail bud, it should be remembered that the path from n. trigeminus or n. vagus to the opposite musculature of the cephalic part of the trunk is through the descending axones of these nerves within the central system, while the path from the caudal nerves to

96 Journal of Comparative Neurology and Psychology.

the same musculature is through the asceudiug axoues of the afferent nerves. This factor will be best considered in connection with the account of reaction to touch on the tail bud.

The most difficult phase of the problem to deal with bj way of anatomical inference or in the framing of a working hypothesis from the point of view of anatomy is the occasional response directed towards the side touched and the period of irregularity in response that precedes the period of regular movement away from the side touched. It is possible that, in such cases, the impulse passes directly to the centers of synapse with the effectors of the opposite side and, in case these centers are inactive, returns by a commissural path to the corresponding effectors of the same side ; or it might be that the connection with the effectors of the same side is through collaterals of axones which themselves pass directly to the opposite side, and that, in case the opposite effectors are inactive, the impulse may flow over into the collaterals and effect a connection with the effectors of the same side. Two observations may be cited in favor of the latter hypothesis: (1) There is a perceptibly lower degree of irritability during the periods of irregularity and asymmetry in response. My experiments are not exhaustive on this point, but they afford a considerable evidence to this effect, and none to the contrary. (2) The irritability of an embryo may vary perceptibly within a comparatively short period of time. This factor has not been definitely correlated with irregularity in response, but it may be the explanation of the occasional movement towards the side touched during the long period of predominant regularity. Also the very rare irregular movement occurring before a period of asymmetry, as observed above, may have its basis in this variable irritability at some point in the neuro-muscular system.

In some such manner as indicated above my experiments permit of a provisional hypothesis to explain the occurrence of the early periods of asymmetry and irregularity in response of some embryos and the occasional movement towards the side touched, and warrant the conclusion that, for a period of about 48 hours, or more, following the first movements in response to a tactile stimulus, the response of a symmetrically developed, normal embryo of Diemyctylus torosus

CoGHiLL, T he Reaction to Tactile Stimuli. 97

is regularly away, from the side touched when the stimulation is applied to the fields of the n. trigeminus and n. vagus.

B. Response to Siunulatlon of the Tail Bud.

There is no marked regularity in the responses to touches on the tail bud. There is a slight general tendency in some specimens towards movemeHt of the head toward the side touched, but no definite significance can yet be attributed to this tendency. It is clear, however, that specimens that are asymmetrical with reference to stimulation on the head are similarly asymmetrical with reference to stimulation of the tail bud, and that ordinarily the asymmetry with reference to the two points of stimulation extends over approximately the same period.

One other fact concerning the reaction to stimulation on the tail bud is established beyond question by my experiments. The first response to such a stimulus in the very young embryo is a head movement, and as the embryo advances in age this movement still begins in the head region and progresses caudad. Ontogenetically, then, the most primitive conduction paths of the medulla spinalis are longitudinal and afterent, and the crossed paths are secondary, excepting possibly in the most cephalic part where the medulla spinalis may be involved in the crossed paths between the n. trigeminus or n. vagus and the opposite musculature of the trunk. The two halves of the medulla spinalis, therefore, seem to be physiologically distinct during this phase of development. This fact of development reveals from a new source the fundamental nature of the longitudinal divisions of the cerebro-spinal system, at least of the somatic components, as they have been conceived by Herrick,^ Johnston* and others on purely morphological and physiological grounds. It also suggests that in their direct connection with the cephalic part of the nervous system the special cutaneous systems of fishes and amphibians accord essentially with the ]irimary plan of the general cutaneous system.

^"Tlie Cranial and First Spinal Nerves of Menidia," Arc-hives of Neurology and Psychology, Vol. II, and The .Tournal of Comparative Neurology, Vol. IX ; also numerous later papers, mostly in this Journal.

'"The Brain of Acipenser," Zool. Jahrh., 1901 ; "The Nervous System of Vertebrates," Philadelphia, 190G ; and other papers in this Journal.

98 'Journal of Co^nparative Neurology and Psychology.

It would be a difficult thing ordinarily to demonstrate that the receptive fields and afferent conductors become functional in an embryo before the effectors do, for through the effectors alone is the functioning of the receptor and conductor demonstrable. But if the skin of a given somite in the tail bud of an amphibian embryo of suitable age be touched there will be no perceptible response in the effectors of that segment, while response will occur in the older somites farther cephalad. Into this given caudal somite, then, impulses are pouring from the external world through the receptors and conductors before the effectors of that segment are capable of making any perceptible response whatever. If this is tiiie of the more caudal somites, it may be assumed to be true of the head seginents also, and the embryo may be regarded as existing under a storm of impulses of the receptive system for a considerable period before it has the ability to give expression through its effectors. Hovv widely this order of development of the receptor and effector may be applicable, as a law, and what its significance may be are questions of interest. It is possible that the summation of subliminal stimuli in neuro-musc\iIar reflexes rests upon this as a fundamental principle of functional development. It is possible, also, that Kappers^ might correlate this precocity of the afferent system with his theory of neurobiotaxis, in which he assumes that the afferent conductors have influence over the effector centers to cause them to migrate, phylogenetically at least, in the direction of the maximal amount of stimulation.

The Swimming Movement.

The movements of Diemyctylus embryos are of two main types ; (1) the flexure, which is a bending of the body in one direction

""Phylogenetisclie Yerlagerungeu der motorisehen Oblongatakerne, ilire Ursaclie und Bedeutuug." Neurol. Centralbl., No. 18, 1907.

"Weitere Mitteiluugeii beziiglicli der pliylogenetiscben Veiingoruiig der iiiotoriscben Hirunervenkerue. Der Bau des antouonien Systenies." Folia Neuro-Biologiea, B. 1, Nr. 2, Jan., 1008.

"Weitere Mitteiluugeii fiber Neurobiotaxis." Folia Neuro-Biologiea, B. 1, Nr. 4, 1908.

"The Structure of the Autouoniic Nervou.s System compared with its Functional Activity." Journal of I'hysiology, Vol. XXXVII, No. 2, 1908.

CoGHiLL, T Jie Reactioji to Tactile Stnniili. 99

only; (2) the "JS" movement or reaction, which is a bending of the more cephalic and the more caudal parts of the body in opposite directions, giving the form of the letter S.

The flexure may occur in several varieties. It may be a "head flexure," which effects a movement of the head only; a "pectoral flexure," which affects slightly more of the trunk than the head Hexure does ) a "mid-trunk flexure," which is effected by the muscles of the middle portion of the trunk only; a "general flexure," which involves the bending of the whole trunk. In the mid-trunk or pectoral flexure the parts cephalad and caudad of the flexed part may assume positions parallel to each other, in the form of the letter U. This may be designated as the "U" reaction. The general flexure may be extended till the body assumes more or less a coiled condition. This movement may be termed the "coiled reaction."

The various forms of the flexure are not to be considered as essentially distinct, for, with possibly the exception of the U reaction, they develop gradually one into the other in the order mentioned. ]S[evertheless, the distinctions are useful for descriptive purposes.

The first member of this series to appear in the course of development of the embryo is the head flexure ; the next is the pectoral flexure, and, as the embryo advances in age, the flexure extends farther caudad untill it involves the entire trunk in a general flexure, and, finally, in a coiled reaction. In ontogeny, then, the flexure develops cephalo-caudad. This is true for responses to stimulation on the tail bud as well as for responses to stimulation on the head.

In the development of any particular flexure, pectoral, general or coiled, the same progression cephalo-caudad is observed. If the n. trigeminus or n. vagus is stimulated by a touch, the normal reaction is a head flexure, and, if the embryo is sufficiently advanced in age, this flexure progresses caudad until the whole trunk is involved.. In like manner, if the touch is upon the tail bud, the response begins in the head region and progresses caudad. The physiological development of a flexure, then, is correlated with its ontogenetic development.

JSTow, so far as my observations go, the S reaction never appears until the embryo is capable of executing an extended general flexure,

100 "Journal of Comparative Neurology and Psychology.

and rarely until it has actually executed a coiled reaction. Furthermore the S reaction is ordinarily first performed by a reversal of the head from an extended general flexure or a coiled reaction before the original flexure is completed in the caudal part of the trunk. This reversed movement of the head, in early stage of the embryo, may simply progress caudal till it reverses completely the original flexure ; but when the movement attains its typical form it is a relatively short, quick movement, and, when performed in series, it becomes the normal swimming movement.

The occurrence of the S reaction in series has its origin, evidently, in a mode of response which appears very early in the course of development. It may be designated as the secondary reaction." This secondary reaction is a movement that is made during the phase of relaxation from a direct response to an external stimulus. It is caused, probably, by a rhythmic process in the motor cells, or, possibly, by stimuli from the proprio-ceptive field. It may be of greater or less extent than the original flexure. It may, for instance, advance a general flexure into a coiled reaction. It is a conspicuous feature in the behavior up to the time when the S reaction appears.

Now, it is obvious that when the head is once reversed from a flexure into an S reaction, the secondary reaction would explain the second reversal, which is simply repetition of the initial movement. The successive reversals of the head may, then, be initiated as secondary reactions and the progression of the successive flexures caudad, in the form of S reactions, propels the animal forward.

Locomotion, therefore, in the amphibian embrj^o is dependent upon the progression of the flexure cephalo-caudad, and the cephalo-candal progression of the individual movement is further correlated with a similar progression in the ontogenetic development of the reaction. Furthermore, it is clear that this order of development of function is correlated with the order of structural development of the central nervous system, as illustrated, for instance, in the order of closure of the neural tube. These correlations naturally suggest, further, that the necessity of locomotion may have been an important phylogenetic factor in determining the order of development of the parts of the nervous system in vertebrates.

CoGHll.L, TJic Reaction to Tactile Stimuli. loi

Emphasis, properly, lias boon placed, by authorities generally, iqion the principle of cephalization as correlated with the organs of speeial sense; but these early movements of the embryo show that, so far as functional development is concerned, the most primitive centralization of the nervous system, ontogenetically, is in direct response to the demands of the motor sj^stem in its relation to locomotion, while the sensory system involved is not the special sensory but the most primitive, diffuse, exteroceptive field. It remains to locate exactly this primitive center of the cerebro-spinal system by correlated anatomical and experimental studies ; but from the experiments alone, this center would seem to be in close relation to the cephalic musculature of the trunk. This is inferred particularly from the fact that a flexure in response to a touch on the tail bud begins in the head region and progresses caudad and is the same in form (without reference to the initial direction of the movement) as the flexure that follows stimulation of the head. All movements, then, regardless of the point of stimulation, must emanate from the same center.* Into this center all impulses W'Ould seem to flow in order to be directed in such a way upon the musculature of the trunk as to give rise to locomotion. Clearly the development of an eye or ear as such in its earliest functional condition has no part in determining this region of centralization. The controlling factor in this centralization is the motor system : a cephalization in response to the prepotency of the requirements of effectors and not in response to the demands of the cephalic receptive fields.

Phylogenetically, then, the most primitive cephalization of the nervous system may have occurred, also, in response to the demands for locomotion and have given rise to a center of control in the region corresponding to the lower portion of the myelencephalon or the upper portion of the medulla spinalis. Quite in harmony with this suggestion is the convincing evidence that Johnston^ presents for the migration caudad of the afferent roots of the cranial nerves. Such a change in their course would lead them more directly into this primitive locomotor center. Upon this hypothesis, also, the economy

'"The Nervous System of Vertebrates," Chapter III.

102 'Journal of Comparative Neurology and Psychology.

of tlie arrangement of the special cntancous nerves of fishes and amphibians is obvious. It is not to be supposed that the cephalization of the locomotor effectors is, in any respect, a direct cause of the cephalo-caudal migration of the special cutaneous receptors and conductors, but such a cejihalization would certainly favor the development of such systems, for, as already suggested, their peripheral conductors hold essentially the same relation to the cephalic part of the central system as do the most primitive central conductors from the trunk.

It should be noted here that a certain amount of locomotion may be acquired by an amphibian embryo by other movements than the S reaction as described above. The body may be flexed, for instance, and straightened by a series of secondary, vibratory movements. Such a reaction propels the animal on its side in a circle or spiral path. Also, a rapid succession of reversed flexures, in which no S reaction can be detected, may give swimming in a zigzag, erratic course. But normal, upright swimming in a direct course is, according to my observations, attained only through the perfecting of the S reaction and its performance in series.

As already suggested, this development of the swimming movement is of interest from the point of view of animal behavior. We now see that swimming, which may be regarded as instinctive in these forms, arises as the elaboration of the simplest known reflex in the embryo, the contraction of the most cephalic trunk muscles. Certain forms of the flexure, such as the TJ reaction and the coiled reaction, do not seem to be in the direct line of the development of tlie swimming movement, being simply intensive or tetanic forms of the ordinary flexures. On the other hand, the other types of flexure develop in a regiilar order and in a remarkably constant manner into the movements of locomotion. Kow none of these simple flexures can be regarded as having any value as trials, since the Diemyctylus swims perfectly upon leaving the egg membranes in the normal course of development, and within them it can gain no practical experience for swimming out of movements of any sort. Instinctive swimming, therefore, and the simplest reflex alike, are inherent in the neuro-muscular system of the embryo, and while the former de

CocHiLi., Tlic Reaction to Tactile Stimuli. 103

velops ill a rcgulai- order out of (lio lallcr, (he luovciiicnls iliemselves, which conform to this order, can have no selective value. The question naturally follows whether in forms which do not admit of such early experiments, such as birds, many quadrupeds and primates, the various forms of locomotion, as w(dl as otlun* forms of behavior, which, in a greater or less degree, appear to develop out of a series of trials, may not conform to the same law. It seems altogether possible that in such cases, also, the so-called erratic movements may have only a trophic value. As such they would be essential to the perfecting of movements, but would have no directive value in the development of responses.

If, moreover, this hypothesis is valid for the ontogenetic origin and development of instinctive behavior it would seem plausible, also, as a theory of phylogenetic development. Its application to phylogenesis, though, would clearly be in opposition to the idea, which is accepted by various psychologists, that instinctive behavior has somehow been reflected back into the race from the intelligent tyjie, — or psychologically expressed, that instinct is a phylogenetic derivative of intelligence. For the latter hypothesis, I am not aware that there is any direct, experimental proof, while we do see, in such A^ertebrates as Amphibia which admit of early experimentation, instinctive behavior (locomotion) developing directly out of the simplest known reflex. However, while we seem to have a definite conception of the psychic parallel of the former (instinct), the concept of the psychic parallel of the latter is much less definite, and largely disregarded by psychologists. Yet it would seem that in the ontogenetic developments of the psychic life of Diemyctylus there must be puite as definite a reflex-psychosis concomitant with the earliest and simplest reflex as there is an instinct-psychosis with the later instinctive behavior in the form, for example, of locomotion ; for, although the neuroses of the simple reflex are evidently not as elaborate as are those of locomotion, they are quite as definite in form. But, however this hypothesis of the relation of the instinct to the reflex may appeal to the psychologist, an adequate knowledge of the behavior of Diemyctj^lus must take into account the origin and development of locomotion from the simple reflex; for this reflex

104 'Journal of Comparative Neurology and Psychology.

represents the simplest kllo^v^l pliysiological unit of the somatic nenro-miiscular system, or of the somatic "action system." The relation of this unit to any of the more complex neuro-muscular processes is certainly an essential factor in the problem of behavior, or of physiology in the broadest sense.

In presenting the mode of locomotion of the amphibian embryo it is not my intention to antagonize the current explanation of the propelling factors of the swimming movement of fishes, ordinarily described as being, in effect, the same as that of a sculling oar. The latter explanation, so far as I am aware, is offered with reference to the adult fish, and it might not apply to an embryonic or very young fish. Quite conceivably, the swimming movement might become modified during growth, in response to changes in body form, modes of feeding and other factors of behavior ; and it is still quite possible that in the adult fish there is a cephalo-caudal ^progression of movement which is obscured by other factors of special adaptation.

This contribution should not be submitted without reference to the splendid work of Paton"^ on the reaction of vertebrate embryos. This is the only paper accessible to me that bears in any respect immediately upon the work in hand. Paton's contribution, however, is chiefly upon the development of fishes, "with merely a reference to Rana and Amblystoma, and is particularly devoted to the spontaneous movements. Such movements w^ould seem to be much more common in embryos of fishes than in embryos of Diemyctylus. The latter, during the early phases of irritability to touch, may be under observation for hours without making a perceptible spontaneous movement of the trunk, cardiac and branchial movements not being taken into account in my work.

My approach to the problem of physiologico-anatomical correlations in the development of the neuro-muscular system of vertebrates differs materially from that of Paton's method. Paton undertakes "to determine in a general, but not in a specific way" how far the reactions are dependent upon "the functional activity of a nervous

'"The Reaction of the Vertebrate Embryo and the Associated Changes in the Nervous System." Mittheilungen a. d. zoologischen Station zu Neapel, Bd. 18, Heft 2 n. 3, 1907.

CoGHiLi., TJie Reaction to Tactile Sinniili. I05

system" and dismisses tlic study of specific reactions as impracticable, on account of the '^apparently conflicting" data ; but my work clearly demonstrates that, in response to the stimulus employed in my experiments, embryos of Diemyctylus have a veiy definite and regular mode of response, during certain phases of development. In fact I have yet to find the first individual that, through any considerable period, reacts contrary to the mode described in this paper, that is to say, no embryo has yet come under my observation that regularly moves its head toward the side touched when the stimulation is on the head. JSTor have I found a single embryo that, observed for a considerable period, has not fallen under one of the three types which I have here described.

SENSATION'S FOLLOWING NERVE DIVISION.

SHEPHERD IVORY FRANZ.

From tJic Lahoralorics of the Government UospUal for the Insane, Washington, D. C.

With Five Figures.

I. TilE PlU<:SSUEE-LIKE SeNSATIONS.

Since the confirmation and elaboration by Goldscheider, by von Frej and by others, of the discovery of Blix that there are jwints or areas on the skin sensitive only to certain forms of stimulation, physiologists have assumed a form of punctate sensibility in the skin. The work of these investigators has been taken to show that in the skin special nerves subserve the following sensations : heat, cold, pain and pressure (touch). On the other hand, the recent work of Head and his co-workers has not only cast considerable doubt on the validity of the broad generalization of punctate sensibility, but it is also plain that the earlier hypothesis is not in accord with the results obtained on man after injury or section of peripheral nerveSi.

Head, it will be remembered, investigated the sensibility to light touch, to different degrees of temperature, to pressure, to dual stimuli, to pain, to size and to movement in patients following injury or section of perij^heral nerves and he carried his inquiries up to the point of recovery for all forms of sensation. The criticism of the older hypothesis, which was made possible because of these recent pathological studies, may well be summed up in the words of Head: When the median nerve is divided, sensation is entirely lost over a considerable part of both index and middle fingers. Over the palm, within the area said by anatomists to be supplied by this nerve, sensation is usually diminished and not completely abolished.

The Journal of Compaeative Neurology and Psychology. — Vol. XIX, No. 1.

io8 ^Journal of Coniparafivc Neurology and Psychology.

In a similar manner, division of the nlnar nerve produces complete insensibility of the little finger, and of a variable part of the palm and the ulnar half of the ring finger. Such is the usual statement of surgeons and anatomists. . . . The most careful examination of the hand fails to show the slightest diminntion in s(,']isation over the median half of the palm in consequence of division of the nlnar nerve. What has always been called the diminished sensibility produced by the division of a nerve is really a condition in which some kinds of sensibility are lost and others retained." ^

The results of careful examinations of about eighty patients, in whom the nei-ves of arm or leg had been divided or injured, led to the following conclusions : "The sensory mechanism in the peripheral nerves is found to consist of three systems :

(A) Deep sensibility, capable of answering to pressure and to the movement of parts, and even capable of producing pain under the influence of excessive pressure, or when the joint is injured. The fibers subser\'ing this form of sensation run mainly with the motor nerves, and are not destroyed by division of all the sensory nerves to the skin.

"(B) Protopathic sensibility, capable of responding to painful cutaneous stimuli, and to the extremes of heat and cold. This is the great reflex system, producing a rapid widely diffused response, unaccompanied by any definite appreciation of the locality of the spot stimulated.

"(C) Epicritic scnsihility, by v/hich we gain the power of cutaneous localization, of the discrimination of two ]wints, and of the finer grades of temperature, called cool and warm." "

The separate sensation elements in each of the three forms of sensibility may be tabulated as follows :

"Loss of epicritic sensibility abolishes: recognition of light touch over hairless parts or parts that have been shaved ; cutaneous localization ; discrimination of compass points ; appreciation of difference in size, including the accurate discrimination of the head from the

Heacl, Rivers and Sberren : The Afferent Nervous System from a New Aspect. Brain, 1905, Vol. 28, p. 100. 'IMd., p. 111.

Franz, Sensations folloiun^g Nerve Division. 109

point of a pin apart from tlie pain of the prick (acuesthesia) ; discrimination of intermediate degrees of temperature, from about 25^ C. to about 40" C.

Loss of pi'otopathic sens'iblUti) abolishes: cutaneous pain, especially that produced by pricking, burning, or freezing, together with that of stimulation , with the painful interrupted current; over hairclad jiarts, plucking the hairs ceases to be painful ; sensations of heat from temperature above 45 ° C. ; sensations of cold from temperatures below 20° 0.

"After destruction of all cutaneous aft'ercnc iil)ers the part is still endowed with deep sensibility, pressure can be recognized, and its gradual increases appreciated ; pain is produced by excessive pressure (measured by the algometer) ; movements of muscles can be recognized ; the point of application of pressure can be localized ; the patient can recognize the extent and direction of movement produced passively in all joints within the affected area."^

It will be seen, therefore, that the examination of patients in whom nerves have been injured or cut reveals many different degrees or qualities of sensation in addition to the four which, from the examination of normal individuals alone, have been supposed to be the only sensory elements. The main points of difference between the old and the new view — so far as the enumeration of sensations is concerned — are as follows: There is apparently a difference between the sensations of hot and warm, and between those of cold and cool ; touch is different from pressure ; there are different kinds of j)ain.

When, for example, the ulnar nerve has been cut, examination of the skin with various stimulating objects shows that over the hairless portions of the fourth and ring fingers light touches with a wisp of cotton wood or with a fine camel's-hair brush are not felt ; the hair-clad parts may or may not react to such stimuli, depending upon the location of the lesion; parts of these fingers will not be sensitive to temperature stimuli, and perhaps not to pressure ; there may, or may not, be sensations from pricks of a pin ; and if the

^Ileacl and Tlionipson : The Grouping of Afferent InLpnlses within the Spinal Cord. Brain, 1900, Vol. 29, p. 551.

no 'Journal of Co?nparative Neurology and Psychology.

fingers can not be moved voluntarily there will be a loss of sensibility to movement passively produced. Some of these effects may be found over a variable extent of the palm and the back of the hand. In rig. 1 is given the condition found in a man following an operation in which part of the ulnar nerve was excised. The area insensitive to light touch and to the intermediate degrees of temperatures included all the little finger, about three-quarters of the ring finger, and about two-fifths of the palm and back of the hand. Part of

Fig. 1. — The extent of loss of sensation following the division of the ulnar nerve at the elbow. The part marked with horizontal lines was insensitive to light touches, and to intermediate degrees of temperatures. The vertical line area (cross-hatched on account of its being included within the area insensitive to touch), was, in addition, insensitive to pressures and to pain. Adapted from Head and Sherreu.

this area was insensitive to deep touch and no sensation was got from the vibrations of a tuning fork. This area was also analgesic.

In this and other cases in which losses of sensibility were found, the ability to appreciate touch was tested with a wisp of cotton wool lightly brushed over the parts. When such a piece of cotton wool is carried over the skin of a normal individual, there is a distinct feeling of touch, which is magnified, perhaps altered, wherever the hairs are touched. The cotton wool should be very lightly grouped in a bundle, not tied, and I have found that on the hand a piece of cotton wool, weighing 55.5 mg., with a bending pressure of from 200 to 300 mg. will be accurately appreciated over the parts

Franz, Scnsatiojjs foUmvnig JSlerve Division. Ill

wliicli are not calloused."^ On the lips and parts of the face, a wisp of cotton wool weighing 20 mg. and bending at 200 to 250 mg., was just sufficient to produce a sensation. At times it is more convenient to use a camel's-hair brush, although Head has objected to the use of this instrument. I find it to be a more constant stimulus, in that it remains the same in bending strength for long periods. With cotton wool it is difficult, if not impossible, to select for each day's series of experiments an amount equal to that used on previous days, and if the same piece be used on successive days, it soon loses its original strength. In the experimental results that follow, I have used both cotton wool and a camel's-hair brush. I selected a long haired brush from which I cut off most of the outside hairs, leaving a brush 24 mm. long from the end of the hairs to the insertion, with about 125 hairs. As thus modified, the bending strength of the brush was 100 mg. for very slight bending, and about 200 mg. for more extensive bending. These figures are to be compared with the bending strength of the wisps of cotton wool mentioned above. I have found that the same results follow the use of the brush as the use of the cotton wool, and since, as mentioned above, it is more constant, it can be used for many patients as well as the same patient at different times.

For further tests I have employed the touch instrument of Bloch, which is illustrated in the accompanying figure. To a piece of wood was attached a spring steel wire A which was bent at a right angle B ; the long part of this wire A measured six inches. The area of cross section was about 0.1 square mm. A scale E attached to the instrument enabled the experimenter to determine the pressure made by the wire in bending. The right-angled "piece of wire was pressed against the subject's skin, care being taken to keep it vertical all the time, and when the patient reported that the pressure was perceived, the reading was taken from the scale and recorded. Two instruments of this character were constructed, one of which was

IIead and Sherren do not give any figures regarding tliese matters. In some later vvorlv, Head used a series of von Frey hairs with bending pressures of 830, 3G0, 230 and 100 mg., respectively. Tlie use of tliese, however, more nearly apjiroaches the use of the esthesiometer which will he described later.

112 'Journal of Comparative Neurology and Psychology.

approximately twice as strong as the other, in reality one division on instrument A equaled 0.6 division on instrument B. The instrument A enabled me to stimulate with tensions from to 2000 me"., while instrument B enabled me to stimulate with tensions from to 4000 mg. Each instrument was divided into arbitrary divisions, circular degrees, and since I was dealing with relative differences I did not translate the figures into terms of tension. All the figures that are included in the following tables are, however, directly comparable.'^

Fig. 2. — Esthesionieter of Blocb. Tbe figure is about one-tbird tbe size of tbe instrument.

The j)resence or absence of pressure sensations was determined by stimulating the parts with the pointed, l>ut not sharp, end of a pencil. On normal parts this stimulus was immediately appreciated. Pain from pressure was determined by an algometer. The surface

'Tbe value of such an instrument for determining differences in toucb sensations was well brougbt out in some experiments, not yet publisbed, on the touch thresholds in different parts of the body. One subject whom I examined carefully showed an increase in threshold values in tbe lower leg and foot on tbe left, although no similar difference was foiuid to exist in other parts of tbe body. This leg was also smaller than tbe right, in circumference. When asked to explain the latter dift'erence tbe subject replied that both knee and ankle of this leg bad been injured at two different times and that for a period of six months six years previously she bad this leg bound up and almost immovable. Another subject, a case of traumatic epilepsy, showed an increase in threshold values for touch on tbe right upper arm and shoulder, on tbe back over the scapula and over tbe chest above tbe breast, although otherwise both sides were the same. This was tbe only sensory or motor change that I found in this woman, and this persisted after tbe operation that was undertaken for her relief. In neither case did tbe testing widi cotton show the least difference between tbe two corresponding segments of tbe body.

Franz, Sensations foUoiving Nerve Division. 113

applied to the skin was circular, 3 mm. in diameter. This instrument was used instead of the usual algometer of Cattell, because it was necessary to make determinations on skin areas so close together that the large stimulating surface of the Cattell instrument would have been impracticable; and since the object of the work was to determine relative amounts of pressure causing pain, the smaller instrument was the more advisable.

The subject of the experiments is II., male nurse in the Government Hospital for the Insane, age 38, of good education, and interested not only in getting well, but also in the results of the ex]:»eriments. For these reasons he was excellent in co-operation and the results were to be depended upon more than those from most patients, who may fear that the knowledge of their condition will be a hin

FiG. 3. — Arm and hand of subject copied from photograph. Area insensitive to light touch marl^ed with horizontal lines.

drance to their obtaining employment. His training in the examination of the insane was a benefit to him in noting his o\vn condition, and his answers showed that his use of terms to describe his sensations was more accurate than that of similar untrained individuals, and the results are to be depended upon more than in those cases in which subjects have not been accustomed to note and to analyze mental symptoms. On June 20th, an inmate of the Hospital during a period of confusion attacked H. with a large pen knife and cut him in the left arm above the elbow, the wound extending 12 cm. from the dorsal side across the triceps muscle to the inner bend of the elbow. The lower part of the wound is to be seen in Fig. 3.

Immediately following the accident the patient was taken to the operating room. The triceps was found to be almost completely severed, and this was stitched toirether after the wound was thor

114 'Journal of Comparative Neurology and Psychology.

ouglilj irrigated. No attempt at this time was made to discover nerve lesions. The wound healed by first intention. The day following the accident I saw the patient. So far as he could tell the thumb felt "dead" on the volar side, but there was no loss of sensation in the forefinger. The second, ring and little fingers also felt dead". The area of anaesthesia to light touch, which was discovered at this time was over the ulnar j^art of the hand, back and front, and included the three fingers mentioned. I did not make careful notes of the condition at the time, and the accurate comparison of the anesthetic area at this time with what was found later can not be made. Following the accident the feeling in the thumb improved, but the sensation ability of the forefinger decreased. On August 12th, an exploratory operation was performed by Dr. G. T. Vaughan. The ulnar nerve was found divided near the olecranum ; the median nerve was nicked, and there were excrescences or clubbed swellings on both edges of the nick. The ulnar nerve was brought together with sutures, the swellings on the median nerve were cut away and pieces of fascia were placed around the nerves to j^revent the ingrowth of any scar tissue from neighboring parts. The results of my later examinations show that, in all probability, the medial antibrachial nerve was severed. This was not examined at the time of the operation. Immediately following the operation there was further improvement in the condition of the thumb, but the patient thinks this improvement was not marked. I did not examine the patient again until October 6th.

At the time of the first careful examination, I found the following condition: The little and ring fingers were totally anesthetic to all forms of stimuli, heat and cold, light touch and pressures. The middle and forefingers were anesthetic for all forms of stimuli over the two distal joints, both on the back and palmar parts. The palmar area corresponding to these two fingers was insensitive to light touch, but pressures could be appreciated and well localized. Pricks of a pin were painful. Up to this time there had appeared from time to time painful sensations following movements of the arm at the elbow, localized poorly, but apparently over the knuckles of the middle, ring and little fingers and over the palm at the place

Franz, Sensations folloirnig Nrrvc Division.

1^5

where these fingers join the pahn, Longitndinally the pains seemed to extend ahont three qnarters of an inch, bnt, as said al)()ve, no very aeenrate localization can be made. In addition, heat (test tube with water heated to 45° C.) did not produce a sensation over the area insensitive to light tonch and over half of the nalmar part of

Fig. 4. — ILiiid sliowiiig areas insensitive to pressure, to li,i,'lit toneli and to temperatures.

the thumb. Cold was not ap])reciated over tlie same area, although the test tube was cooled to 0^ 0., but there seemed to be a dissociation of the areas concerned with hot and cold sensations, for on the thumb, with the exception of the tip, no hot sensations were evoked, although cold was properly appreciated over the thiunb with the exception of the first joint. Fig. 4 illnstrates this condition.

Il6 ^Journal of Comparative Neurology and Psychology.

Movement sensations were tested as follows: The eyes of the patient were closed or turned away so that he did not see what was being done. The left Avrist was moved in the four possible directions and he properly duplicated these mo^-ements with his right wrist. Similar experiments on flexion and extension were made with the thumb, forefinger and second finger and similar results obtained. When, ho^ve^'er, the ring and little finger were moved the patient gave no indication of the movement of these parts. ]\[oreover, when the thumb and fingers were moved together so as to make a fist, the patient moved only the thumb and the first and middle fingers of the right hand. It was evident, therefore, that the jxatient had no sensations of movement from the little and ring fing(n-s. Voluntary movement of the fingers of the left hand were poorly executed. The thumb could be moved only slightly, and this principally in flexion. The forefinger and middle finger also could be slightly flexed, while the third and fourth fingers could not be moved to the slightest degrees. J^o voluntary power to flex or to extend the wrist could be found. During the next ten days at times poorly localized, rather radiating, pains were felt over the ring and little fingers and on the back and outer aspect of the hand, i^o pain was felt on pressure of the fingers, nor on movement of the individual fingers. Pain, but slight, was felt on j)assive extensor movements of the wrist, but none on adduction and abduction. Pain at elbow on extension. During this period the subject complained of tingling sensations extending from the hand to the elbow, at times this pain was so severe that it prevented sleep. The flexion at the wrist was about 45°, extension about 20°. Flexion at elbow almost to the limit, extension about 150°. Very slight voluntary flexion of the first, second, and third fingers, but none of the fourth finger. A week later movement of the first and second fingers and thumb were properly sensed, movements of ring finger were sensed as movements, but patient could not tell which finger was being moved. ISTo sensations from movements of the fourth finger. Side to side movements of the middle finger were sensed, but in attempting to duplicate them on the right side they were always made in the opposite direction, although the similar movements of the first finger and thumb

Franz, Sciisdt/ons followmg Nerve Division. 117

were jiroiici'ly expressed with the corresponding members of the uninjured hand. These results indicate clearly, I believe, that the motor sensations from the little finger were entirely unappreciated, those of the tlnnnb and forefinger w^ere appreciated in a normal manner or dc^gree, while there w^as a change in appreciation of the movement sensations from the middle and ring fingers. The fact that the movements of the ring finger were sensed as movements does not necessarily imply that movement could be apj)reciated. Such movement sensations may have been due to the movement of the skin covering other fingers and palm, which gave more normal sensations, or they may have been due to movement of tendons or other tissues in the palm or back of the hand. It appears probable to me, however, that souie movement sensation persisted in this finger, or, rather, w^e should say that the nerves conveying the impulses for movement sensations had sufficiently regenerated to enable a sensation to be produced. The results with the middle finger, especially the sideway movements, are extremely suggestive of an improvement in this part of the hand.

During the days that experiments were carried out, many times the patient reported that light touches, touches with cotton wool or a brush, over hairy parts were different in character than those over other normal parts. The way this was always expressed was that the sensations were stronger, or clearer. From the reports of Head's work, one may understand that there is a rather sharply defined line of separation of the parts sensible to cotton wool from those insensible to the same sort of stimulation. That this is not always so, w^as shown in my experiments. The patient would often feel the stimulus, and then again miss it in certain places, while in neighboring regions the stimulus often failed, but occasionally did produce a sensation. That there is no sharp line dividing the 'epicritic' and 'protopathic' areas is shown also by the results of a different sort of stimulation. If, instead of cotton wool or the light brush, we use an instrument that will enable us to give different degrees of stimulation, we find the area insensitive to cotton wool does not give a sensation on stimulation, even with considerable amounts of stimulation. N^eiffliborinc: areas in which there can be

Ii8 Journal of Comparative Neurology and Psychology.

no doubt of the presence of both cpicritic and protopathic sensibilities do not show a normal amount of sensibility to this form of stimulation. There is a gradual decrease in the loss or in the dullness of sensibility from the area of cpicritic loss, Ave may sav from area of complete sensibility loss, to the area in which no sensory change, according to Head's methods, can be demonstrated. The accompanying figure illustrates the conditions found on the hand.

The area tested was carefully examined for hairs, and wherever hairs were found, the ])art was carefully shaved. The part of the area insensitive to cotton wool was mapped out and recorded on the

Fig. 5.— To illustrate the sensibility of the hand determined by Bloch's instrument, compared with the loss of pressure and light touch sensations.

hand with red ink. The sensibility to pressure and to forms of stimuli denoting the presence or absence of protopathic sensibility Avas also tested and the areas of loss were marked on the part. Points were then selected separate from each other by 1 cm. These were also marked with red ink so that the stimuli could be eiven at approximately the same spots each time. These points were then examined with the help of Bloch's instrument illustrated above. Five experiments were made on each point and on corresponding points on the right hand. Table I gives the results of this form of examination. The letters indicate the corresponding areas on the left (nerve cut) and right hands. The figures give the averages of

Franz, Sensations folloiumg Nerve Division.

119

five determinations of the just perceptible determined by one of the Bloeli instruments.

TABLE I. Touch Thresholds on Hand. Block Instrument.

Points.

Left hand.

Right hand.

A

6.0

B

8.6

C

02.*

6.0

D

29.8

10.6

1-]

27.8

14.8

F

13.4

7.4

Approximate, since instrument reads only to 60. The point A on the left hand was within the area in which pressure (pencil) was not felt. B was within the area in which pressures w^ere felt, but no sensations accompanied stimulation of cotton wool or a caiuers-hair brush. C, J), E, and F, were areas in which the epicritic sensibility was found intact. It will be noticed from the table that the amount of stimulus to produce a sensation at point C was twice the amount of that necessary on the other areas, D, E, and F, and that it was greatly in excess of that required on the normal hand. The points D, E, and F, supposedly normal, showed an increase in the threshold value^ — approximately double the normal. On the fingers similar results were found. On the parts in which pressure of the pencil was immediately perceived, but no sensibility to cotton wool, the limit of stimulation with the Bloch instrument was not felt, and in the areas in which epicritic sensibility was retained there was an abnormally great threshold value. Only in those parts where cotton wool could be felt was it possible to make any determination with the Bloch instrument, and when the results in these experiments were compared with the results of the right hand, the difference between the apparently normal area of this left hand with the known normal area of the right, are as strikingly marked as those in the table just given, and those in the table mentioned below. On apparently normal points of the thumb of the left hand, sixteen exi^eriments with instrument A gave

120 'Journal of Comparative ~N eurology and Psychology.

a threshold average of 29, while a similar, nuiiiber of corresponding points of the right thumb gave an average of S.T. On the ball and on the back of the third joint of the thumb eleven experiments on the left hand averaged 42,1, and on the right, 20.5.

On the upper part of the forearm, a section including the area in which different forms of sensibility were altered, was carefully crossmarked in centimeter square divisions. This section of the area was carefully shaved so that these experiments could be carried out. Each of the thirty-five squares was then carefully tested five times. The averages were calculated and the results of these examinations are given in Table II.

TABLE II. Touch Thresholds on Volar Side of Forearm.

Bloch Instrument B.

Points.

A

B

C

D

E

F

G

Aver, threshold

1QQ

25.5

34.9

54.1*

46.*

54.*

Ap])roxiinate. The results for the five horizontal centimeter squares are grouped in this table. Below G the arm did not respond to temperature stimuli, nor to the stimulation of the hairs. Areas A, B and C reacted to all forms of temperature stimuli with sensations of hotness, warmth, coolness and cold, while area I) responded to only some of these. These areas also gave results to brushing and pulling the hairs, while plucking the hairs in areas E, F and G gave no sensations. A full account of the results on the hair and temperature sensations in these areas will be found in a subsequent article. It is sufficient for the present to know that areas E, E and G were clearly devoid of epicritic sensibility, M'hen areas A, B and C were, in accordance with Head's methods of examination, clearly normal.

In G and F the stimulus was perceived only once, although tlu; instrument was pressed to its highest point. Twice in area E stimuli were perceived, seven times in area D, and always in areas C, B and A, w^hen the pressure of the tension of the instrument became sufficient. A comparison of these results with the results of a similar examination of the right arm show that the stimulus necessary to

Franz, Sctisations follrnvnig Nerve Division. 121

122 'Journal of Comparative Neurology and Psychology.

the ulnar side from a line drawn tlirongh the middle of the ring finger; the superior radial nerves presumably the radial side to a line dra\^al through the middle of the thund); while the remainder of the palmar part of the hand is innervated by the median nerve. All of these areas must have, according to Head, deep, protopathic and epicritic sensibilities. But Head's work shows that Ave seldom find areas of these different fomis of sensibility occupying the same position on the hand. For the median, for example, there may be a loss of deep sensibility in the first and second fingers; but the rest of the area which is assigned to the median retains its deep sensibility. The retention of this sensibility must be due to fibers in other nerves. So also we find there is an overlapping of the epicritic sensibility, and it seems to me the results on H, as well as the results olitained by Head, clearly show such an overlapping effect. On the arm, the areas A, B and C, Fig. 3, were supposed to retain all their epicritic sensibility, whereas it is evident that areas B and C are not nearly so sensitive as area A.

The tests for threshold of pain sensations brought out no new facts. The areas which were carefully examined for touch thresholds, and which gave the results recorded above, were examined, but, with perhaps a slight increase in threshold over the whole lower arm on the left, there was nothing distinct. The measurements made with the algometer described above showed that in the area near the elbow (rectangle in Fig. 3) for, each horizontal centimeter above the line of loss of deep sensibility there was no difference. The averages of three experiments in each longitudinal area are as follows: 9Y8, 905, Y79, 820, 970, and 923 grams. Similarly on the hand there were no sufficiently noticeable differences between the threshold in area B (Fig. 5) from that in area F.

The pains on movement of fingers, wrist and elbow are of a different character from those produced by pressure, and they have been mentioned in an early part of the article. They were not located at the place where the movements occurred, but were always referred to the fingers or hand. It seems likely that these were due to pressure on the nerve trunks, but this is purely speculative and is supported by only a few observations as follows: At times

Franz, Sensations follownig Nerve Division. 123

the i)aticiit reported jiain sensations produced by draAving his shirt sleeve up above the elbow ; when the arm was pressed, and especially Avhen pressure was exerted along the course of the nerve trunks, the subject felt pains in the fingers and palm, of a radiating character and apparently widespread.

The results of this study may be summarized as follows: There is a widespread disturbance in touch sensations following injury to the peripheral nerves, not only in the loss of certain forms of sensation but also in the increase of threshold values in much larger areas of the body segment than those in which there is sensation loss. This decrease in sensibility is marked for touch sensations, but is also apparent for the sensations that arise from stimulating the hairs and for the temperature sensations. The hair and temperature sensations will be more fully discussed in a subsequent article. With the exception of the areas in which no pain from pressure was felt, the pain thresholds are not altered, at least not in the same way as those for touch.

Note : — Tlie results of the careful exaniiuatioiis of Dr. H. Head's arm in which the radial nerve was cut for the purpose of testing the sensibility have .inst been received (Rivers and Head: A Human Experiment in Nerve Division. Brain, 1908, Vol. 31, pp. .32.'>-4.~)0 ) . Owing to the recent appearance of this work, a note of the main features of difference cannot be added to the present article, but will be included as part of the second artice of the series.

[Received for publication, December 5, 1908.]

ALTERATIONS IN THE SPINAL GANGLION CELLS FOLLOWING NEUROTOMY.

Bt

?!. WALTER RANSOX.

From the Anatomical Lahoratonj of the Unlccrsity of Chicago.

With Six Fioures.

Introductiox.

The work of Nissl, Liigaro, and a considerable nnmber of other investigators has priced iis in possession of the essential facts concerning the iiner strnctnre of the spinal ganglion cells both nnder normal and pathological conditions. It is not the purpose of this paper to supplement in any way the excellent cytological studies of these ear'ier investigators, but rather to answer certain questions suggested by an investigation recently published from this laboratory under the title of "Retrograde Degeneration in the Spinal Nerves" (Ranson'06).

It was shown that two months after the section of the second cervical nerve in young rats 52 per cent of the nerve cells had disappeared from the corresponding ganglion. The average loss was 4412 nerve cells, the average normal ganglion containing 8451 such cells. Since there w^ere only 1500 medullated fibers in the nerve at the time of the operation on 12-day-old rats, it is remarkable that its division should have caused the destruction of so large a number of cells.

Still more difficult to explain is the fact that the degeneration of 52 per cent of the ganglion cells was accompanied by a loss of only 17 per cent of the dorsal root fibers. In the hope of finding an explanation for these results, the literature dealing with the architectural relations of the afferent elements entering into the formation of the spinal nerves, has been carefully reviewed in another paper (Ranson '08).

From a study of the literature, it is evident that the cells of the

The Journal of Compakative XEUitoiAXiY and Psychology. — Vol. XIX, No. 1.

126 'JoiirnaJ of Comparative Neurology and Psychology.

spinal ganglion slion'd be divided into two types represented by the large and the small cells which diU'er both anatomically and physiologically from each other. The large cells, which in the second cervical ganglion of the white rat constitute a third of the total number, are connected with medullated fibers. The remaining twothirds are small cells and these are associated with non-medullated fibers which divide injto central and peripheral fibers after the manner of a T or Y. These non-medullated fibers can be traced into the dorsal root and toward the periphery as far as the junction point of the afterent and efferent fibers.*

It is believed that these fadts, together with the observations in the present paper, make it possible to explain the conflicting resuHs presented in the paper on Retrograde Degeneration." At the time that paper w^as written there were certain technical difficulties in the way of deciding whether any particular type of cell was chiefly affected by the destructive process. This question had

After this paper had gone to the press the importaut book of Dogiel, "Der Bau der Spinalganglien," came to hand. He finds eleven tj'pes of spinal ganglion cells according to the character of their processes, with the greatest conceivable wealth of collaterals and dendrites. Nevertheless in seven out of the eleven types the fundamental character of the spinal ganglion cell is maintained in that the axons divide after the manner of a T or Y into central and peripheral fibers. Our belief that the vast majority of the fibers arising in the spinal ganglion divide after this manner is thus confirmed, although a hurried examination of the book might lead to the opposite conclusion.

In the four remaining types (III, IV, VIII and XI) he was unable to demonstrate this division and the destination of these axons remains hypothetical. For the piu-poses of the present paper these types take the place of his old Type II which he has abandoned. He thinks it probable that these cells are not connected with fibers in the peripheral nerve. In this paper we have attempted to show that there are cells in the ganglion which because of their failure to react to lesions of the peripheral nerve can have no axon in the nerve. These non-reacting cells would correspond as well to his new Types III, IV, VIII and XI as to his old Type II.

He still maintains that the small cells give rise to non-medullated fibers and the large ones to medullated fibers. On page 33 he states that the axon "after its exit from the capsule becomes covered with a sheath of myelin or in the case of the small cells and fine fibers it remains non-medullated even to its bifurcation." In Type III, which consists of large cells, all the axons are medullated, and in Type X, which consists of small cells, all the axons are non-medullated.

R ANSON, A If ('rations in Spinal Ganglion Cells. 127

to be answered, however, before any solntion of the problems, already mentioned, could be reached; and, accordingly, the prime purpose of the present investigation is to follow the various stages of axonal reaction in sections, prepared for that purpose, in order to determine whether it is chiefly the large or the small cells which undergo complete disintegration.

ISTaturally, there were also other problems which engaged the attention as the investigation progressed. The more important of these can be stated briefly as follows :

1. Can the cells of Dogiel's Type II, or his new types III, IV, VIII, and XI, which send no axon into the nerve, be identified by their failure to show chromatolysis after the section of the peripheral nerve ?

2. What proportion of the spinal ganglion cells react to the section of the nerve ? When it is remembered that there are on the average three spinal ganglion cells for each medullated afferent fiber in the nerve, and that most observers have foimd nearly all the cells showing the axonal reaction, this question becomes a very pertinent one.

3. If all the cells react, is there any indication that the reaction in some is secondary to that in others ? That is to say, do the small cells, which do not have medullated axons in the nerve, react so much later than the large cells that one might interpret their alterations as due to a j^urely intra-ganglionic disturbance ?

4. What is the nature of the degenerative processes which, as has been shown, lead to the disappearance of about half the cells ?

5. Is there any special locality in the ganglion where the cell destruction is most marked, or any part which remains intact ?

6. Can the cells wdiich survive be followed through the stages of repair until they are again of normal appearance ? .If so what type of cell is most likely to undergo repair ?

Technique.

The second cervicail nerve of the right side was cut under aseptic precautions in white rats 12 days old. The technique of the operation was the same as that described in the paper on "Retrograde

128 Journal of Coinparafive Neurology and Psychology.

Degeneration." The operated ganglia, along with the control ganglia of the opposite side, were removed after varying periods and fixed in Hatai's flnid (Ilatai '01). After penetration with paraflin, they were cnt in serial sections 6 microns thick and stained with toluidinblue. The 18 rats were allowed to survive the O'peration for different lengths of time, which were as follows: 5, 0, 7, 8, 1), 10, 12, 16, 17, 20, 25, 27, 31, 34; 37, 41, 44, and 57 days.

Types of Cells.

A most exhaustive account of the cytology of the spinal ganglion cells and of the axonal reaction, resulting from the lesion of their peripherally directed processes, is to be found in the publications of Lugaro. His contributions to this subject, which have appeared from time to time during the past ten years, have been summarized by David Orr ('04) from whose excellent review the following citation has been taken. According to Lugaro, there are five cell types in the spinal ganglia of mammals: (1) Large clear cells, with granuliform chromophile elements scattered almost equally, throughout the protoplasm with a slight increase toward the cell periphery. The nucleus is large and clear and there is a definite j^erinuclear and peripheral clear zone. (2) Large and medium sized cells with very fine chromophile elements, which are larger, however, at the periphery of the cell. Other characteristics as in Type 1. (3) Small dark cells with very fine chromophile elements, which are larger around the nucleus with which they are in contact. The fundamental substance and the nucleus are diffusely colored.

(4) Small and medium sized clear cells with chromophile elements somewhat large and few in number. The nucleus is clear and separated from the chromophile substance by an achromatic zone.

(5) Large clear cells with elongated chromophie elements lying concentrically around the nucleus in parallel planes. The nucleus is always excentric.

There can be no doubt but that cells corresponding to all these types can be seen in the spinal ganglia of various animals ; but the number of transitional forms is so great that anv such elaborate

Ranson, AUcraiions in Spinal Ganglion Cells. 129

I classification, however valuable it iiiaj be as an aid to description,

must be more or less artificial.

A simjDler classification has been given by Cox ('98) who makes size the basis of differentiation; the large and the small cells form his two main groups. The large cells fall into two types. The cells of Type I present granules of irregular size and shape, which only in the periphery of the cell have an elongated form. There is no concentric arrangement of the granules, and the nuclei are approximately centric. Type II comprises cells with large elongated granules which have a tendency to arrange themselves in rows. The nucleus is excentric. The small cells resemble, so far as the character of their chromatic substance is concerned, the large cells of Type I. Hatai (1901) has also emphasized the importance of size as a basis of classification of the spinal ganglion cells.

Warrington and Griffith ('04) have adopted a classification which is, in the main, very satisfactory and which really dift'ers but little from that of Cox. These authors merge Lugaro's Types T and II into one and call them the "clear cells." These correspond to Cox's large cells of Type I. Lugaro's Type III is accepted and comprises the "obscure cells." These correspond to Cox's small cells. Lugaro's Type IV is divided into two groups. The larger ones, which are all of medium diameter, are called the "coarsely granular cells" and correspond in all probability to Cox's large cells of Type II. The small cells of Lugaro's Type IV are called the "smallest clear cells." They are rare, but differ markedly from the other small cells in that the protoplasm is not diffusely colored. These cells are not given a place in Cox's classification.

According to Warrington and Griffith, the large, clear cells represent about 25 per cent of the total number; they vary in size from 35 to 100 microns. The coarsely granular cells are much less numerous, re])resenting about 4 per cent ; they are of a rather uniform diameter of 35 to 50 microns. The obscure cells varying in size from 10 to 5(5 microns constitute G8 per cent or about two-thirds of all of the cells. In the following table are given the size and relative number of the various types of cells as Warrington and Griffith found them in the second cervical ganglion of the cat. The corresponding types in the classification of Lugaro and Cox are also indicated.

130 journal of Comparative Neurology and Psychology.

It should be stated that, while Lugaro used a variety of mammals, Cox confined his studies to the ganglia of the rabbit and Warrington and Griffith used only the cat. It is very probable that this is in part res2)onsible for the fact that there is not perfect agreement in results.

TABLE I. Types of Cells in the Second Cervical Ganglia of the Cat.

Warrington and Griffith.

Lugaro.

Cox.

Size.

Clear cells !TypesI&II Type I 35-100

Obscure cells Type III Small cells 10-56

Coarsely granular cells . . . .' Type IV , Type II (?) | 35-50

Smallest clear cells ! Type IV 10-25

Relative Number.

25.% 68.1% 4.3% 1.9%

In considering such a classification the question arises as to just what significance is to be attached to the grouping adopted. Do the cells belonging to such a group retain permanently their special characteristics; or may a cell pass from one tyj^e into another, as physiological conditions change ? In other words, does a type represent a group of cells anatomically distinct or only a set of cells which happen to be at the moment of fixation in the same physiological phase ? Lugaro is of the opinion that his five types of cells must be considered as specifically distinct from an anatomical point of view. But the study of the reparative changes in the ganglion after section of the nerve has led me to regard it as extremely probable that the arrangement of the Nissl granules is quite as much an expression of the functional phase of the cell as of its anatomical type. Thus, in the repair following chromatolysis the cells can be followed as they pass in the course of a few days from one well-marked type of finely granular cells into another well-marked type of coarsely granular cells. The contrast between these two states of the same cell is much greater than that between the cells of Lugaro's Types I and II.

We shall, therefore, not attempt to classify the spinal ganglion cells according to the arrangement of the tigroid substance alone,

Ranson, Alterations in Spinal Ganglion Cells. 131

but, accepting DogieFs classification based on the external morphology of the neuTOiies, attempt to fix upon the arrangement of the chromatic bodies characteristic for each of his type of cells. This leads ns to a classification which is almost identical with that of Cox and which agrees in all essential points with that of Warrington and Grifiith. We distinguish the following grouj)s of cells:

1. Small cells.

2. Large cells.

(a) Medium sized coarsely granular cells. (h) All other large cells.

There can be little doubt as to the justification for making size the prime criterion for. the separation of the spinal ganglion cells, since this classification is supported by a variety of other observations. We find that the large and small cells differ from each other in their reaction to the protoplasmic dyes ; the small cells are associated with non-medullated fibers, the large cells with medullated fibers (Dogiel) ; and while the latter react to electrical stimulation of the nerve, the former do not (Hodge). It will also be shown that they differ in their reaction to section of the nerve. All the facts justify us in regarding the large and small cells as fundamentally different from one another. It should not be supposed, however, that any particular diameter of cell body can be fixed upon as a dividing line between the two types. There are medium sized cells which in all other respects resemble the large cells, and others of the same size which present all the characteristics of the small cells.

1. The Small Cells. The characteristics of the small darkly staining cells have been sufficiently described in the citations already given. Their, relatively large nucleus and scanty cytoplasm, staining deeply with the acid dyes, separate them quite sharply from the large cells. The arrangement of the ITissl granules presents nothing peculiar. Their tigroid substance is most often in the form of very fine granules, but is sometimes aggregated into masses of medium size which may be most abundant either near the nucleus or at the ]ieri])hery of the cell. The small clear cells described by Warrington and Griffith are not present in any considerable number in the s]3ina]

132 journal of Comparative Neurology and Psychology.

ganglion of the white rat. The small cells constitute about twothirds of all the cells in the ganglion. They correspond to those described by Dogiel, and according to him possess non-meduUated fibers. A full consideration of the small cells will be found in a previously published paper (Ransou '08).

2. The Large Cells. The large clear cells are characterized by their clear protoplasm. The size, shape, and arrangement of the chromatic granules presents such an infinite variety that any attempt at classification based on these alone would be largely artificial. We know, however, that among the large cells are some — called by Dogiel cells of Type II, or his new types III, IV, VIII and XI, — which do not send any axon into the peripheral nerve; and, since these would not show axonal reaction after division of the nerve, it should be possible to identify them by their normal appearance in a ganglion in which all the other large cells show chromatolysis. Suggestive observations have been made in this connection by Cox, and Warrington and Griffith. The two latter investigators found that after section of the nerve at a point just distal to the spinal ganglion all the cells in the ganglion with the exception of the coarsely granular medium sized cells and the very smallest cells in the ganglion showed chromatolysis. jSTo lesion of the peripheral nerve would produce any alteration in these coarsely granular cells. In the cat these cells present a rather constant diameter of from 35 to 50 microns. Similar observations were made by Cox upon the rabbit, but the non-reactive cells which he found differed from those of Warrington and Griffith in the concentric arrangement of the elongated tigroid masses and in the excentric position of the nucleus. In short, they were the cells which constitute his Type II. Cox regards his Type II as identical with Dogiel's Type II, since the component cells are few in number and do not give evidence of any injury 'to their axon after section of the peripheral nerve close to the ganglion. W^arrington and Griffith suggest the same possibility for their coarsely gi'anular cells.

It seems probable to the writer that these two types are really the same and that in some animals the tendency to concentric arrangement is more marked than in others. This is borne out bv the fact

Ranson, Alter ati on s in Spinal Ganglion Cells. 133

that the cells with elongated conceutric granules are not present in the cat, where Warrington and Griffith find the non-reactive cells to be of the coarsely granular form.

In the white rat, according to my own observations, the concentric type is very rare, oidy t^^'o or three typical instances have been found. On the other hand, a number of the medium sized coarsely granular cells show a sufficient elongation of the granules to suggest somewhat the other type. These cells in the spinal ganglion of the w^hite rat, therefore, resemble closely the non-reactive coarsely granular cells described by Warrington and Griffith in the cat, but also show some affinity with the cells of Cox's Type II ; and it is of considerable interest to note that, like these cells of the other investigators, they also fail to react to lesion of the nerve.

Since in my experiments only the dorsal ramus of the second cervical nerve was cut and about 13 per cent of the total number of afferent fibers running in the ventral ramus escaped lesion, there are a few cells of all types which fail to react. While this tends somcAvhat to obscure the picture, it is possible to determine that the unaltered coarsely granidar cells are present in nearly as great numbers as in the normal ganglia (Fig. 3) and it is clear that few if any of the cells of this type have suffered alteration as the result of the section of the nerve. These non-reactive cells in the spinal ganglion of the white rat are all of medium size and present a clear protoplasm with large chromatic granules which in some cases show a tendency toward a concentric arrangement. These cells are not very numerous, constituting only a small percentage of the total number of cells.

It is, therefore, clear that among the large cells of the ganglion there are a few cells of a fairly definite type which fail to react to a lesion of the nerve. These cells vary slightly according to the animal used in the experiment. I believe that these are varieties of a single type of cell and am inclined to accept the suggestion, made by Cox, and Warrington and Griffith, that they represent the cells of Dogiel's Type II (new types III, IV, VIIT, XI).

There does not seem to be any adequate reason for a further classification of the large cells so far as the arrangement of the

134 journal of Comparative Neurology and Psyshology.

chromatic granules is concerned. The descriptions of Lugaro give an idea of some of the various pictures that may be presented.

ClIROMATOLYSIS.

A general review of the literature on chromatoljsis would have but little value for us. The observations of the previous investigators will be discussed in connection with the particular problems as these are taken up. It will, however, be necessary to present at this point the procedure adopted in the more important investigations, as a basis for the comparison of results in the subsequent pages. Lugaro ('96) cut the sciatic nerve in dogs at the level of the hip joint, and also resected the brachial plexus in both dogs and rabbits. The animals were allowed to live for periods varying form 2 to 240 days. Cox ('98) resected the brachial plexus in rabbits and allowed them to live for a period varying from one day to a year. Cassirer ('99) removed a piece of the sciatic nerve at the point of its exit from the pelvis, in rabbits which he killed 5 to 63 days after the operation. Koster ('03) also resected the sciatic nerve immediately after its exit from the pelvis in cats, dogs, and rabbits. The animals lived from a iew days to a year after the operation. All these investigators prepared the spinal ganglia associated with the injured nerves by some modification of Nissl's method, most often staining with tolui din-blue. By the study of ganglia removed at different periods after the operation they have been able to follow the changes in the ganglion cells through the various phases of chromatolysis. The following general statement concerning these various phases is necessary as a preface to a discussion of the problems which each presents.

Somewhere from 1 to 4 days after a nerve has been divided changes become noticeable in the cells of the associated spinal ganglion. There occurs a progressive solution of the tigroid substance beginning either near the nucleus or at a point intermediate between the nucleus and the periphery of the cell. The cell becomes swollen and the nucleus more or less displaced toward the periphery. These changes characterize what will be called the ])hase of reaction. After a time which varies within wide limits according to the con

R ANSON, Ahcrati()}is in Spinal Gdiiglioii Cells. 135

ditions of tlic experiment the secondary or consecutive alterations make tlieir appearance. These are of three kinds: (1) repair, which occurs in a considerable proportion of the cells and consists of a restoration of the Nissl bodies and a return of the nucleus to the center, of the cell ; ( 2 ) atrophy, which occurs in nearly all the cells that undergo repair and results in a very considerable shrinkage of the cell body and its nucleus; (3) progressive degeneration leading to the complete destruction of all those cells Avhich fail to undergo repair.

The phase of reaction has been most carefully studied and all the essential features of the intra-cellular changes have been described many times. Accordingly, we shall have chiefly to consider at this time the relative susceptibility of the different types of cells, and pay but little attention to the finer details of chromatolysis. The measurements of Lugaro and Fleming supply us with satisfactory data as to the atrophy taking place in the cells ; but the observations on the phases of repair and degeneration are of the meagerest sort. It is with regard to these two late phases that the present investigation has given the most suggestive results.

The Phase of Reaction. Due in part, perhaps, to the peculiarities of the type of animal but more to the immaturity of the individuals used for the experiments, chromatolysis occurs very early in the second cervical ganglion of the young white rat. Five days after an operation performed on a ]*at 12 days old chromatolysis is already far advanced and is, in fact, at its highest point. Even in the largest cells there remains but a narrow peripheral ring of undissolved Nissl-granules. The tumefaction of the cells and the peripheral dislocation of the nuclei are as pronounced on the fifth day as at any subsequent time (Fig. ?>).

All authors, Lugaro, Fleming, Cox, Cassirer, Koster, and Warrington and Griffith, agree that the vast majority of the cells in the spinal ganglion react to a lesion of the peripheral nerve. There is some difference of opinion as to the exact proportion ; while Koster asserts that all the cells react, the remaining authors make an exception of the very smallest cells in the ganglion, and Cox, Warrington and Griffith also find a small percentage of the large cells that do

136 'Journal of Comparative Neurology and Psychology.

not react. Making allowance for these differences, we may saj that it is a fact, accepted by all who have studied this question, that somewhere from 85 to 100 per cent of the spinal ganglion cells show chromatolvsis after the section of the peripheral nerve close to the ganglion. The preparations of the ganglia of the rat also show the greater number of cells in the various stages of reaction. The extensive chromatolvsis, which seems to be established beyond any possibility of doubt, is very difficult to harmonize with the results obtained by the numerical investigation of the architecture of the ganglia, results which are also beyond cavil.

It has been clearly demonstrated by many observers, Hodge, Biihler, Lewin, Hardesty, Ilatai, and Eanson, that there are several times as many cells in the spinal ganglion as there are medullated afferent fibers. This well established fact has been discussed in detail in a previous j^aper and we do not need to consider it here. It is very interesting, however, to note that we have here two well established facts which seem contradictory to each other. Whereas at least 85 per cent of the cells show what has been interpreted by all observers as a typical axonal reaction after section of the peripheral nerve, the numerical results show that on the average only 33 per cent of the cells are connected with the medullated fibers in the nerve. In ^vhat way are we to explain the axonal reaction in the other 52 per cent of the cells ? Lugaro and Koster, who had not themselves worked with the numerical method, felt justified in doubting the correctness of the numerical results, because they conflicted with facts that they knew to l)e correct. But having worked with both methods and being convinced of the correctness of both sets of results, I cannot so lightly set aside either in order to avoid the dilemma. There seem to be but two alternatives : either the reaction in this 52 per cent of the cells is not an axonal reaction at all but is secondary to some intra-ganglionic disturbance, or there must be a very large number of non-medullated fibers in the nerve. It has been shown (Hanson '08) that the histology of the ganglion would not exclude the possibility of the first alternative. The spinal ganglion is not to be regarded as an aggregation of more or less spherical cells each independent of the others, and

R ANSON, Altcratiuiis ill Spinal Ganglion (^rlls. 137

connected only with its central and peripheral processes ; but is in reality a complicated mass containing the ramifications of dendrites and axis cylinders, forming exceedingly intricate intercellular meshworks and pericellnlar baskets, the cells in this w^ay being brought into close functional relation with each other." We will now study the reaction in the different tyjies of cells to see if there is any evidence that the reaction in some cells is secondary to that in others.

It has been shown in a previous paper that the evidence points to the small cells as those not associated with medullated fibers, and it is, therefore, in these cells that w^e would expect to find evidence of the secondary nature of the chromatolysis. ISTevertheless, the evidence shows that the small cells are the ones most susceptible to a lesion of the nerve. According to all observers wdio have made any statement in this connection, the small cells are the first cells in the ganglion to react. This fact is entirely at variance with the idea that their reaction is secondary to an axonal reaction in the large cells.

According to Lugaro the small dark cells are rapidly altered, the reaction reaching its height by the fourth day, while the reaction in the other cells does not reach its maximum until the fifteenth day. The cytoplasm of the small cells is pale, especially at the center, and the nucleus has been displaced to the periphery. Cox found that the small cells showed alterations as early as twentyfour hours after, the operation and by the end of the fourth day most of them were very much altered, wdiile the large cells were just beginning to show chromatolysis. All the other observers agree that the small dark cells show chromatolysis but do not say at w^hat time the reaction occurs. There can, therefore, be no doubt that most of the small dark cells react to the injury of the nerve and do so earlier than the large cells. It must be borne in mind, however, that not all of the small cells react in this way. As will be remembered, a few of the smallest cells do not present a dark cytoplasm, but are clear cells w^th a few large chromatic granules. Lugaro, Warrington and Griffith agree that these cells never show chromatolysis. According to the last two observers there are also a few of the

138 Journal of Coiiiparafive Neurology and Psychology.

smallest dark cells tliat fail to react. In the cat no cells nnder 25 microns, dark or clear, showed chromatolysis. They estimate these non-reacti\'e cells as constituting from 7 to 13 per cent of all the cells, according to the ganglion studied.

In my preparations from the rat, which came to autopsy five days after the operation, the chromatolysis is at its highest point and the majority of the cells are greatly altered. It is therefore not possible to determine whether the large or small cells were first altered. It is in the small cells, however, that the most extreme alterations are found. Fig. 3 illustrates some of these changes. The nucleus is strikingly excentric, in most of the cases it causes a distinct bvilging of the cell-outlines, and in many it appears to be indenting the cell from without. The chromatic substance is completely dissolved except for a dense ring which persists at the perijihery of the cell and a small clump sometimes found near the nucleus. Even at this stage, five days after the operation, it is clear that some of these small dark cells have disintegrated, but of this more will be said in connection Avith the phase of degeneration.

From what has ])een said it will be obvious that the changes wliich appear in the small cells are characteristic of axonal reaction, and there is nothing in the finer details of the chromatolysis to indicate that it is due to any other cause than a lesion of the axons of these cells. If we are to continue to place any confidence in the conclusions based on Nissl's axon-reaction; we are forced from these facts to admit that the majority of the small cells possess axons in the peripheral nerve, the numerical results to the contrary notwithstanding.

Concerning the phase of reaction in the large cells there is very little to add to that which has been described by previous observers. We have already mentioned the facts concerning the non-reacting coarsely granular cells. There is but one other point of interest in the chromatolysis of the large cells, namely, the absence of any trace of clumping of the tigroid masses about the nucleus in the preparations of the reacting ganglia of the rat. Fleming, Cox, and Kleist have each seen and described this phenomenon; and there

R ANSON, Altcr(iti()}is III spinal Gaiighon Cells. 139

can be little doubt about the correctness of these earlier observations, since the figures and descriptions given by these authors are very clear, and agree in all essential points with each other. In some of the illustrations it seems as if almost the entire quantity of chromatic substance is accumulated in one solid mass which more or less completely encircles the nucleus. Why this appearance was not to be found in a single cell in my preparations is hard to understand. It may be due to the fact that my animals were of a different kind, but more probably is to be explained on the basis of the rapidity with which the reaction occurred in my specimens. It seems probable that the chromatolysis was so rapid that there was no time for the clumping of the granules to occur before they were dissolved. Or it may be that clumping occurred in the earlier stages but gave place before the fifth day to complete chromatolysis.

The Phase of Degeneraiion. Very little attention has been paid to the degenerative changes which lead to the ultimate disappearance of the cells. Lugaro found that scattered cells underwent complete degeneration and became surrounded and penetrated by capsular nuclei. This occurred from 15 to 40 days after the injury. Vacuolar degeneration seemed to him to play a small part in the cell destruction. Fleming speaks of a "disintegration of the protoplasm" that occurred in many cells 6 to 18 Aveeks after the operation. Cox gives an excellent description of the degeneration by vacuolation. According to him the vacuoles are sometimes small, sometimes large and multiple and in the latter case the l^issl-bodies have almost entirely disappeared from the cell. If the vacuoles are very large the nucleus vanishes, so that the cavity which represents the remains of the earlier cell contains only the membranous ]iartitions between the vacuoles. These membranes are impregnated with little granules. In addition to this form of degeneration Kleist mentions another, namely the gradual progressive atrophy of the cell resulting at last in its complete disappearance. Koster is of the opinion that the vacuolation is not the result of a pathological process but is due to an error in technique. The cell destruction according to him does not occur until late, mostly after the 284th day. One sees then violet-stained protoplasmic remains which contain only

140 'journal of Cojjiparative Neurology and Psychology.

degenerate nuclei or none at all, and whieb are surrounded bj proliferating connective tissue cells of tbe capsule.

It is quite remarkable, considering bow large a number of cells disappear from tbe ganglion (52 per cent according to counts on tbe second cervical ganglion of tbe wbite rat), tbat tbe strictly degenerative cbanges are not more obvious in sections of tbe reacting ganglia. One reason wby my preparations bave not given altogetber satisfactory data on tbis point is tbat most of tbem were not counterstained. Wbile tbis gives tbe best material for study of cbromatolysis, it does not bring out tbe protoplasmic cbanges wbicb indicate tbe disintegration of tbe cell. In tbe preparations of ganglia removed seven and eigbt days after tbe operation, wbicb were counterstained, a small percentage of tbe cells can be seen undergoing deiinite disintegrative cbanges. Two well-marked types of disintegration can be recognized. By far tbe most frequent and im]iortant is tbat in wbicb tbe cell becomes penetrated by proliferating fibroblasts from tbe capsule wbicb form nests of vesicular nuclei suggestive of tbe intracapsular proliferation in rabies. Tbe early stages in sucb a process are sbown in Fig. 5. Tbe small dark capsular nuclei bave given place to large vesicular ones. Tbese are seen grouped about tbe cell. Five bave penetrated into tbe cell. Tbere is no trace of tbe ISTissl-bodies, tbe nucleus presents an indefinite crenated border and contains tbe outlines of a barely recognizable nucleolus. A furtber increase in tbe number of tbe fibroblasts gives rise to tbe formation of nests of sucb cells in wbicb one can just recognize tbe outlines of tbe original spinal ganglion cell. Sucb nests are relatively frequent after tbe seventb day. By tbe transformation of tbese into adult connective tissue witb tbe consequent sbrinkage, tbe spaces formerly occupied by tbe ganglion cells are obliterated so tbat tbe ganglion cells tbat survive are brougbt to lie almost as close togetber •IS in tbe normal ganglion. Tbe amount of intervening connective tissue is surprisingly little after two montbs.

Tbe otber form of degeneration, vacuolation, is more rare but seems to play a certain part in tbe cell destruction. It may go on to sucb an extent tbat tbe cell is converted into a single large vacuole, surrounded by a tbin ring of protoplasm (see Fig. 6). It is true

Ranson, Alterations in Spmal Ganglion Cells. 141

that one sometimes meets Avitli vacuolation in the cells from normal ganglia ; but the process does not reach the same extent nor is it nearly as freqnent as in the ^'operatcnr' ganglia.

While this is all^that can he said positively concerning the disintegrative processes, it seems probable that many of the small cells become so ninch swollen that the nucleus is extruded, after which the cell rapidly disintegrates. Such a process as this cannot of course be observed, but it may be supposed to occur, because of the extreme peripheral position of the nucleus in the cells (see Fig. 3). The nuclei often are so placed that they appear half outside the cell, and in some more marked cases as if the nuclcTis were a separate structure indenting the cell-body from without.

In connection with the degenerative changes two other problems demand attention. Is there any special })art of the ganglion in which the cells seem i^articularly susceiDtible to degeneration after section of the peripheral nerve ? Is there any particular type of cell which is more likely to disappear than the others i The first question is suggested by observations of Bumm, who worked with the second cervical nerve of the cat, and of Kleist working with the same nerve in cats and rabbits, both of whom found after section of the dorsal root a region on the posterior aspect of the proxinuil part of the ganglion in which the degenerative changes were much more marked than in the rest of the ganglion. Both Bumm and Kleist believe that there are situated here cells which are associated with a fiber in the dorsal root but not with one in the nerve. They consider these as relay neurones inserted between the sympathetic and the central nervous system. It is to be borne in mind, however, that these authors base their conclusion purely on the effect of cutting the dorsal roots, and that they did not attem])t to determine whether these supposed cells could be demonstrated l)y their failure to react when the peripheral nerve was cut. The descriptions and illustrations given in their papers indicate that the destruction in this dorso-proximal quarter of the ganglion involves a large part of the cells and give the impression that these cells must be quite numerous. N^ow if the idea that these cells possess no fiber in the peripheral nerve is correct, they should not react to a section of the nerve and

142 ^Journal of Comparative Neurology and Psychology.

one would be justified in expecting to find a considerable number of unaltered cells in this region of the ganglion. It does not seem probable that such a condition could have been overlooked by all those who have studied chromatolysis in the sjunal ganglion after section of the peripheral nerve; and, having Bumm's observation in mind, I have again gone carefully through my preparations, but have been unable to see that this dorso-proximal region of the ganglion showed any greater number of unaltered cells than are to be found in any other part of the ganglion. This fact argues strongly against the assumption of Bumm and Kleist that the dorso-proximal part of the ganglion is the locus of cells which send a fiber into the dorsal root but none into the nerve. Just why the cells in the dorso-proximal part of the ganglion should be more susceptible to a lesion of the dorsal roots than the other cells of the ganglion is hard to say. It should, however, be borne in mind that it is just this portion of the ganglion that woidd be most ex])osed to direct trauma in an operation on the dorsal roots, and that it would also be most affected by any anaemia produced by the division of the arteriole accompanying the dorsal root. It is not probable, however, that these are determining factors, and the observations of Bumm and Kleist may still have an important but as yet undetermined significance.

It has been shown that there is no special area which is most affected, but there still remains the question whether any particular type of cell is more susceptible than another. One of the clearest observations that can be made on the material from the white rat, is that the cells which disappear belong for the most part to the small cell type. It is true that it is very difficult to follow the different steps in the disintegration of these cells, and it would be impossible to reach this conclusion by direct observation of the process of degeneration. The only hint in this direction which such a study gives is the fact that the nuclei of these cells are much more excentric than those of the large cells, so much so that the nucleus often appears as a second sphere attached to one side of the cell body. It is clear, however, that as the length of the post-operative period increases, the number of cells diminishes, spaces are left filled with

Ranson, Alterations in Spinal Ganglion Cells. 143

very loose cellular connective tissue, and the cells which remain are predominantly large cells. As time goes on the connective tissue contracts, the spaces are obliterated, and the large cells, no longer separated by so many small cells, come to lie much closer together than in the normal ganglion. This condition is very obvious by the 20th day, at which time the degeneration is almost complete and the majority of the surviving cells have returned to their normal appearance. Figs. 1 and 2 are representative areas from the control and operated" ganglia of a rat which survived 20 days. The large cells have undergone some atrophy. The disappearance of the small cells and the approximation of the large ones is obvious, yet the contrast is less marked in the illustration than that which one finds in following through a series of sections of these two ganglia.

Perhaps a more satisfactory method of showing this relation is by means of a differential count. A difficulty in the way of such a count is that size is not the only point of differentiation between the two types of cells. For, as has been said, there are cells of medium size which because of other characteristics belong with the large cells, and other medium sized cells which belong with the small ones. On this account it is not possible to make an altogether satisfactory differential count on the basis of size alone. ISTevertheless, it is only by the use of such a rigid objective criterion that one can rule out all possibility of personal bias. Accordingly, in making the count a mean diameter of 20 microns was accepted as an arbitrary dividing line between the two types. The majority of the large cells have a greater diameter than this, the majority of the small cells less. In making this count use was made of an ocular net micrometer ruled in squares in such a way that with the combination of the Zeiss oil-immersion lens and 12 eye-piece the sides of these squares corres^wnded to the chosen diameter. In this way it was possible rapidly to measure each cell and determine to which class it belonged. Two counting machines w^ere used, the large cells being registered on one, the small cells on the other. Four sections from each ganglion were thus subjected to a differential count. These sections were taken at random from different parts of the ganglia.

144 'Journal of Comparative Neurology and Psychology.

Table II shows that the decrease in the number of cells is due chiefly to the loss of cells under 20 microns in diameter. There is also no doubt a slight loss of cells of the larger size, but this is compensated by the fact that the remaining large cells lie closer together, so that a single section shows about the same number as there are in a similar section of a normal ganglion. But the point to be emphasized is that the cell destruction affects chiefly the smaller cells. That this fact escaped the attention of all the previous investigators is probably due to the slow repair of the large cells in their specimens, which permitted a considerable amount of atrophy to occur before the cells again became of normal appearance.

TABLE II. Showing the proportion of large and small cells in the normal and "operated" second cervical ganglion of the white rat 20 days after section of the ramus posterior of the corresponding nerve.

Norm

\L,

Operated

Large cells

Small cells

Large cells

Small cells

32

48

28

16

32

35

36

24

26

47

21

18

28

38

31

20

Sum 118

168

116

78

Average . 29.5

42

29

19.5

It is necessary in making such a differential count to make it at just the proper period after the operation. The careful measurements of Lugaro and Fleming have shown that there is during the phase of active chromatolysis a considerable swelling of the cells, and that repair is accompanied by marked atrophy which continues to progress for some time and leaves the cells very nmch reduced in size. It is clear that cither swelling or atrophy would tend to obscure the relations brought out in Table II. In making the counts it was necessary to select a period when the repair of the cells was just complete and tlieir a?dematous condition had subsided, but before they had begun to show marked atrophy. For this purpose the preparations from the rat killed 20 days after the operation

Ranson, Alterations in Spinal Ganglion Cells. 145

were \evy well suited. At the end of two months so mucli atrophy has occurred that the relations given in Table II can no longer be made out.

These observations bear out the conclusion reached in a previous paragraph that the reaction in the small cells is not secondary to that in the large. If it were, it would be difficult to see why the large ones should survive and the small ones disintegrate. These small cells which do not have medullated axons in the nerve are the first to react, show a typical axoual reaction, resulting in very extensive cell destruction. These facts are very important for a proper understanding of the architecture of the spinal ganglion, but their bearing on that subject can best be discussed in another place.

Tlie Phase of Be pair. We come now to the study of the reparative processes by which the surviving ganglion cells regain their normal appearance. For this study the preparations -from the young rats proved especially fitted. Because of the rapidity with which the ganglia passed through the various phases, the repair came very much earlier in these specimens than in any previously described. Repair began on the eighth or ninth day and was almost complete in 20 days. During this interval specimens were taken every two or three days and a complete series representing the reparative changes was thus obtained. In the preparations of other investigators taken from the ganglia of other, and older, animals the repair occurred late when the specimens of the series were taken weeks apart.

Nine days after the operation the nuclei of the large cells begin to recede from their peripheral position and come more and more in the course of the next few days to occupy the center of the cells. There is little change as yet in the character of the chromatic granules. In most of the cells the stainable substance forms only a narrow ring about the periphery of the cell ; the remainder of the cytoplasm stains a difi^use light blue. Twelve days after the operation only a few of the large cells show peripheral nuclei and there is a very noticeable augmentation in the quantity of the tigroid masses. As this accumulates in the cell it disposes itself in two ways. A part of it goes to increase the breadth and

146 "Journal of Comparative Neurology and Psychology.

density of the jDerijjheral ring that has resisted solution during the earlier phases of chromatolysis. The larger part, however, is distributed in the form of very fine granules through the remainder of the protoplasm. At this stage the majority of the large cells correspond quite well to Lugaro's description of the cells of his second type. They are large and medium-sized clear cells with very fine tigroid masses which are larger at the periphery of the cell. They do not, however, long conform to this type, since the chromatic substance is rapidly increasing in quantity and is laid down in the form of large granules scattered uniformly throughout the protoplasm. By the seventeenth day the central portions of the cells contain Nissl-bodies of considerable size, and by the twentieth day the distinction between the coarsely granular peripheral ring and the rest of the cell has disajDpeared. The cells present a uniformly coarsely granular appearance. In this way we are able to trace the large cells from the height of chromatolysis through the stages of gradually increasing chromatic substance, until they present the pylcnomorphous appearance characteristic of the large cells 20 days after the operation (Figs. 2 and 4).

The importance of thus following the transformation is two-fold. In the first place, it shows that the large cells undergo repair; and in the second place it shows that the large cells present in the 20day preparations and counted as such in the enumerations given in Table II are in reality the same large cells as were originally present in the ganglion, and cannot l)e regarded as hypertrophied small cells. In this way we can be sure that the conclusions derived . from Table II are not misleading, namely, that it is the small rather than the large cells which undergo complete degeneration.

Conclusions.

The more important results of the investigation may be summarized in the form of answers to the problems suggested in the introduction.

1. It is ]n'obable that the medium sized coarsely granular cells which fail to react to a lesion of the nerve close to the dorsal root ganglion, belong to neurones which send no axon into the nerve (Dogiel's Type II, or new types III, IV, VIII and XI).

Ranson, Alterations in Spina! Ganglion Cells. 147

2. While only a part of the cells (52 per cent) undergo complete degeneration, nearly all of the cells (at least 85 per cent) show chromatolysis after an operation severing all branches of the nerve; and this occurs in spite of the fact that not more than 33 per cent of the spinal ganglion cells are associated with medullated fibers that would be injured in cutting the nerve.

3. There is no indication that the reaction in the small cells is secondary to some intra-ganglionic disturbance; it possesses all the characteristics of a true axonal reaction and occurs at least as early, and according to Cox and Lugaro somewhat earlier, than the reaction in the large cells.

4. The . process of degeneration varies in different cells. In some cases the invasion of the degenerating ganglion cells by proliferating fibroblasts is the most striking feature, in others vacuolation appears to be the cause of the cell disintegration. In the small cells it is probable that the extremely excentric nuclei may, as the turgescence increases, be finally extruded.

5. There does not seem to be any part in which the reaction is more marked than in the rest of the ganglion, and the dorsoproximal portion shows as many altered cells as are to be found elsewhere. This fact may be taken as an argument against the assumption of Bumm that this portion of the ganglion is the locus of cells which do not possess a fiber in the peripheral nerve.

6. The large cells of the ganglion can be followed through the various stages of repair until they again present central nuclei and are filled with a very large amount of coarsely granular chromatic substance.

7. It has been shown, moreover, that it is the small neurones which are most seriously injured, since it is from among this group of elements that the loss of cells chiefly occurs. Since the small cells show typical axonal reaction and degenerate in by far, the greatest number, we are forced to conclude that they possess axons in the peripheral nerves, and that their non-medullated processes traced by Dogiel as far as the junction of the ventral and dorsal root, extend beyond that point into the nerve.

This conclusion has recently been confirmed by the demonstration

148 Journal of Comparative Neurology and Psychology.

of very large numbers of non-meclullated fibers in the spinal iierve.s of the rabbit by the method of Cajal. These non-medullated fibers, an account of which will appear in another paper, allow us to cxplaiu with ease the following otherwise inexplicable facts :

1. The axonal reaction in the small cells which are not connected with medullated fibers in the nerve.

2. The results of Lugaro, Fleming, Cox, Cassirer, Koster, Warrington and Griffith, and myself which show that the vast majority of all the cells in the spinal ganglion react to a lesion of the peripheral nerve although only a small part of these cells are connected Avith medullated fibers in the nerve.

3. The degeneration of 4412 ganglion cells after the section of a nerve containing 1500 medullated fibers. These 4412 cells were chiefly small cells and were associated with non-medullated fibers in the nerve.

4. The intact condition or only slight degeneration of the dorsal roots after the degeneration of 52 per cent of ganglion cells, the average loss in the dorsal root being 17 per cent. This is easily understood when it is remembered that the large cells which alone are associated with medullated fibers in the dorsal root pass rapidly through the phases of reaction and repair to complete restitution.

BIBLIOGRAPHY.

BUMM, A.

'0,3. Die pxiun-imeutelle Diu-ebtvi'inuin;,' dor vordern und liintern Wurzel des zweiten Halsnerveii bei der Katze luid ilire Atrophiewirkuug avif das zweite spinale Halsgauglion. BHz. Bcr. Gcs. MorpJi. Physiol.. Miinclieii, B. 18, p. G.5.

Cox, W. II.

'98. Der feinere Ban der Spinalgauglienzelleii des Kanincliens. Aiutf.

Hcfte, Abtb. 1, Bd. 10, Heft 31, p. 73. '98a. Beitriige zur patbologischen Histologie luid Pbysiologie der

Ganglienzellen. Internat. MoiKitsscJir. fiir Aiiat. und Pliys.,

Bd. 1.5, p. 241.

Cassirer. IX.

'99. Ueber Yeriinderungen der Spinalganglienzelleu nnd ibrer centralen Fortsiitze nach Dnrebscbneidung der zugeborigeii peripberen Nerveii. Deutsche Zeitxchr. fiir Nervciihcil., Bd. 14. p. 151.

Ranson, Alterations in Spinal Ganglion Cells. 149

Fleming, R.

'07. The effect of ascending degeneration on tlie nerve cells of the ganglia, and the posterior nerve roots, and the anterior cornua of the cord. Edtnhurrih, Med. Jour., vol. 43, p. 270.

IIATAI, S.

"00. The finer structure of the si)inal ganglion cells in the white rat. Jour. Conii). Kcur., vol. 11, p. 1.

Kleist, K.

'04. F].\perinjentell-anatomische Untersm'hnngen iiher die Beziehnn geu der hiutereu Riickenniarkswurzeln zu den Spina Igan glien. Yirchoii-'s ArcJiiv, Bd. 175, p. 381. 'Go. Die Veranderuug der Spinalganglienzellen nach der Durcli schneidnng des peripherischen Nerven nnd der liinteren

^^'nrzel. Yirchov's Archie, Bd. 173, p. 4GG.

KUSTER. G.

'03. Ueher die verschiedene biologische Werthigkeit der hinteren Wurzel nnd des sensihlen peripheren Nerven. Neurol. CenIralhJ., Bd. 22, p. 1003.

LuCiARDO, 10.

'04. On the pathology of the cells of the sensory ganglia. A'/r. dl Fatal, nerv. c incnt., vol. 5, nos. 4, G, 0; vol. G, no. 10; vol. 7, no. 3; vol. 8, no. 11; Ref. Rev. of Neurol, and Psych., vol. 2, p. 228.

Warrington, W. B., and Griffith, F.

'04. On the cells of the spinal ganglia and on the relationship of their histological strnctnre to the axonal distribution. Brain, vol. 27, p. 207.

Ranson. S. AV.

'0(i. Retrograde degeneration in the spinal nerves. Jour. Comi).

Neur. and Psych., vol. IG. "OS. The architectural relation of the afferent elements entering

into the formation of the spinal nerves. Jow. Comp Neur.

and Psych., vol. IS.

150 journal of Comparative Neurology and Psychology.

Fig. 1. Zeiss, Ocular Ii, Objective 8. — Drawing traced from photo-micrograph of a transverse section through the control second cervical ganglion of a young white rat twenty days after the operation, showing the characteristics of the normal ganglion. In the center is a clear area representing the dorsal root fibers. Note the size of the large cells, also the large number of small cells.

Fig. 2. Zeiss, Ocular 4. Objective 8. — Drawing traced from a photo-micrograph of a transverse section through the "operated" second cervical ganglion of a young rat twenty days after the operation. This section illustrates the alterations which have occurred in the ganglion as a result of division of the nerve. It can be readily seen by comparing Figs. 1 and 2 that there have occurred both an atrophy of the ganglion as a whole and a decrease in the number of the cells. It is apparent at a glance that the cells in the operated ganglion are predominately of the medium size. None are as large as the largest cells seen in the normal ganglion, due to the fact that the cells have already begun to show some atrophy. The most striking feature is the loss of the small cells.

R ANSON, Alterations in Spinal Ganglion Cells. 15 1

Fig. 1.

Fig. 2.

152 "Journal of Covi^arafive Neurology and Psychology.

Fig. 3. Zeiss, Ocular If, Objective 1/12. — Drawing of a small area from a transverse section through the "operated" second cervical ganglion of a young white rat five days after the operation. With the exception of one mediumsized cell, all of the cells show more or less extensive chromatolysis. The medium sized non-reacting cell is distinctively of the coarsely granular type. The usual features of chromatolysis can be seen in the reacting cells. Notice the extreme peripheral position of the nuclei of the small cells.

Fig. 4. Zeiss, Ocular 4, Objective 1/12 — Drawings of a small area of a section through the same ganglion as that represented in Fig. 2. It represents, the condition in the ganglion twenty days after the section of the nerve. The most striking feature is that the large cells have almost regained their normal appearance. The nuclei are centric and there is a large amount of chromatic granules distributed in the normal manner throughout the protoplasm.

Fig. 5. — Showing a degenerated cell penetrated by proliferating cells from the capsule, from the second cervical ganglion of a rat eight days after the operation.

Fig. G. — Showing a cell distended by a large vacuole, from the second cervical ganglion of a rat seven days after the operation.

R ANSON, Alterations in Spinal Ganglion Cells. 153

■J

Fig. 3.

% .^'■"^^^^^

'. ' y t"^^

-"**.

'MM'.^^y

Ha^haifititHDi 'o.

Fig. 4.

Fig. 5.

Fig. 6.

The Journal of

Comparative Neurology and Psychology

Volume XIX May, 1909 Number 2

Ol!^ THE RELATIO:^ OF THE BODY LENGTH TO THE

BODY WEIGHT AND TO THE WEIGHT OF THE

BRAIN AND OF THE SPINAL CORD IN THE

ALBINO RAT (MIIS NORVEGICUS

VAR. ALBUS).

HENRY H. DONALDSON.

Profcsfior of Nci(rolof/ij at The Wi!<f<ir Institute.

With Three Figures.

In a recent paper (Donaldson, '08) the relations of the body weight to the weight of the brain and of the spinal cord in the albino rat have been described.

In addition to the determination of the body weight it was stated in the paper jnst cited (pp. 340-7) that measnrements had also been made on the body length (trnnk and head) of some of the rats, bnt to avoid confnsion the discussion of this character and its relations was reserved for the present paper.

The reasons for making a series of linear measurements on the albino rat were briefly the following: —

1. To obtain a second general measure of the body growth of the albino rat in terms other than those of weight.

2. To gather data by which to determine the body weight and body length ratio for the variety measured.

This ratio is valuable because it gives a notion of the general shape of the animal and also enables us to state whether there are

The Journal of Comparative NErROLOGY and Psychology. — Vol. XIX. No. 2.

156 'Journal of Comparative Neurology and Psychology.

differences in this relation according to sex, as well as to make comparisons with other forms.

It also permits the determination of the influence of dwarfing and other modifying conditions on the weight-length relation.

f3. Both the weight of the brain and of the spinal cord can be related to the body length, and the measurement on body length thus made to furjiish an additional datum from which the weights of the brain and of the spinal cord can be inferred. As we shall see, this datum is a much better one than body weight, especially in those cases where, for one reason or another, the animal has become emaciated.

4. If we consider the body length of the rat to correspond in a general way with the sitting height in man, we have one more means of comparing the growth changes in the two forms.

In the following pages we shall discuss these points, so far as they have been worked out. For the mathematical treatment of the results I am indebted to my colleague, Dr. Hatai, who is publishing at this same time some notes on the formulas previously used by both of us (Hatai, '08; Donaldson, '08), as well as giving a new and more general formula for determining the weight of the brain from the body weight (Hatai, '09).

The technique of weighing and measuring was that described in the earlier paper (Donaldson, '08). A number of complete records on the albino rat have been added to those on hand at that time. Moreover, for the relation of body weight to the body length alone, additional records have been obtained by weighing and measuring animals which had been anesthetized lightly.

It was my first intention to print the full series of individual records (233 males, 173 females) in a general table at the end of this paper. I have, however, decided not to do so for the following reasons : —

First. — Printing such a general table would involve repeating a number of the records already published in a former paper (Donaldson, '08), and would in turn need to be again repeated in a forthcoming paper on the change in the percentage of water during the growth of the nervous svstem.

BRAIX AND SPINAL CORD OF RAT.

240 220 200 180 160 140 120 100 80 60 40 20

1

BODY LENGTH

1

1 1

1

1

. * .1^

— ^

.'. ...^^^-r-^-'^^'^^^

^T— "^^

/ • . _:

. ■' .'-r^f^V-CT^' * '"'

,

. - .ii^sSiST '

• , ' ' .

^ •••n-^

J-* — •• •

..>-'^ • ■

~

•^^Kj* •

• .^^--^

'^, ■■ (^

•.:">^"

..^l'-! ■

IJ*"

<(-^.

./■

.%

- f

- 4

- 4

1 1 1 1 1 1

1 1 1

1 1 1 1

BODY WEIGHT 1 1 1 1 1 1 1 1 1

II 1 1 1 1 1

i 1 1

Gms.

20

60

200

80 100 120 140 160 180

To show in the ablno rat, the body length in millimeters according to body weight in grams. Records for 170 males «, and 148 females X. The theoretic curve for the

is based on formula (4). iHE JonnNAi, or Compaeative Nedeolooy and Pstcholoqi. — Vol. XIX, No. 2.

300 320

sexes combined

156 'Journal of Comparative Neurology and Psychology.

Donaldson, Brain and Spinal Cord of Rat. 157

Second. — The individual records have been tabulated and are on file at the Institute. They are therefore available for use by other investigators, and may be had by application to the Director of The Wistar Institute.

Third. — It is hoped that this condition will be only temporary, and that when this group of investigations is completed, the entire series of individual records employed for them can be printed in the form of tables in a special brochure, thus making them generally available. At this time only the mean values of the observation are tabulated.

We turn at once, therefore, to the consideration of the special questions : —

1. The body length of the albino rat according to body weight.

On Chart I, so far as is possible without confusion, the individual records for body length (170 males and 148 females) are entered according to the body weight. The continuous line on the chart shows the theoretical curve. As can be seen, the distribution of the records is such as to fit a theoretical curve that rises with diminishing rapidity, and so far as it can be plotted, is still bending towards the horizontal. A distinction between the sexes in the relation of body length to body weight, though present, is hardly to be seen on Chart I. The mean values for the body lengths are given in Table 2. Making use of these data, the weight length ratios have been determined for the series in hand.

Table 1 gives the numerical expression of the relations obtained by dividing the calculated body length (for both sexes combined, see Table 2 ) by the body weight.

The ratios thus obtained are given in Table 1, and these show that the albino rat becomes relatively shorter as its weight increases.

By means of a correlation table based on groups differing by 10 grams in body weight and 10 mm. in body length, the mean statures for given body weights have been calculated. This has been done for each sex separately, as well as for both sexes taken together, and the final values obtained are given in Table 2.

When the means for the males are compared with those for the

1^8 'Journal of Comparative Neurology and Psychology.

females (see Chart II, based on 179 males and 160 females) it will be observed that the latter rnn slightly below the former. The difference, thongh small, has significance, as we shall show later. However, for the general discnssion at this time the resnlts are not separated according to sex, but are treated together.

TABLE 1.

The Ratios Obtained by Dividing the Body Length by the Weight in

THE Case of Mus Norvegicus Var. Albus.

Body length

Body weight

mm.

Ratios.

gms.

Both sexes combined (See Table 2.)

5

51.9

10.. 38

15

77.6

5.17

25

94.8

3.78

35

109.1

3.11

45

120.5

2.67

55

130.6

2.37

65

137.7

2.11

75

144.9

1.93

85

152.0

1.78

95

157.7

1.66

105

163.4

1.55

115

167.7

1.45

125

173.5

1.38

135

177.7

1.31

145

180.6

1.24

155

184.9

1.19

165

189.2

1.14

175

192.0

1.09

185

194.9

1.05

195

197.8

1.01

205

200.6

.97

215

203.5

.94

225

206.3

.91

235

209.2

.89

245

210.6

.85

255

213.5

.83

265

216.4

.81

275

217.8

.79

285

220.6

.77

295

222.1

.75

305

224.9

.73

315

226.4

.71

325

227.8

.70

The theoretical curve which most closely represents the change in body length with increasing body weight, is given by the formula (4)

y = 143 log (x + 15) — 134

where y represents the body length and x the body weight.

This is a formula of the same type as those used for determining the weight of the brain and of the spinal cord in relation to the

BBAIN AND SPINAL CORD OF RAT. HENRT H. DONA

mm.

240 220

BODY LENGTH

BODY WEIGHT

J I I I I I I I I

I I I I I

J L

G^s. 20 40 60 80 100 120 140 160 180 200

To show in the albino rat, the mean values for the body length according to body weight, sexes separated;

220 240 260 280 300 320

males, females. The theoretic curve is not drawn,

as it would confuse the other lines.

The JonnNAL op CoiipAniTivE Nbceolooy and Psychology. — Vol. XIX, No. 2.

Donaldson, Brain and Spinal Cord of Rat.

159

weight of the body, and the type has been already discussed in a previous paper (Donaldson, '08, p. 350).

In this connection, however, there are some points to be corrected and further discussed. The consideration of these points is taken up in a paper by Dr. Hatai which apj)ears at this time.

TABLE 2.

Mean l)()(ly leiij^tli :u-ci)rdiiig to body weight in Mus iiorvogicus var. albus. The body length as given in the last coUnnu has been calculated by the fomnila (4), y — 143 log (x + 15)— 134.

Body Length Observed.

Body

Body length in

Weight Gms.

Frequen

Mean

Frequen

Mean

Frequen

Mean

mm. calculated by formula (4)

cies.

mm.

cies.

mm.

cies.

mm.

M.

M.

F.

F.

M. + F.

M. + F.

5

12

59.2

12

58.3

24

58.8

51.9

15

15

76.3

24

75.4

39

75.9

77.6

25

11

97.7

8

96.3

19

97.0

94.8

35

8

108.8

S

105.0

16

106.9

109.1

45

5

121.0

U

121.4

16

121.2

120.5

55

7

130.7

13

125.0

20

127.9

130.6

65

9

134.0

5

131.0

14

132.5

137.7

75

9

141.6

5

139.0

14

140.3

144.9

85

8

152.5

4

147.5

12

150.0

1.52.0

95

6

156.6

10

154.0

16

155.3

157.7

105

6

165.0

9

159.4

15

162.2

163.4

115

12

166.7

9

165.0

21

165.8

167.7

125

7

172.1

15

171.0

22

171.6

173.5

135

9

176.1

4

175.0

13

175.6

177.7

145

5

181.0

4

180.0

9

180.5

180.6

155

9

184.0

6

185.0

15

184.5

184.9

165

7

188.0

183.3

13

185.7

189.2

175

4

192.5

2

190.0

6

191.3

192.0

185

5

193.0

3

188.3

8

190.7

194.9

195

2

195.0

2

195.0

197.8

205

7

200.7

2

195.0

9

197.9

200.6

215

4

202.5

4

202.5

203.5

225

2

205.0

2

205.0

206.3

235

209.2

245

2

210.0

2

210.0

210.6

255

3

215.0

3

215.0

213.5

265

2

220.0

2

220.0

216.4

275

1

215.0

1

215.0

217.8

285

220.6

295

1

205.0

1

205.0

222.1

305

224.9

315

226.4

325

1

225.0

1

225.0

227.8

The co-efficient of correlation between the body weight and body length, the records being grou})od as stated al)Ove, is found to be .90.

It is possible, therefore, to infer the body weight from the stature, and vice versa, provided the body weight is normal.

At the same time it is evident that body weight is much more open to fluctuations than is the body length, and therefore the body leuo'th is the better standard.

l6o Journal of Comparative Neurology and Psychology.

2. The relation of the weight of the brain and of the spinal cord to the body length.

We shall consider each division of the central nervous system separately.

(a) The relation of the weight of the brain to the body length.

When the data on brain weight are plotted according to the body length, we obtain the distribution of individual entries (196 males,

TABLE 3.

Calculated Brain Weights and Spinal Cord Weights According to Body

Length in Mus Norvegicus Var. Albus.

Data for Both Sexes Combined.

Body weight

Brain weight

Spinal cord weight

Body length

gms.

gms.

gms.

mm.

Calculated by

Calculated by

Calculated by

Formula (4).

Formula (8).

Formula (3).

50

4.5*

.204*

.031*

55

6.6

.409

.047

60

7.8

.522

.059

65

9.7

.660

.077

■ 70

11.7

.827

.088

75

13.9

.962

.106

80

16.6

1.065

.129

90

21.8

1.191

.159

100

28.2

1.288

.194

110

36.6

1.379

.235

120

44.7

1.442

.270

130

55.1

1.504

.305

140

67.4

1.561

.346

150

81.6

1.612

.381

160

102.4

1.675

.428

170

118.7

1.714

.463

180

142.0

1.760

.498

190

169.5

1.811

.539

200

201.3

1.851

.580

210

239.1

1.897

.621

220

283.6

1.942

.656

225

324.0

1.977

.691

Since the formulas do not allow of extrapolation toward the lower end of the curve, the averages of the observed values are here employed.

137 females as shown on Chart III. The difference betAveen the two sexes is slight, and in this instance therefore the data for both sexes will be treated together.

The theoretic ciir\'e which fits the means most closely has been obtained in the following manner: —

For the body lengths given in Table 3, the body weights were calculated by formula (4) transposed as follows: —

BRAIN AND SPINAL CORD OF RAT. HENRY H. DONALDSON.

'..'..-'•ir'i'h-tiA'^^^^^' '^ '

i I I L

BODY LENGTH

1 I 1

mm. 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200

To show In the albino rat, the brain weight and spinal cord weight according to body length. (1) Upper entries, brain. Individual records, 196 males «, 137 females X. curve is based on formula (8). (2) Lower entries, spinal cord, 189 males •, 137 females X. The theoretic curve is based on formula (3).

210 220

The theoretic

The Jocrnal of Compabativb NEDEOLoat and Pstcholoot. — Vol. XIX. No. 2

Donaldson, Brain and Spinal Cord of Rat. i6i

y + 134

143

10 - 15 (4')

where X represents the body weight and y the body length.

On the basis of the body weights thus determined the weight of the brain can be culeulated by the revised forniuhi (8)

1.5G /.. u TN .569 n-jiR =^ / ^.1.56

log X 1-^^ (x - 8.7) -^^^ - 0.316 V

y = — "

/ X 1-^^ \

2 o V ^ (x - 8.7)

r^ '- . ' .1 (8)

Ll + (logx) « 1 + (logx) "-'J

as given by Hatai, '09, in this number of this journal, in which y represents the weight of the brain and x the body weight. The computation is simpler, however, if we use y := .554 +' .569 log (x — 8.7) ... (1) (Donaldson, '08) when X > 10, and a special formula

y — 1.50 log (x) — .87... (7) (Hatai, '09)

when X < 10,

The results obtained from these two formulas are identical with those from formula (8), and are given in the third column of Table ?>. The corresponding curve is shown by the continuous line on Chart III.

When the means are determined by the aid of a correlation table, in which the records are arranged in groups diifering by 10 mm. in body length and 0.1 gms. in brain weight, the co-efficient of correlation between body length and brain weight is found to be .86, which is high.

(b) The relation of the weight of the spinal cord to the body length.

When the individual records for the weight of the spinal cord are plotted in relation to the body length, we obtain results which are surprisingly regular. See Table 3 and Chart III (189 males, 137 females).

As in the case of the determination of the brain weights, the

i62 'Journal of Comparative Neurology and Psychology.

body weights used were those calcuhited by formula (4), and then the theoretical curve which fits these results most closely has been obtained by the use of the formula (3). (Donaldson, '08.)

y = .585 log' (x + 21) — 0.795 (3)

where y represents the weight of the spinal cord and x the weight of the body.

This curve apparently forms a straight line, though in reality it is a trifle convex towards the base line.

From the correlation table based on groups dilfering by 10 mm. in body length and .04 gms. in spinal cord weight, we obtain a coefficient of correlation which is .99, being almost f»erfect.

It will be seen from the foregoing that the weight of the spinal cord can be inferred from the body length with a high degree of accuracy.

In this connection an application of the foregoing data can be made at once. It was noted in a previous paper (Donaldson, '08, p. 3G0) that for rats of the same body weight, but of different sex, the central nervous system in the male was slightly heavier than in the female. The question naturally arises, therefore, whether there is any somatic character with which this diiference in the weight of the central nervous system according to sex can be connected. I shall endeavor to show that in the sex difi^erence in body length we find such a character.

It has been ])ointed out in the present paper (p. 158) that for the same body weight the males have a slightly greater body length than the females. It will be of interest, therefore, to determine whether this diiference in body length is sufficient to account for the difference in the weight of the central nervous system.

It is to be remembered in this connection that when males and females of like body weights are compared, the brain in the male is absolutely heavier, but the spinal cord is absolutely lighter. (Donaldson, '08.)

The relative difference is slightly greater in the case of the spinal cord, but the absolute mass of the brain is so much greater than that of the cord that as a final result the entire central nervous system is found to be heavier in the male.

Donaldson, Brain and Spinal Cord of Rat. 163

If we turn now to the preceding Table 2, we find the percentage difference between the body lengths for the two sexes (as determined from the average of the percentage differences between the five pairs ranging from 155 to 205 gms. in body weight) to be 1.74 per cent, in favor of the male. That is, on the average, mature males of a given body weight exceed by 1.74 per cent, in body length females of a like body weight.

If now we select the body length of 103 mm., which is that for the male having a body weight of 185 grams (see Table 2), and consider that this body lengih is 101.74 i)er cent, of the corresponding female body length, we find by calculation that the body length of the latter is 189.7 mm., thus giving an absolute difference of 3.3 mm. in favor of the male. In order to determine what dift'erence in the weight of the central nervous system would correspond to this difierence in body length, we may refer to the preceding Table 3, where the weight of the nervous system (both sexes combined) is given according to the body lengths. From this table it is possible to determine how much increase in the weight of the nervous system corresponds to an increase of 1 mm. in body length. Taking the entries from the body lengths of 180 to 210 mm., we obtain the following : —

Average increase in From Increase in body length the weight of the

central nervous system

180-190 1 mm. .0092

191-200 1 mm. .0081

201-210 1 mm. .0087

Average 0087 gms.

If the average difference in weight for 1 mm., as shown by the table, is .0087 gms,, 3.3 mm. would imply an absolute difierence of .02871 grams. This amount is 1.20 per cent, of the weight of the nervous system for a rat 195 mm. in body length (this is the mid-value between 180 mm. and 210 nun., the limits taken in the foregoing table). In Table 6, of the previous jiaper, Donaldson, '08, it appears as an average of all the groups taken in pairs, that for rats of like body weight, but difi'ereut sex, the entire central

164 'Journal of Comparative Neurology and Psychology.

nervous system in the male exceeds that in the female by 1.13 per cent/

It will be seen from the foregoing that the increase in the weight of the nervous system in the female, when the body length is made equal to that of the male, is 1.20 per cent, and the anticipated difference is 1.13 per cent. It follows that the difference according to sex in specimens of like body weight is accounted for by the difference in stature, the female having the smaller central nervous system because the stature of the female is less than that of the male.

When, therefore, the influence of body weight and of stature is taken into account, the weight of the entire central ner\-ous system in the two sexes is similar. It still remains true, however, that there is a characteristic division of this total weight according to sex, whereby the male has a slightly heavier brain, but a lighter spinal cord. These results are in accord with the more recent observations on the human nervous system. (Brain: Blakeman, '05; Lapicque, '08. Spinal cord: Mies, '93; Pfister, '03, and Donaldson, '08.)

COMPAKISON OF THE BoDY LeNGTH OF THE AlBINO KaT WlTIl

THE SiTTiiN^G Height of Man.

The objection is often made that the length measurements on the lower mammals cannot be compared with the measurements of stature in man because of the differences in the relation of the head to the trunk, and of the trunk to the legs.

As a matter of fact, however, the body length (trunk -f- head) which we have taken in the rat involves measurements of the pelvis, vertebral column and the skull quite comparable with those made in determining the sitting height in man. The chief difference is in the case of the skull which is measured from base to vertex in man, while in the rat the measurement is along the fronto-occipital axis, and so includes the nasal bones. These latter grow a trifle more rapidly than the cranium, especially in the male (Ilatai, '07), but

'The value of 8 per eeut given in Donaldson, '08. iiage 300, ninth line, is an error. The correct value is 1.13 per cent as given above.

Donaldson, Brain and Spinal Cord of Rat. 165

the difference becomes insigiiificaiit in comparison with the other parts of the skeleton which contribute so much more to the total result.

We, therefore, conclude that a comparison between the body length of the albino rat and the sitting height of man may be properly made.

The purpose of making such a comparison is to determine whether the rat is similar to man in the way in which this character changes with age.

It is not a character which at the time needs to be studied in detail and so only very general statements are necessary.

In his study on the growth of school children at Worcester, Mass., West ('92) made records for the sitting height in both sexes between the ages of 5 years and 21 years. The results are charted in his Fig. 1 (p. 32) and given in his Table 1 (p. 35).

If we take the average values of the sitting height in man for the two sexes, first at 19 years of age and again at 5 years of age, we find the following: —

Sitting height at 19 years 873 nmi.

Sitting height at 5 years , 595 mm.

Difference 278 mm.

Percentage gain, 47 per cent.

For comparison it is necessary to determine the increase in body length in the albino rat during the corresponding interval.

Computing from birth as the zero age, and taking the time unit for the rat on one-thirtieth of that for man (see Donaldson, '06), we obtain the following: —

. Nineteen years of human age correspond with 220 days of rat age.

Five years of human age correspond with GO days of rat age.

Table 9, in Donaldson, '08, shows that 220 days correspond with an average body weight of 234 grams, and of 60 days, with 78 grams. The corresponding body lengths in the rat, as shown in Table 2, are for

1 66 'Journal of Comparative Neurology and Psychology.

234 grams 209 mm.

Y8 grams 147 mm.

Difference 02 mm.

Percentage gain, 42 per cent.

It appears, therefore, that while the sitting height in man increased 47 per cent during the greater portion of the active growing period, the body length in the rat increased 42 per cent during the corresponding period.

Though not exactly alike, these tigures represent changes of the same order, and this is all that we desire to show at the present time. The value of this determination, so far as it can be foreseen, is to indicate that the spinal cord during growth is subject to approximately the same relative amount of passive lengthening in both man and the albino rat.

Conclusions.

1. In the albino rat the ratio obtained by dividing the body weight by the body length diminishes as the body weight increases.

2. Among rats of the same body weight, the males have a slightly greater body length than the females.

3. The correlation between body weight and body length is high, being .90.

4. The correlation between body length and l)rain weight is high, being .86.

5. The correlation between body length and the weight of the spinal cord is nearly perfect, being .99.

6. The greater weight of the central nervous system in male, as compared with female rats of like body weight, is completely explained by the greater body length of the males. This result agrees with the more recent observations on man.

7. The relative increase in the body length of the' rat during active growth is similar to the increase in the sitting height of man during the corresponding period. Hence, in both forms, the

Donaldson, Brain and Spinal Cord of Rat. 167

spinal cord is suhjeet to a coiTCspoiidiiig ainoimt of passive lengtli

eniiio'.

'5 8. The body leiigtli is a l)etter datum than the body weight from

which to infer the weight of the Itrain or of the spinal cord. This is especially true when there is any reason to suspect emaciation of the body.

9. A mean of the two determinations of the weight of the brain or of the spinal cord (1) from the body weight (when normal) and (2) the body length, will give better a]i]>roximations than the determination based on either datum alone.

BIBLIOGRAPHY. Blakeman, J.

190.5. A study of tlie biometric constants of English brain-weights, and their relationships to external physical measureaients. Biometrika, vol. 4, pp. 124-100. Donaldson. H. H.

1000. A comparison of the white rat with man in respect to the growth of the entire body. Boofi McmorUil VoJkiiic, pp. .5-0.

1908. A comparison of the albino rat with man in respect to the growth

of the brain and of the spinal cord. Joiini. Coiiip. NciiroL,

vol. IS, no. 4, pp. 345-.302. Hatai, S.

1907. Studies on the variation and correlation of skull measurements in

both sexes of mature albino rats (Mus norvegicus var.

albus). Alitor. Joiirn. Anat., vol. 7, no. 4, pp. 423-441. 190S. Preliminary note on the size and condition of the central nervous

system in albino rats experimentally stunted. Jourii. Coiiip.

NcuioL. vol. IS. no. 2, pp. 151-155.

1909. Note on the fornuilas used for calculating the weight of the brain

in the alI)ino rats. Joitrii. Coiiip- ^^ciiioL, vol. 19, no. 2. Mies, J.

1S93. Ueber das Gewicht des Ruckenmarkes. Centrum, f. NcrvenheilJitindc ti. Psychiatric, S. 1-4.

PriSTER, H.,

1903. Zur Anthropologie des Riickenmarks. Neurol. Central!)., Bd. 22, S. 757 und 819. West, G. M.

1892. Anthropometrische Untersuchungen iiber die Schulkinder in Worcester, Mass., Amerika. Archiv f. Anthropologic, vol. 22, pp. 13-48.

NOTE ON THE FOKMULAS USED FOE CALCULATING THE WEIGHT OF THE BRAIN IN THE ALBINO RATS.

BY

SHINKISHI HATAI.

Associate in Xciiyolofju at The Wistar Institute.

In previous papers (Hatai, '08; Donaldson, '08) the formnlas for calculating the weight of the brain and of the spinal cord in relation to the body weight were determined on the assumption that the amount of increment to the weight of these parts is proportional to the reciprocal of the body weight plus a constant or

^=h— ^ (1)

a^ (x + a) ^ ^

where y is the weight of the brain or spinal cord in grams and X the weight of the body in grams.

Integration of (1) gives at once the value of y.

Thus :

y=h f (x + a) ^^"^^ ^"^ ^^ + a) + c

(X + a) or in our previous notation (Donaldson, '08) :

y = A + C log (X + ;^) (2)

This type of logarithmic formula has been used by the present writer ('08) and Donaldson ('08) and was found to be very satisfactory for representing the relation between the body weight and the weight of the brain or spinal cord.

This type was further employed by Donaldson ('09) to represent the relation between the body weight and body length and was proved by him to be satisfactory.

The .ToiRXAL of CoMr-vnATivE NEfnoLouv and rsYciior.dOY. — Vol. XIX, No. 2.

1 70 Jotirnal of Comparative Neurology and Psychology. The formula in each case was as follows:

Brain weight or y = .569 log (x- 8.7) +0.554 (3)

Spinal cord weight or y = .585 log (x +21 ) -0.795 (4)

Body length or y = 143 log (x + 15 ) - 134 (5)

Although the formulas (4) and (5) are entirely free from theoretical objections within the interval x = (5 grams, 325 grams), the formula (3), however, has two defects when we apply it to the case of x < 8.7. The first defect, which appears when x, the body weight, is less than 8.7 grams, is due to the fact that the resulting value of (x — 8.7) becomes a negative quantity and the logarithm of such a quantity is necessarily imaginary'. The difficulty thus presented is, however, merely a theoretical one, since for the purpose of computation the following method may be employed.

Let us consider the two cases when x is greater than a and when X is less than a then we have

(A) ^ = ^ when x > a

dx (x-a)

(B) ^^ = — ^ when x < a

dx (a — x)

Then integration of (A) leads to the foi'mula (3) which we have already obtained, that is

y=A+Clog (x-B)=.554 + .569 log (x -8.7)

while the integration of (B) becomes

(C) y = A - C log (/? - x) = .554 - .569 log (8.7- x)

The formula (C) thus obtained gives results identical with those obtained when we compute the value of y from the formula

(D) y = .554 + .569 log (-C)

In this case, of course, with an understanding that log ( — C)

should be treated as equivalent to ■ — log C.

As long as the results obtained by the formula (C) agree with those obtained by the formula (D), the following procedure is justified.

Hatai, Weight of the Brain. 171

In the formula {?>), when the variable x is less than a constant C, or in this case 8.7, we can take the logarithm of the real positive nnmber (C) and pnt a negative sign before it, i. e., .569 log( — C) =

— .509 log C where — C = (x — 8.7).

With the foregoing nnderstanding, the formnla ean thns be applied even in the case of a rat, the body weight of which is less than 8.7 grams.

There remains, however, a second defect in this formula (3) which cannot be overcome.

When the value of x lies between 7.7 and 9.7 grams, the formula fails to represent the observed values on account of sudden change in the course of the resulting curve. Although this interval is very small when we consider the whole extent of the curve, yet it prevents the general ai^plication of the formula.

In Chart I, Plate II, in the paper by Donaldson, '08, the curve representing the change in the brain weight between the body weights of 5 and 10 grams was completed by simply joining the two points, both of which had been carefully calculated by the formula (3), and it was not until we came to consider the formula in another connection that we appreciated the impossibility of applying it to this interval.

I have now obtained a revised formula which is free from the foregoing objections. At the same time it should be stated that the values obtained by this new formula do not differ from the values so far as computed by the previous formula (3), or as given by the ideal line by which the curve was previously completed.

I shall jDresent first the theoretical considerations touching the revised formula. Let us consider the series

2 o ^^ ^-^Ll+z" l + z-'-'J

(6)

where i|;(z) and (^(z) are (some) functions of z. The sum of the first n terms of S becomes obviously

" ^^ '^ l+z°

172 'Journal of Comparative Neurology and Psychology.

When (z) < 1, the limit of z" is zero for n = oc and consequently

s = i//(z).

On the other hand, where (z) > 1, z" tends to cc and therefore in this case S = (f) (z). (See Jordan "Conrse d'analyse," Tome I, p. 320.)

We have shown already that the brain weights in rats in which the body weights are greater than 10 grams, can be calculated by the formula

y = .569 log (x-8.7) +0.554 (3)

Later we found that the brain weights in rats in which the body weight lay between 5-10 grams may be calculated l)y a special formula for this portion of the curve, namely:

y = 1.56log(x)-.87 (7)

and therefore in the two formulas (3) and (7) y can be considered as the function of log x.

The values calculated by the latter formula (7) agree perfectly with the ideal line which completes the brain-weight curve between 5 and 10 grams of body weight.

As has been shown already, the formula (6) is perfectly general in its application when two conditions are satisfied; namely, when |z| > 1 in one case and |z| < 1 in the other.

We also found that not only are the two formulas (3) and (7) functions of log x, but that (1) is applicable to rats in which the body weights are more than 10 grams or | log x | > 1, while formula (7) is applicable to rats in which the body weight is less than 10 grams or | log x | < 1. This satisfies all the necessary conditions.

Thus a combination of the two formulas (3) and (7) will enable us to calculate the brain weight for any given body weight from 5 grams to 320 grams. (Extrapolation may be used towards the upper end of the curve).

The final formula is represented by the following: —

1.56 .569

logx (x - 8.7) - 0.316 '^ /, xl-56

^ o ^ (x-S.7)

r-^—- ^— 1 '.(8)

Ll + (logx)" l + Clogx)"-' J

in which y represents the brain weight and x the body weight.

Hatai, Weight of the Brain. 173

As to the actual use of the above formula, I may add the following remarks.

As was mentioned already, the series reduces to

<l> (log x) =.569 log (x-8.7) +0.554

when (log x) is greater than 1 ; while, on the other hand, the series reduces to

yjj (log x) = 1.56 logx-.87

when (log x) is less than 1. Therefore it is only necessary to note whether we are treating rats in which the body weights are greater or less than 10 grams.

If the l)ody weight is greater than 10 grams, we can simply use

<^ (log x) or y = .569 log (x- 8.7) +0.554

and if it is less than 10 grams, the other formula

\p (log X) or y = 1.56 log x - .87

Of course, one can determine the brain weight directly from the formula (8) after some laborious calculation; nevertheless such a procedure has no advantage over the simpler process described above.

The present formula (8) is desirable simply, first, because it is free from the theoretical objections ; and, second, because by it we can express the complicated relations existing between the body and brain under a sinele generalized form.

BIBLIOCxRAPHY. Donaldson. H. H.

1908. A comparison of the albino rat with man in respect to the growth of the brain and of the spinal cord. ./. of Comp. Neui-ol., vol. IS, no. 4. 1009. On the relation of the body length to the body weight and to the weight of the brain and of the spinal cord in the albino rat (Mus norvegicus var. albus). ,/. of Comp. Neurol, vol. 19, no. 2. Hatai, S.

1908. Preliminary note on the size and condition of the central nervous system in albino rats experimenlally stunted. ,/. 0/ Comp. Neurol. , vol. 18, no. 2.

THE NERVUS TEEMINALIS (NERVE OF PINKUS) m THE FROG.

BY

C. JUDSON HERRICK.

From the Anatomical Laboratory of the JJnircrsitij of Chicuyo.

With Ten Figures.

A ganglionated nerve connected with the forebrain and intimately associated with l^he nerviis olfactorins has been described in nearly all gronps of fishes. The first clear description of snch a nerve is that of Pinkns ('04) for Protopterus. It was termed the nervus tcrmhialis by Locy, in 1905, and accurately described in twenty genera (27 species) of selachians, and it was mentioned by Allis ('97) as occurring in Amia, Brookover ('08) has described it more fully in Amia and Lepidosteus and at the meeting of the Association of American Anatomists in Baltimore, December, 1908, Brookover and Sheldon reported the presence of a similar nerve in the teleosts. Further literature on the subject is cited by the authors mentioned.

Ernst de Vries ('05) described a transitory ganglion on the vomeronasal nerve of mammals and suggested that the nerve of the organon vomeronasale (Jacobson's organ) of higher vertebrates is homologous with the nervus terminalis of fishes. Since, however, the organon vomeronasale of mammals is lined with sensory epithelium of the same type as the undoubted olfactory parts of the nose and gives rise to nerve fibers indistinguishable from other fila olfactoria (Read, '08), it is probable that its innervation does not differ from that of the other parts of the olfactory organ. In this case it is difficult to see how the nerve of the organon vomeronasale can be compared with the nervus terminalis of fishes, for the latter fibers are not known to connect with the specific cells of The Journal of Compaiiative Neukolouy and rsYCHOLooy. — Vol. XIX, No. 2.

S

176 Journal of Comparative Neurology and Psychology.

the olfactory mucous membrane, they hear a ganglion on their course and centrally, in most if not in all cases, they do not connect with the olfactory bulbs but with the brain farther caudad in the vicinity of the recessus preopticus or lamina terminalis. It may, therefore, be concluded that, while the nervus terminalis occurs in fishes generally, its presence has not hitherto been demonstrated in the adults of any forms above the fishes.

In examining preparations of the brain of the frog prei)ared by the Golgi method I found an impregnation of nerve fibers which conform so closely to the central course of the nervus terminalis of selachians and dipnoans that I have no hesitation in considering them homologous. In the first series of sections in which this nerve was seen its fibers were completely impregnated on both right and left sides from a position rostrad of the olfactory bulbs to their decussation in the lamina terminalis ; and, since the olfactory nerves and tracts were for the most part unimpregnated, the course of the nervus terminalis could be followed with precision. These findings were subsequently verified in several series of adult and larval frogs, as follows:

Transverse sections of adult Kana i)ipiens by the Golgi method (the series referred to above, Figs. 1 to 7).

Sagittal sections of adult Ttana ])i])i('ns by the Golgi method.

Transverse sections of the adult Rana pijiiens by the Weigert method. In this series the ])rocess of decolorization of the sections was incompletely carried out, leaving considerable color in the background, so that, though the intra-cerebral course of the nervus terminalis is unmedullated, the course of the tract could nevertheless be followed with precision. Other series of Weigert sections permitted the nerve to be identified where it enters the brain, but not through the brain substance, on account of the complete decolorization of its fibers.

Transverse sections of a half-grown frog tadpole by the Golgi method, illustrating the w^hole central course of the neiwe and its free terminal arborizations in the lamina terminalis (Figs. 9 and 10).

Horizontal sections of an old larva of Rana catesbiana 30 mm.

Herrick, ISfervLis Termirialis of Frog. 17

k-iig, stained with Delafield's liaMiiatoxylin and erjthrosiu (Fig. 8). ^ ^

In these five specimens all, or nearly all, of the intra-cerebral course of this nerve was followed on both the right and the left sides. In several other specimens, both larval and adnlt, smaller portions of the nerve were also seen. For all of the sections on which this work is based I am indebted to the skill of my assistant, ]\Ir. P. S. McKibben. The findings are briefly these.

In the series of transverse sections made by the Golgi method through the brain of the adult frog first referred to, at the level of the olfactory bulbs (Figs. 3 and 4), there is impregnated a compact fascicle of a few (probably less than 40) unmedullated nerve fibers on each side, lying between the meninges and the ventral surfaces of the olfactory bulbs. This is the nervus terminalis. When followed caudad the nerves of the right and left sides are found to be similar; but rostrad they exhibit slight differences.

On the left side at the extreme rostral end of the olfactory bulb the nervus terminalis has separated from the meninges and joined one of the small fascicles of fila olfactoria on the ventral border or the nervus olfactorius (Fig. 2). None of the fila olfactoria of this fascicle are impregnated, so that it is easy to follow the nervus terminalis separately on this side. Scattered fila olfactoria are impregnated in other parts of the nervus olfactorius and these are indistinguishable in appearance from the fibers of the nervus terminalis. Less than 1 mm. rostrad of the olfactory bulb (Fig. 1) the nervus terminalis passes from the ventral to the medial aspect of the nervus olfactorius, still embedded in its marginal layer, and here it disappears from view. At about this level the impregnated fila olfactoria also lose their stain, so that it is proba})le that the nervus terminalis continues farther rostrad in the nervus olfactorius, though its fibers are not farther impregnated in my preparations.

On the right side the relations are essentially the same, but not so clearly demonstrable ou account of the fact that impregnated fila olfactoria mingle in some places with the fibers of the nervus terminalis- and the latter is not so compact a fascicle. Its fibers

1 78 Journal of Comparative Neurology and Psychology.

lie somewliat more deeply embedded in the nerviis olfactoriiis than those of the left nerve. They curve dorsad and mesad before disappearing, as on the other side.

Passing caudad, the fila olfactoria enter the bulbus olfactorius, but the fibers of the nervus terminalis separate from them ventrally and constitute a small compact bundle of fibers lying in the meninges ventrally of the olfactory bulbs (Figs. 3 and 4). This position is maintained until they have passed caudad of all of the formatio bulbaris (glomerular formation) of the olfactory bulbs, where they turn abruptly dorso-mesad and enter the substance of the cerebral .hemisphere at its ventro-medial border. The point of entrance of these fibers varies somewhat in different specimens. It is in the adult aways farther caudad than any of the formatio bulbaris of the ventral and medial aspects of the olfactory bulbs, and in all but one of the observed cases farther caudad than the formatio bulbaris of the bulbulus accessorius on the lateral aspect of the olfactory bulb.

Having entered the brain, the nei-vus terminalis passes caudad (Fig. 5), turning slightly dorsad and laterad in its course, embedded in the ventral part of the hemisphere about midway between the ventral border of the lateral ventricle and the medial wall of the hemisphere. The fibers of the tractus olfactorius medialis lie ventrally and medially of it and those of the median forebrain bundle dorsally of it. Upon reaching the lamina terminalis (Fig. 6), it ascends more rapidly between the lateral forebrain bundle and the crossed portion of the medial forebrain bundle to enter the middle part of the anterior commissure complex dorsally of the pre-optic recess, where it decussates (Fig. T). The fibers can be clearly traced across the meson in a compact bundle, but their exact place of termination has not been determined. These relations were confirmed in every detail in the transverse Weigert series, and in everything except the decussation in the anterior commissure in the sagittal Golgi series.

In one of my series of transverse sections through the brain of the adult R. pipiens prepared by the method of Cajal the nervus terminalis can be followed in its course through the cerebral hemi

Herrick, Nerviis Terminalis of Frog. 179

sphere on one side. Neither the fila olfactoria nor the fibers of the nervus terminalis are stained, and therefore the peripheral relations cannot be determined. The nervus terminalis (unstained) is seen to detach itself from the olfactory nerve under the olfactory bulb and to pass back to the lamina terminalis exactly as already described. Within the brain it is surrounded by a dense mass of deeply stained fibers belonging to the secondary olfactory and other systems, so that it can easily be followed back to its decussation as a clear yellow area surrounded by a dark field of impregnated fibers.

Professor J. B. Johnston informs me that in 1905 he observed a similar nerve in Golgi sections of the adult frog brain; but since he had no control of this single observation, it was not published.

In the Golgi sections of the young larva (Figs. 9 and 10) the nervus terminalis is seen to enter the lamina terminalis and there its fibers arborize, some of the free termini crossing the meson and some remaining uncrossed. Other histological preparations of the larva show that the cells in the region of these arborizations are much crowded, forming the nucleus medianus septi. It is probable tliat in the adult also the nerve ends in the nucleus medianus septi, either wholly crossed or partly crossed and partly direct, as in the tadpole.

In the horizontal series of sections through the brain of an old larva of Rana catesbiana stained with hsematoxylin and crythrosin very nearly the whole intra-cerebral course of the nervus terminalis is sho^\m in four consecutive sections, as seen in Fig. 8. The nucleus medianus septi does not appear here. It lies immediately dorsally of the plane figured as a dense mass of cells which crosses the median plane in the lamina terminalis directly ventrally of the foramen of Monro.

The relations of the nervus terminalis of the larva are essentially similar, so far as observed, to those of the adult frog, save that the nerve enters the brain relatively farther rostrad and farther laterad in the larva. Fig. 8 shows it penetrating the formatio bulbaris rostrad of the l:)ulbulus accessorius. It seems probable to me that the point of entrance of the nervus terminalis remains relatively

l8o "Journal of Comparative Neurology and Psychology.

fixed, the changed relations of the adult being due to a farther growth of the olfactory bulbs rostrad rather than to a recession of the nervus terminalis caudad. In the adult the olfactory capsules lie far rostrad of the olfactory bulbs, while in the larva they lie in about the same transverse plane, the olfactory nerves passing out to them almost laterally.

In none of my preparations have I been able to trace the fibers of the nervus terminalis distally more than about 1 mm. beyond the olfactory bulbs. I have not examined the peripheral relations of the olfactory nerve and nasal capsules in the adult frog. In several preparations of frog tadpoles (probably R. pipiens) I have found cells scattered along the peripheral course of the nervus olfactorius which differ from the sheath cells of the fila olfactoria. The clearest case observed is a large tadpole taken just before the metamorphosis, which was prepared by the method of Cajal and cut into horizontal sections. Scattered along the ventral surface of the olfactory nerve in the middle part of its course are more than 100 nuclei which differ conspicuously from the sheath nuclei among which they lie, being round or broadly oval and twice as wide as the narrowly oblong sheath nuclei. They are scattered along the olfactory nerve from its foramen through the skull to the point where it lu'eaks up to spread over the olfactory mucous membrane. From the similarity of these nuclei to those found by Brookover on the nervus terminalis of ganoids and teleosts I incline to regard them as belonging to ganglion cells of the nervus terminalis of the frog, though I have not been able to demonstrate their fibrous connections.

The exact central connection of the nervus terminalis also demands further investigation. The single impregnation of the terminal arborizations in the lamina terminalis of the voune; lan^a is not altogether conclusive and this observation must be verified and extended before much weight can be given to it. One is, however, struck by the similarity between this observation and the descriptions of Locy of the central relations of the nervus terminalis in the selachians.

In conclusion, it seems clear that the nerve here described in

Herrick, Nerviis Trnninalis of Frog.

I«I

the frog is iii(3rplio]<)<.-ieally similar to the iiervus toriiiinalis of fishes, so far as our information extends.

LITER ATUK K CITED. Aixis, E. I\

1S07. The cranial muscle.s and cranial and tirst si)inai uerves in Aniia calva. Juiini. Morttli., vol. 12, no. o. Bkuoko\ei!, C.

1!)US. Pinlai.s's nerve in Aniia and Lepidosteus (Abstract.) ISvivncc.

N. f^., vol. 27, no. 702, p. Ulo. LocY, W. A.

VM-j. On a newly-recognized nerve connected witli the forebrain of

selachians. Anat. Aiiz.. vol. 2(), nos. 2 and 3.

I'lNKUS. F.

1894. Ueber einen noch nicht beschriebenen llirnnerven des I'rotopterns annectens. Aimt. An.:., vol. 9, no. LS, pp. G02-5GG.

liKAU, El-FIE, A.

1908. A eontribntion to the knowledge of the olfactory apparatus in dog, cat and man. Am. Jouni. of Anat., vol. 8, no. 1.

DK VRIES, E.

1905. Note oil the ganglion vonieronasale. Proc. of the ^Section, of Science, Icon. AlcmJ. van Wettcnscluii)i)cn, Amsterdam, vol. 7, part 2.

Figs. 1 to 7. A series of transverse sections through the hrain of adult Kana pipieus prepared by the Golgi method, to illustrate the central course of the nervus terniinalis. All except Fig. 1 are drawn to the same scale.

Fig. 1. Through the olfactory nerves about 1 mm. rostrad of the olfactory bulbs. X 60.

Each olfactory uerve is broken up into numerous fasciculi, some of the larger of which are indicated by the dotted outlines. A small proportion of the fibers of the olfactory nerve (fila olfactoria) are impregnated, some of these fibers being scattered singly among the fasciculi of the nerve, others aggregated into more or less definite Inuidles. The fascicle marked n. tcrmiiiaUs on the left side is unmixed with flla olfactoria ; on the right side the two fascicles so designated are probably of mixed character.

'n. terniinalis'

Fig. 2. Through the olfactory nerves just at their entrance into the olfactory bulbs. X 30.

The most rostrally jilaced glomeruli occiipy the dorsal part of the section. The fila olfactoria occupy the middle and ventral parts of the section and only a few of them are impri'gnated. The nervus terniinalis is embedded in the most ventral part of each olfactory nerve and all of its fibers are imi)regnated.

Herrick, Nerviis Term'maJis of Frog.

183

om

eruti

n. term l-riaUs

Fig. 3. Tlirougli the olfactory bnlbs at the level of the rostral ends of the rhiiiocoeles. x 30.

The stippled area surrounding each rhinocoele indicates the extent of the layer of granule cells, most of which are not impregnated in the preparation. A few impregnated granules are drawn on the right side. Several typical neurones of the mitral cell layer are drawn on the left side. Fibers of the tractus olfactorius pass dorsad from all parts of the mitral cell layer. The glomerular layer lies farther ventrally, while the layer of fila olfactoria occupies the extreme ventral part of each bulb. The nervus terminalis has separated from the fila olfactoria on each side and lies between the latter and the meninges.

184 'Journal of Comparative Neurology and Psychology.

'( facto ri US med

■MpstolfdctoriLLS qr$niMle ceils \n.terniLndiLS Fig. 4. Through the olfactory bulbs at the level of the rostral eud of the bulbulus accessorius. X 30. The layer of grauule cells is indicated by the stippled area, ventrally of which is the layer of mitral cells (unimpregnated). The dorsal half of the section is occupied by secondary olfactorj^ cells, two of which are imperfectly impregnated. The fibers of the lateral secondary olfactory tracts are not impregnated. They lie chiefly along the dorso-lateral border of the section external to the dotted line. The olfactory glomeruli are limited to the extreme ventral part of the section. The nervus terminalis lies still farther ventrally close to the meninges. Herrick, Nnvus Tcrwinalis of Frog. 185 olfdcl onus septL tnolfdctori as ined. nied. fore bra in bundle [at fore bra in bu^ndle n.termLndUs vetitro-ldterdlis Fig. T). Through the cerebral hemisphere hetween tlie olfactory hnlh and the lamina termiualis. X 30. The lateral and medial basal forebrain bundles are impregnated and the neurones marked a and It send their axones into these two bundles respectively. The tract marked tr. olfactorius incdialis contains also other elements, particularly olfactoi-y fibers of the third order for the nucleus medianus septi. The ventral fibers of the lateral secondai'y olfactory tract are impregnated, but not the dorsal fibers of this tract. The nervus terminalis lies ventrally of the median forebrain bundle and laterally of the tr. olfactorius medialis. 1 86 Journal of Comparative Neurology and Psychology. _,^'ir olfcictorius sepii "^N^. "'^-la.t. forebrairv bundle ^^^/a: olfactorius veniro-lat. Fig. G. Through the most rostral part of the hiniiua terminalis. X 30. The tract marked medial forchrain hividle on the right side is composed chiefly of the crossed portion of this tract (cf. flg. 7). Fibers of the uncrossed portion of this tract arise from the cells marked J) on the left side. As in fig. 5, the tract marked tt: olfactorius mcdialis contains also tertiary olfactory fibers for the nucleus medianus septi. Hfrrick, Nervus TerrninaJis of Frog. .87 . -^conimiss] /hippocampi id t. fore bra in bundle Fig. 7. Section iiniuecliately rostrnl to the (lecnssation of the nervns terminalis in the Inminn teminalis. X 30. The decussation of the medial forehrain hundle {mcd. f. 7>. 6.) occupies the ventral part of the lamina terminalis. The commissura hippocampi (dorsal commissure) is approaching the lamina terminalis from the dorsal side. The other elements of the anterior commissure complex lie farther caudad. l88 yoiirtjal of Comparative Isleiirology and Psychology. Fig. S. Composite drawing of liorizontal sections tlirough the brain of a tadpole of Rana catesbiana aboiit 30 niui. long, to show the central course of the nervus terminalis. X '5^^ The specimen was beginning the metamorphosis when preserved, having the hind k'g buds about (i mm. long. Sections wore cut in the horizontal plane 30 microns thick and stained with Delafield's hjematoxylin followed by erythrosiu. In these sections the nuclei of the cells are clearly stained and some of tlie forebi'ain tracts. Among the latter is the nervus terminalis on both sides. Distally this nerve can be followed only a very short distance after leaving the brain, its fibers being mingled with fila olfactoria and Indistinguishable from them, both being unmedullated. Centrally the nerve can be followed back to the lamina terminalis, where it plainly decussates in tlif anterior commissure. All of the details of this figure are taken from section 19 of the series, except parts of the nervus terminalis which are taken from the neighboring sections whose numbers they bear. Section 10 shows the nerve at its point of entrance into the rostral end of the hemisphere and also its decussation in the lamina terminalis. With the aid of the camera lucida I have projected upon the outline of this section the remainder of the lutra-cerebral course of the nerve, which is all included within the three sections lying next ventrad (sections 20, 21 and 22). Three elements of the anterior conunissure complex are shown, the decussation of the nervus terminalis, the decussation of the lateral forebrain bundle and the commissure of the corpora striata. The decussation of the medial forebrain bundle lies in the plane of the section, but it is not stained in the preparation ; cf. fig. 7. Herrick, Nervus Terminahs of Frog. 189 ,n.terniina.li^ -/jy.latjrdt forekillSraL biindCe. r^ ' ■ ^-^Lcof'p. striatum '-^^ '/:l^'W^ -xi?^i recessiis fp; preoptlcus optic \ chiasma 190 Journal of Cojuparative Neurology and Psychology. tr.ol facto Fig. 9. A transverse section through the braiu of a half-grown frog tadpole, taken just behind the olfactory bulbs. Golgi method, x ^^ The nervus terminalis is shown immediately after its entrance into the cerebral hemisphere. A few fibers of the tractus olfactorius are impregnated ventrallv of it. Fig. 10. The section following immediately caudad of the one shown in fig. 9. X TO. The section is very thick and shows almost the whole central course of the nervus terminalis and its ending by free arborizations in the nucleus medianus septi. The tractus olfactorius lateralis is also impregnated. THE NERVUS TERMINALIS IN THE CARP. R. E. SHELDON. From the Anatoinkal Luburutury of the University of Chicago. With Seven Figures. Fritsch figured in 1878 the stump of a nerve arising mesad of the olfactory from the rostral aspect of the brain of Galeus canis. This he called an "iiberzahliger Nerv." Our present knowledge concerning it is, however, due almost entirely to the work of the last few vears, during which its existence has been demonstrated in group after group among the lower vertebrates. Pinkus, '94, '95, in Protopterus was the first to trace and describe the entire course of the nerve which he called simply, ein neuer ISTerv." Allis, '97, found the nerve in Amia but added nothing concerning its structure and connections, naming it, however, the nerve of Pinkus. Tocy, '99, in Acanthias described a nerve, closely associated with the olfactory, as in the cases previously reported, but ganglionated. Pinkiis had founds cells in connection with the nerve in Protopterus, but hesitated to call them a ganglion. In 1902 Sewertzoff described in Ceratodus embryos a similar ganglionic nerve which he named the nervus preopticus owing to the fact that it appears to arise near the preoptic recess in Dipnoans. Later, Burckhardt, in 1905, found the same nerve in adult Ceratodus. Locy in several papers, '03, '05a, '05b, takes up in detail its occurrence in different groups of selachians, its peripheral and central connections and its embryonic history. He pointed out its homology with the nerve of Pinkus in Protopterus and Amia and the nervus preopticus of Ceratodus, proposing for it the name of nervus terminalis. Brookover, '08, demonstrated a ganglion for the nerve in Amia and Lepidosteus, adding also to our knowledge of its peripheral connections. The Jouenal of Co.mi'Arative Nelt.ology and Psychology. — Vol. XIX, No. 2. 192 'Journal of Comparative Neurology and Psychology. Until very recently, however, its presence in forms other than the selachians, ganoids and dipnoans has not been demonstrated. At the Baltimore meeting of the Association of American Anatomists, in a joint paper, Brookover and I reported its existence in teleosts, describing its ganglion, and peripheral and central connections. At the same time Herrick showed that it is also present in both the larval and adult frog. This paper takes up in detail the central course of the nerve in the carp, Cyprinus carpio. It consists entirely of a tract of unraedullated fibers which was traced by means of the following methods and material. 1. Weigert method. (a) Four transverse series of the entire olfactory crura, bulbs and nasal capsules of adult individuals about 50 cm. in length. (&) Two longitudinal series through the olfactory bulbs and capsules of similar adults. (c) Two transverse series through the cerebral hemispheres of adults. (d) One transverse series through the entire head of a youngcarp about 2 cm. in length. These were stained by a modification of the straight Weigert method which left the unmeduUated fibers a reddish brown. This method was particularly valuable, as the tract is, for most of its extent, surrounded 1)V meduUated fibers from which it stands out quite distinctly. 2. Vom Rath method. One transverse and one longitudinal series through the olfactory crura, bulbs and capsules of an adult about 30 cm. in length. This method also gave good results, as the unmedvdlated fibers appear lighter in color than the medullated. 3. Cajal method. Two transverse series through the cerebral lu'mispheres of an adult about 35 cm. long. In my preparations tlie nervus terminalis is an orange yellow, while the uu^dullated fil)ers suiToniiding it are nearly black. These series were of especial value in showing the decussation of the nerve in the anterior commissure. 4. Toluidin blue and thionin methods. Sheldon, Nervus Tcninnalis in the Carp. 193 (a) Two transverse series tliroiigii the bulbs and capsules in an individual of 35 cm. length. (h) One longitudinal series through the bulbs and capsules of a fish oO cm. long. (c) One transverse series through the hemispheres of an individual 40 cm. in length. By these two methods one can easily demonstrate the peripheral ganglion and the nucleus in which the fibers end centrally. As noted in the paper reported before the Association of American Anatomists, numerous scattered ganglionic cells are found on the ventro-median side of the olfactory nerve about half way between the formatio bidbaris and the olfactory capsul^. Cells of the same type are also found caudad to the glomerular region and rostrad to the nasal capsules, diminishing rapidly in numbers, however, as one passes caudad or rostrad from the main group of cells. It was also noted that Cajal preparations show coarse fibers which can be traced from these cells rostrad to the olfactory mucous membrane, where they are distributed to the epithelium with the olfactory nerve fibers. The tract which forms the central course of the nerve is easily distingaiished about half way between the caudal and cephalic ends of the olfactory bulb on its ventro-median side. Here it is surrounded by the medullated fibers of the tractus olfacto-lobaris medialis as shown in Fig. 1, Rostrad of this point, however, it is soon lost, as the medullated fibers either end or seek a new position, leaving the nerve surrounded only by the unmedullated fibers of the olfactory nerve. Scattered among these fibers in this region are the cells of the peripheral ganglion. As the tract passes caudad from the bull) it continues to hold its position on the ventro-median aspect but migrates peripherally until it lies next the meninges, surrounded on three sides, however, by the tractus olfacto-lobaris medialis (Fig. 2). This relation holds throughout the length of the olfactory cms (see Fig. 3). On reaching the cerebral hemispheres the tract turns dorso-laterad through the tractus olfactolobaris medialis to lie, for a time, between the latter and the radix olfactoria medialis propria (Fig. 4). It holds this position for some 194 'Journal of Comparative Neurology and Psychology. distance lying partly enclosed by the tractus olfacto-lobaris medialis. When the anterior commissure is reached, however, it turns abruptly mesad (Fig. 5) and largely decussates in the mid-line (Fig. G) in the most rostral part of the commissure. Part of the libers apparently do not cross but end on the same side. The exact termination of the fibers could not be demonstrated. It is certain, however, that they end close to the mid-line, without doubt in the dense nucleus of small cells shown in Fig. 7. x\s was stated in the earlier report, it should be noted that the connection between the central tract and the peripheral ganglion with its fibers has not been established and cannot be except by fortunate Golgi or Cajal preparations. There can be little doubt, however, that this is the nervus terminalis for the following reasons. The ganglion and peripheral distribution of the fibres are identical with the condition found by Brookover in Amia and Lepidosteus in connection with what is undoubtedly the nervus terminalis. It is also similar to that described for Protopterus by Pinkus, for Ceratodus by Sewertzoff and for selachians by Locy. The course of the central tract' corresponds to that shown for the nervus terminalis of selachians by Locy, who worked out the central termination in some detail. Still stronger support comes from the findings of Herrick in the frog. In this form the neiwe takes the same course through the brain, including the decussation. Peripherally, however, the nerve leaves the brain to run in the meninges rostrad past the formatio bulbaris so there can be no question as to its character. Summarizing: there is little doubt that there exists in the carp a nerve comparable morphologically to the nerve of Pinkus or the nervus terminalis of other fishes and the frog. Hull Laboratory of Anatomy, The University of Chicago, January 18, 1909. Sheldon, Nervus Terminalis in the Carp. 195 LITERATURE CITED. Allis, E. I'. 18U7. The cranial muscles and cranial and first spinal nerves in Aniia calva. Jour. Morpli., vol. 12, no. 3, March, 1897, pp. 487-808, pis. XX-XXXVIII. BiNG, Robert and Burckiiardt, Rudolf. 1905. Das Centralnervensysteni von Ceralodus forsteri. Hcintia, Zoolog. ForscJiunf/srciscu hi, Ausfral. i(. il. Malaij. Arch., pp. 513-584, Taf. XLII, 35 Fig. iu Text. Brookover, Chas. 1908. Pinkus's nerve in Amia and Lepid^)steus. /S'e/c'»c'C, N. S., vol. 27, no. 702, June 12, 1908, pp. 913-914. Fritsch, G. 1878. Untersuchungeu iilier den feineren Ban des Fischgehirus. Berlin. 187S, pp. 1-94. i-xy, 13 pis. IIerrick. C. Judson. 1909. The nervus terminalis (nerve of IMnkus) in the frog. Rcitnrt at Baltimore meeting, Assoc. Amcr. Anat., 1909; and Journ. Conii). Ncvrol. and Psych., vol. 19, no. 2. LocY, W. A. 1899. New facts regarding the develoimient of the olfactory nerve. Anat. Anz:. Bd. IG, no. 12, pp. 273-290. Fig. 1-14. 1903. A new cranial nerve in selachians. Mark Annivcrsanj Vol., article III, pp. 39-55, pis. V-VI, 1903. 1905a. A footnote to the ancestral history of the vertehrate hrain. Science, N. S., vol. 22, no. 554, Aug. 11, 1905, pp. 180-183, 5 figs. 1905h. On a newly recognized nerve connected with the forebrain of selachians. With 32 figs. Anat. An.z., Bd. 2G, pp. 33-G3, 111 123, 1905. I'lNKLs, Felix. 1894. I'eber einen noch nicht beschriebenen Ilirnnerven des Protop terus annectens. Anat. Anz.. Bd. 9, no. 18, pp. 5G2-5GG, 4 Abb., 1894. 1895. Die Ilirnnerven des Protopterus annectens. Moriih. Arh., Bd. 4, zw. lift., pp. 275-346, Taf. XIII-XIX. Sewertzoff. a. N. 1902. Zur Entwickelungsgeschichte des Ceratodus forsteri. Anat. Anz.. Bd. 21, no. 21, Aug. 1902, pp. 593-008, 5 Abb. Sheldon. R. E., and Brookover, Chas. 1909. The nervus terminalis in teleosts. Rviiorl at BaJliniore meeting, Assoc. Amcr. Anat., 1909. 196 Journal of Comparative Neurology and Psychology. tv-te -^Ac-ra^olf.. UV. ■P,, Fig. 1. Transection tlirongli the middle of the right olfactory bulb of a carp 50 cm. iu length. Weigert method. X ^"^ (Zeiss oc. 2, obj. AA, reduced to two-thirds). Shows the uervus termiualis imbedded among the fibers of the tractus olfacto-lobaris medialis. Most of the stippled periphery is filled with the unmedullated fibers of the olfactory nerve which are ending iu glomeruli iu this region. An especially prominent mass of such fibers appears dorso-laterally forming a protuberance. The nervus termiualis is lost a short distance rostrad of this level among the fibers of the olfactory nerve; n. term., uervus termiualis; olf. nerve, olfactory nerve, fibers of which are scattered about the periphery at the points noted ; rad. olf hit., radix olfacloria lateralis; tr. olf. loh. iiied., tractus olfacto-lobaris medialis. eti\tKe\\a\ root Tr.o\f .loli.mcd rv.fcrm -raA o\^. lat, caA . o\f . me c} .Wob rr.o\f\olj. Tig.S. o\t. UK. Fig. 2, Transection (liruugli the middle of tlie right olfactory crus of a 50 cm. carp. Weigert metliod. X 21 (Zeiss comp. oc. S, ohj. A* reduced to twothirds), n. term., nervus terminalis ; md. olf. hit., radix olfactoria lateralis; rad. olf. mod. prop., radix olfactoria mediaiis propria; tr. olf. loh. incd., tractus olfacto-lobaris medialis. Fig. 3. Transection through tlie caudal part of the right olfactory crus of a carp 50 cm. in length. Weigert method. X ^^ (Zeiss comp. oc. G, obj. AA, reduced to two-thirds). This is a section innnediately rostrad of the cerebral hemispheres and shows essentially the same features as Fig. 2. 198 Journal of Comparative Neurology and Psychology. Fig. 4. Transection tbrongh the rostral part of the cerebral beuusi)beres of a 50 cm. carp. Weigert metbod. X 21 (Zeiss couip. oc. 8, obj. A* witb tbe pointer at lU. Reduced to tbree-fiftbs). Sliows tbe nervns terminalis l)artly imbedded in tbe tractus olfacto-lobaris medialis wbicb it bas passed tbrougb dorso-laterad. Compare its location in Fig. 3. Note tbat tbe n. term. now lies Itetween tbe tr. olf. lob. med. and tbe rad. olf. med. prop.; n. tcnii.. nervus terminalis; irid. olf. hit., radix olfactoria lateralis; nul. olf. vied, prop., radix olfactoria medialis projiria ; 1r. olf. loh. med., tractus olfactolobaris medialis; tr. istrio-tlial., tractus slrio-tbalamicus. Shrldon, N ervus Tennnialis in the Carp. 199 Fig. 5. Transection through the hemispheres of a carp 50 cm. long, Immediately rostrad of the anterior commissure. Weigert method. X 21 (Zeiss comp. DC. 8, obj. A*, with the pointer at 10. Reduced to three-fifths). Shows the nervus terminalis separated from the tractus olfacto-lobaris medialis preparatory to its decussation, com. interiiilb., commissura interbulbaris ; 11. term., nervus terminalis; rail. olf. lat., radix olfactoria lateralis; tr. olf. loh. mecl., tractus olfacto-lobaris medialis; tr. strio-tlial., tractus striothalamicus. 200 'Journal of Comparative Neurology and Psychology. --Lmedian cau\tv^ ^ j-^^^'-^^j-w-teft-^ n Te r m . ^H^l V. .' ii^ ^:^.V^-/itrr olf loir. med. j^ %feom interuulLr. Fig. 6. Transection through the rostral part of the anterior commissure of a carp about 35 cm. in length. Ramon y Cajal method. X 156 (Zeiss comp. oc. G, obj. 8 nnn. Reducecl to two-thirds). Shows the entire decussation of the nervus terminalis of the right side and part of that of the left. The cells of the nucleus of termination are not shown, com. interbull)., commissura interbulbaris ; n. term., nervus terminalis ; tr. olf. loh. med., tractus olfacto-lobaris medialis. Sheldon, Ncrvus Tcrmniahs in the Carp. 201 ^ ^/S.^.'^-' /< r&^g^ ^^-?j. i^^^_" e^a ' •'«*.^%» -///^^^ |:|^.lPf.*.olUolr.mc<^. »:' _^&i p*^ rf" T,g. 1 )<.AlKav.n«, M.ll. iVl. Fig. 7. Transection tliroiigh the rostral part of the anterior commissure of a carp about 40 cm. in lengtli. Toluitlin blue method. X l-'j*^ (Zeiss comp. DC. 6, obj. 8 mm. Reduced to two-thirds). Detail cells. X ^^G (Zeiss comp. DC. 18, obj. 4 mm. Reduced to two-thirds). All drawings made with a camera lucida at level of stage. This section is through the same region as Fig. G and shows the numerous small cells of the nucleus of termination of the nervus terminalis. Two of the cells are drawn at a higher magnification. The two tr. olf. lob. med. appear almost colorless in the preparation, but are here colored black in order to orient the group of cells with reference to Fig. G. THE CKITEEIA OF HOMOLOGY IN THE PERIPHEKAL NERVOUS SYSTEM. BY C. JUDSON IIEKRICK. From the Anatomical Lahoratunj of the University of Chicago. Even a cursory survey of the literature of comparative neurology reveals a confusion of usage among different authors who have described the same organ under different names or ap]ilied the same name to different organs. This confusion in some cases is so great that it is necessary to add to the name of tho.^part the name of the author whose usage is followed, in much .^^■i same way that zoologists add the name of the authority after .he name of every species. The confusion in many cases rests upon an imperfect knowledge of the facts ; but in others it arises from differences in the interpretation of commonly accepted anatomical and physiological data. In so far as the latter is the case it suggests the necessity for an analysis of the factors upon which homology rests and an attempt on the part of working anatomists to come to an agreement as to the relative value of these factors. It will, I think, be generally agreed that true homology always rests, in the last analysis, upon genetic relationship of the parts homologized. In the case of serial homology, or homodynamy, too, I doubt not that the principles of homology of organs from species to species will be found to apply with but small change in the meristic comparison of organs from segment to segment in a metameric body. This cardinal principle of homology requires that the parts so homologized must have had a common origin phylogenetically, or in the case of meristic organs have sprung from a common segmental type. No functional or structural similarity, however close, which has been brought about by convergence from diverse ancestral 'IHB Journal of Comi-arativb Neukolooy and PsycHOLOGy. — Vol. XIX, No. 2. 204 "Journal of Comparative Neurology and Psychology. conditions due to the action of similar environniiental agencies or any other cause can be regarded as having weight in determining homology. The question then narroAvs itself down to the problem of the recognition and evaluation of evidences of genetic relationship among organs. There is theoretically no limit to the diversity of the forms which homologous parts may show in their transformation from type to type in the course of phylogenetic history. So long as the sequence can be traced in unbroken series with no admixture of foreign elements into the organ complex the homology remains perfect. Practically, however, such ideal relations seldom prevail, for most organs are complexes of diverse tissile elements, some of which may disappear in the course of a long phylogenetic history to be replaced perhaps by others originally foreign to this organ. How far this process of substitution may be carried and leave the individuality of the organ unimpaired is certainly a debatable point. A peripheral nerve, for instance, may show extreme variation in its distribution without change in its functional composition and present no difficulties to the morphologist so far as its homologies are concerned. But a nciwe whose composition varies from species to species may be incapable of any simple morphological treatment and homology, even though its area of peripheral distribution is practically constant through the whole animal series. This would be the case if some functional systems represented in the nerve in the higher members of the scries can be shown not to have been developed out of those of the lower members, but to have entered the nerve as alien structures. Thus, the ramus lateralis vagi is a nerve of simple composition which is present throughout the Ichthyopsida, but with the widest possible variation in the details of its distribution. Its homology throughout the entire series is free from uncertainty in all but a very few cases. When, however, we find that the ramus lateralis accessorius of the facial nerve, which when present in teleosts typically runs an entirely separate course into the body, in some cyprinoid fishes is joined by an intra-cranial anastomosis to the ramus lateral vagi and that the two nerves pass into the body fused Herrick, Criteria of Homology. 205 in a single trunk, the homology is immediately disturbed. For, not only does the foreign admixture come from a different cerebral segment, but it is of totally different functional composition, connecting with a different type of sense organ peripherally and with a different cerebral coordination system. A survey of the phylogenetic history of the facialis nerve presents constantly recurring phases of the problem. This nerve was in primitive vertebrates a branchiomeric nerve, supplying a gillbearing segment and containing at least four components. The stages in its metamorphosis into a nerve supplying chiefly the superficial mimetic facial musculature of man can be clearly read by the comparative anatomist. In the course of this phylogeny some new components are added, some are totally lost and the survivors experience manifold changes of function and rearrangement of rami. Though the identity of .the nerve as a segmental unit is never lost, yet its perfect homology throughout the series is certainly open to discussion ; and the morphological position of some of the rami whose composition varies from type to type by reason of peripheral anastomoses with other segmental nerves, such as the trigeminus and glossopharyngeus, is still more ambiguous. I would suggest as a basis for further discussion the following rules for the fixing of homologies in the peripheral nervous system of vertebrates : (1) If a nerve is a member of a meristic series (cranial or spinal), the preservation of its individuality in the comparative . series of animals requires that its roots must come from the same segment or segments throughout the phylogenetic series. If there has been a shifting of one or more roots to another segment the homology is thereby to that extent impaired. (2) But if the nerve in question is a composite of roots from several segments (like the hypoglossus), the individuality of the nerve is not necessarily destroyed by a shifting of the whole series forward or backward, or by the inclusion of more or less segments of the same kind in the complex, provided the relations to adjacent segmental nerves are not fundamentallv altered. Neverthe 2o6 Journal of Comparative Neurology and Psychology. less such variations of this nerve cannot be regarded as iierfedly homologous with each other. (3) Perfect meristic or serial homology (homodynamy) requires that the members of the series shall repeat the same pattern, both in the components represented in the roots and in the peripheral distribution. Typical illustrations of this serial homology are seen in some of the segmental nerves of annelid worms and in some of the spinal nerves of lower vertebrates. The human body probably presents few such exact repetitions of a segmental pattern except in some peripheral rami of a part of the spinal nerves. (•i) In general, homology requires that the nerves concerned shall have similar segmental relations, similar components and similar distribution. (5) If, however, in the course of the phylogeuy a new component is differentiated within a given segmental nerve from a more ancient unspecialized element, this nerve does not thereby lose its homology with the primordial unspecialized nerve ; for the genetic relationship remaius unbroken. Accordingly, if it should prove (as seems probable) that the gustatory component of the facial nerve was in the early phylogeuy differentiated from the preexistent unspecialized visceral system of that segment, no disturbance of the homologies would result. (6) Peripheral rami which are defined primarily with reference to other peripheral non-nciTOUS organs, like the sciatic nen'e, may be regarded as homologous in different animals so long as they possess the same functional components and maintain essentially the same relations to the organs with reference to which they are defined, even though the segmental relations of the roots and plexuses from which they are derived may vary. (7) But if any peripheral nerve is a ramus by definition of a definite segmental nerve (such as the r. hyoideus facialis), its homology is not perfect unless in the types compared it is composed wholly of fil)ors derived only from its own segment. Any admixture of fibers from another segmental nerve to that extent destroys the honiologj^, no matter how perfectly the mixed nerve may follow the same course as the unmixed nerve. For instance, Herrick, Criteria of Homology. 207 in primitive vertebrates there is in front of tlie trigeminus a general cutaneous nerve belonging to a different segment, the profundus nerve. The profundus nerve is rarely preserved in the adult, though vestiges of it can be recognized in several selachians and ganoids, Avliere there is evidence that the profundus nerve has fused witli the ophthalmic rami of the trigeminus. These rami, therefore, are to be regarded as trigeminus jjlus j)rofundus nerves in all cases where it can be shown that the profundus elements are preserved. Similarly, in fishes branches of the lateral line and gustatory roots of the facialis often anastomose peripherally with trigeminal branches. It is evident that the mixed ramus thus constituted has no longer the same individuality as before. It cannot be classed simply with the trigeminus or facialis ; it is both. It should be given both names, or an entirely new name, or else some arbitrary rule should be laid down regarding the selection of a single name already current. The past usage in such cases has been most varied and confusing, and the confusion has, in many cases, been worse confounded by ignorance of the fact that there was any difference in the composition of the anastomosing rami, or l)y indifference to this fact even when recognized. The result is that to-day the synonymy of the rami of the cranial nerves of lower vertebrates, where such anastomoses are frequently and very diversely developed, is in worse confusion than that of the pre-Linnsean herbals. (8) A given ramus of a seginental nerve which contains more than one component is not perfectly homologous with a ramus of the same nerve in a different species which has a similar peripheral distribution, but lacks one or more of the components or has an additional component, even though the added component comes from the same segmental nerve. Thus, the hyomandibular trunk of the cod is not perfectly homologous with this neiwe in Menidia ; for the cod lacks the visceral sensory component of this nerve which is present in Menidia, though the other relations are practically identical. (9) From the preceding considerations it follows that the composition of every nerve and rannis must be accuratxsly known before its homologies can be understood. Dissimilar and unrelated func 2o8 'Journal of Comparative Neurology and Psychology. tional systems must never be homologized either in a phylogenetic or nijeristic series. (10) When the composition of a segmental nerve is fully known each root and its ganglion (in case of the sensory roots) should have a separate name and be treated as a functional and morphological unit. In determining the homologies of such a unit regard must be had primarily for the function which it performs, as determined by its terminal relations, i. e., the type of peripheral end organ and the location within the central nervous system of the primary nuclei of origin or termination of its fibers. These considerations take precedence over all others in doubtful cases. (11) The peripheral relations of nerves to other organs along their courses are also important in determining their homologies ; but resemblances in such relations must not outweigh differences in functional composition, where these two factors are in conflict. (12) No inflexible rules can at present be laid do^vn for the nomenclature of peripheral rami of mixed or variable composition. In the selection of names for peripheral nerves or rami priority should rule among competing terms, other things being equal. But if the prior term implies a false morphology, or is unnecessarily cumbersome or otherwise objectionable, or if it has been long obsolete, it may be discarded in favor of a better one or one better known. Peripheral rami of mixed composition can often be analyzed into an ancient or primary branch of one nerve and cenogenetic additions by peripheral anastomosis from a different segmental nerve. In such cases the ramus may be named as a branch of the nerve with which its palingenetic connection is made, even though it is not exactly homologous with the nerve so named in other species which lack the peripheral anastomotic addition. Thus the r. mandibularis trigemini is typically composed of general cutaneous and motor fibers. In some vertebrates gustatory fibers enter it by peripheral anastomosis from the geniculate ganglion of the facialis. The mixed nerve so formed may for convenience still be termed the ramus mandibularis trigemini, even though it is morphologically partly facialis, provided the imperfection of the homology is explicitly recognized. In the same way the lingual nerve of man may be assigned to the trigeminus. Herrick, Criteria of Homology. 209 in spite of the admixture of facialis fibers tlirough the chorda tympaiii, provided the trigeminal element can be sho^vm to be phylogenetically the older. The principles outlined above for guidance in determining homologies in the peripheral nervous system can be applied, mutatis mutandis, to tracts within the central nervous system. It will not be necessary to make the application here in detail. The treatment of the grey nuclei and correlation centers Avill also be controlled by similar rules, the aim being to homologize only such structures as are genetically related and to use functional cormections wherever possible as guides to homology. LITERAKY NOTICES. Margaret Floy Washburn. The Animal Mind. New York, The Macmillan Co., 1908. Pp. x + 333.$1.60. (Second volume of the Animal Behavior Series, edited by R. M. Yerkes.).

During the past few years the prohlems of animal behavior have attracted the attention of nnnierous zoologists and psychologists. The older "anecdotal" school has finally given place to a school of strictly experimental investigators. The result of a decade of experimentation is an accumulation of data which seems destined to provide a secure foundation for a science of comparative psychology ; but as yet these data are, in many instances at least, so fragmentary and so ill-organized that writers wholly fail to agree upon their interpretation. Several obstacles are encountered by the reader who attemiits to keep in totich with the worlv which is being done in this field. The investigations have been concerned, in the main, with circumscribed and isolated I)roblems ; and no thorough-going attempt has ever been made to correlate the various groups of experimental findings, or to present a systematic resume and interpretation. Then, too, the data are scattered through a great number of psychological and biological periodicals which are not readily accessible. Moreover, with the advance of scientific achievement in this field there has been developed a refinement of technic and of method which must, of course, be mastered before one can hope to evaluate the results or discover their significance. And, it may be added, the literature is replete with controversial clashes between opposing factions, who advocate a more mechanical or a more anthropomorphic interpretation of observations upon animal behavior.

This, in brief, is the situation which confronted Professor Washburn when she undertook to prepare a volume on "The Animal Mind." In her attempt to clear up the situation she sunnnarizes numerous investigations, evaluates their resiilts in the light of the experimental methods employed, and she discusses the bearing of these results upon the general question : What must be the characteristics of the animal mind, — granting that such a mind exists." The magnitude of the author's task may be inferred from the fact that she cites 4TG references from the literature ; and her presentation of the results of other investigators is but a small fraction of this exceedingly valuable contribution to the science of comparative psychology.

After an introductory discussion dealing with the difficulties and the methods of comparative psychology (pp. 1-2G), and with the evidences of mind (pp. 27-36), she proceeds to the specific question of the protozoan mind (pp. 37-57). The author's attitude toward her problem is illustrated by the following quotation (pp. 36-7) : "We know not where consciousness begins in the animal world. We know where it surely resides — in ourselves ; we know where it exists beyond a reasonable doubt — in those animals of structure resembling ours which readily adapt themselves to the lessons of experience. Beyond this point, for all we know, it may exist in simpler and simpler forms until we

212 'Journal of Comparative Neurology and Psychology.

reach tbe very lowest of living beings. * * * No one can prove the absence of consciousness in even the simplest forms of living beings. It is therefore perfectly allowable to speculate as to what may be the nature of such consciousness, provided that the primitive organisms concerned possess it." She does not present the arguments which may be advanced for and against the ascription of mental processes to the lower animals ; nor does she indicate how probable or how plausible is her assumption that the lower animals possess a consciousness. This may or may not be a serious omission, — probably many readers who have followed the discussions will agree that it it not. But it does seem paradoxical enough that an author should devote whole chapters to the description of something which, in the opinion of many reputable scientists, does not exist ; and whose existence the author herself is not willing to vouch for.

An examination of the motor reactions of ameba and Paramecium is believed to warrant the inference that the hypothetical protoizoan mind differs from tbe human mind in certain essential and clearly definable features. The mental stock-in-trade of the protozoon probably amounts to not more than three or four qualitatively different sensations; there is an utter absence of mental imagery (or revived sensations), and of anything correlate with attention. (This inference, however, does not seem to be justified, in the case of stentor, by Jennings' observations.) The mental life of the protozoa cannot therefore be a continuous "stream of consciousness," but only a succession of discrete and isolated experiences of the most primitive sort. From this humble beginning of mind, the author sketches in broad outline the developing consciousness through the entire animal series. "The reactions of animals to stimulation show, as we review the various animal forms from the lowest to the highest, increasing adaptation to the qualitative differences and to the spatial characteristics of the stimuli acting upon them. It is therefore possible to suppose that the animal mind shows increasing variety in its sensation contents, and increasing complexity in its spatial (and other) perceptions. But besides this advance in the methods of responding to present stimulation, the higher animals show in a growing degree the influence of past stimulation."

The author presents a detailed description of this increasing complexity in animal response to stimulation. Three chaptei^ on sensory discrimination trace the development of sensory equipment (pp. 58-147). This is followed by a discussion of spatially determined reactions and space perception (pp. 148-204). Here are considered the question of the adaptation of animal reactions to the spatial relations of stimuli (light, gravitation, and the like), and the question of animal perception of space. It is inferred that orientation in the lower animals is probably due to an experience of impleasantness or uneasiness, — and that no spatial perception need be assumed to account for the reaction. But certain responses to genuine visual stimuli (/. e., where eyes are present) may be due to a consciousness of spatial relations.

Chapters on "The Modifications of Conscious Pi'ocesses by Individual Experience" (pp. 20,5-209) describe the various labyrinth and puzzle-box exjteriments, and discuss the elimination of useless movements in the acquisition

Literary Notices. 213

of motor habits, llefo, and in the succeeding chai)tor on the "memory idea" (pp. 270-284), the author reports that an examination of the learning process in animals fails to discover any conclusive evidence for the presence of "ideas" (excepting perhaps in the case of monkeys). "The behavior of the lower forms of animal life, at least, can be fully explained without supposing that the animals concerned ever consciously recall the effects of a previously experienced stimulus in the entire absence of the stimulus itself." The closing chapter (pp. 285-294) deals with the biological significance of attention.

Professor Washburn's book is the pioneer in its field. It will unquestionably prove to be a time-saver to the student of animal behavior, and will be a welcome adjunct to the work oi the class-room. Many of its discussions are carried through with a thoroughness and an insight which render them of [laramount value. This is particularly true of the learning process. The treatment of spatially determined reactions and spatial perception and of tropisms is, however, in the opinion of the reviewer, too vague and too inconclusive to be of value to the student.

It is suggested that future editions could be improved by the addition of a more inclusive index. To cite but a few omissions, such important topics as memory, mental image, and experience are not mentioned in the author's index. The bibliography might be extended to include Wundt (writings since 1892), Sanford, Brehni, .Jourdan, and Ribot, and a complete list of the work done in this field by Darwin, Mcjebius, Wasmann and Watson.

J. W. Baied.

University of Illinois.

The Journal of

Comparative Neurology and Psychology

Volume XIX June, 1909 Number 3

ON SENSATIONS FOLLOWING NERVE DIVISION.

BY

SHEPHERD IVORY FRANZ. From the Laboratories of the Qovernment Hospital for the Insane,

Washington, D. O.

With Seven Figures.

II. The Sensibility of the Hairs.

Examinations of the sensibility of the skin from the standpoint of punctate sensibility indicate that the hairs are closely associated with those points that are stimulated by pressures and that react by giving a sensation of pressure or touch. The hairs are assumed to have a form of sensibility allied to or the same as that of neighboring parts sensitive to touches or pressures, although on thi's point most authors are silent.^ The experiments carried out by Head and Sherren indicate, however, that the hairs are independently sensitive to stimuli and that they react in a different manner than do the socalled pressure points. Numerous observations in cases of nerve lesions show that the sensations evoked by stroking or pulling the hairs differ from those of pressure and touch of the skin. For example, in an individual whose radial nerve has been cut the parts

'That this is probably not so is indicated by recent studies of the nerve endings (presumably sensory) in or near the hair bulbs of "touch" hairs in animals. The structures which have been found are different to those on hairless parts endowed with sensibility to touch and pressure.

The Journal of CowrARATivB Neurologt and Fsycholooy. — Vol. XIX, No. 3.

2l6 "Journal of Comparative Neurology and Psychology.

which remain endowed Avith nidtor nerves are sensitive to pressures and jarrings of the skin, but in his own case Head found that in this region "when the hairs are puUed the elevation of the skin produced no effect upon consciousness," land also', that "pressures which had previously caused a sensation were no longer appreciated when applied to the skin lifted from the subcutaneous structures to form a ridge."" It is evident, therefore, that the sensations produced by stimulation of the hairs are not pressures and do not belong to the class of "deep sensibility." On the other hand. Head and Sherren found that with the return of ]n-otopathic sensibility — the ability to appreciate prick as such and to respond to ice and to water at about 50° C. — "the hairs began to react to cotton wool, and this stimulus evoked a curious radiating sensation with a characteristic quality. True localization was impossible and the skin over the same parts became, when shaved, entirely insensitive to cotton wool."^ In another place we are told "under certain conditions the hairs may regain a peculiar form of sensibility at the time when the affected parts are sensitive only to prick and to the extremes of heat and cold. , Plucking a normal hair, will, in most cases, cause pain, and it is this sensibility to pain that returns to the hairs when they react in this manner to stimulation with cotton wool."^ These facts indicate that the sensations from the hairs are to be grouped not with the epicritic, although they react to cotton wool, but among the protopathic forms of sensation.

The subject of my experiments is an individual in whose arm the median and ulnar ner\'es had been cut about four months previous to the examinations.^ To assist in the definition of the area in which protopathic sensibility remained and from which the epicritic sensibility had departed, I carefully examined the hairy parts of

^'H. Head, W. H. R. Rivers and J. Slierren. The Afferent Nervous System from a new Aspect. Brain, 1905, vol. 28, p. 303.

"H. Head and J. Sherren. The Consequences of Injury to the Peripheral Nerves in Man. Brain, 1905, vol. 28, p. 241.

Head and Sherren. Op. cit., p. 242. 'For an account of the lesions and general sensihility changes see my article in Joiir. Cotnp. Neurol, and P.sijrhol., 1909, vol. 19, pp. 107-124.

Franz, Sensations FolUnving Nerve Division. 217

the forearm and hand, plucking individual hairs, and also brnshing the hairs with cotton wool or with the light earners hair brush. The results obtained on this patient are sufficiently different from Head's reports to warrant a i-ather full description of the sensibility of the hairs.

Over the parts of the hand and forearm which are quite normal, immediately a hair is touched there is felt a sensation, apparently similar to that when the skin is lightly touched with a blunt instrument, such as a pencil. This sensation results, 1 find, from the movement of the hair, and it is emphasized when more than one hair is moved or when more than one hair is grasped with forceps. After a hair has been tirndy grasped with forceps and slight traction is exerted upon it, the sensation becomes clearer, or more intensive, and when sufficient traction is exerted a distinct feeling or sensation of pain supervenes. The pain appears to differ in character from that produced by extremes of pressure, as for example that produced by an algometer, for it is rather burning in character. Observations such as these Avere made by H. (the subject of the experiments to be reported here) and b}^ me, and confinned by repeated experiments at various sittings.

On the other hand, as may be expected from the results of Head's experiments, over parts which are not normal the hairs react in a totally different fashion. On the \o\i\v side of H's forearm, near the bend of the elbow, I marked off an area which included subareas in which the different forms of sensibility were altered. First I examined carefully the sensibility of the hairs to traction. I went over the whole forearm and hand A\'herever hairs were found, and marked in red ink a line between the area which was sensitive and that which was insensitive to such stimulation. The hand and arm were then photogTaphed and from the photograph was traced Fig. 1, which is here reproduced. The area in which traction on the hairs was not accompanied by a pressure-like or by a pain sensation is that marked with vertical lines. The upper part of the arm, beyond the ellK)w, I did not carefully investigate, for it seemed that some of the change to be found there might be, and probably was, due to the cutting of superficial skin nen'es both at the time of the accident

2l8 Journal of Comparative Neurology and Psychology.

and at the subsequent operation. The extent of this area is much larger than the area of loss of protopathic sensibility as determined by the methods of Head, and the extensive character of the change led me to a more careful examination of parts of the area. Since time did not permit the careful mapping of the whole area, I selected for more careful work two areas, one on the volar side of the forearm near the bend of the elbow, the other on the dorsal-ulnar side of the arm near the wrist.

Fig. 1.

Fig. 1. — Diagram of hand and forearm of subject in whom uhiar and median nerves were cut at elbow. Scars of original accident and of operation shown above elbow. Area marked with vertical lines represents area in which hairs are insensitive to plucking. The diagram was made from a tracing of a photogi'aph, and the distortion of the hand is due to the peculiar position in which the patient holds the fingers.

An illustration of the upper of these two areas as it Avas marked in red ink on the arm of the subject is given in Fig. 2. The line separating the area of pain-on-pulling-hairs from that of no-pain-onpulling-hairs is approximately that between areas D and E. In general contour, however, it is more irregular and more nearly approaches the shape of the lowest line on the figure. The extent of this area in which pulling the hairs did not result in a sensation of

Franz, Sensations Following Nerve Division. 21 g

pain or in a pressure-like sensation was determined as accurately as possible. Once the general line was found, I went carefully over all the hairs within a centimeter of the line and mapped out the exact area in which plucking did not result in pressure-like or in pain sensations. Within the areas F and G, I found none of the hairs to be sensitive to plucking. Throughout this area and below G, the usual pressure-like and pain sensations were not produced by pulling the hairs, even when two or more hairs were pulled out with their roots. In area E the plucking of only a very few hairs on the upper

Fig. 2.

Fig. 2. — Diagram of upper inner part of forearm. Horizontal areas G, F. and E were insensitive to plucking of hairs. Below G pressures were not felt, likewise traction and brushing of hairs. Horizontal areas A, B, and C had all forms of sensibility intact. Areas E, F, and G were sensitive to brushing of hairs.

border of some of the squares resulted in sensations, while stimulation of some of the hairs at the lower border of area D was not accompanied by the proper kind of sensation. The extent of the lack of proper sensations in the hairs may then be said to be in the horizontal areas E, F and G. Although the plucking of hairs in area E was not accompanied by a normal sensation of pain and the pressure-like sensation, such as are produced by plucking the hairs over normal parts, when several, say five or six, were pulled at one time a feeling of an indefinable (to the patient) character was obtained. The subject

220 'Journal of Comparative Neurology and Psychology.

could not make any good comparison with any other known kinds of sensation, for it was different from pressure and was not like the sensations obtained by brushing or plucking the hairs over normal parts. The feeling (or sensation) could not be localized beyond the general (upper or lower) part of the arm. From numerous stimulations and repeated experiments the feeling or sensation was found to depend upon the movement of the skin in neighboring regions. When great care was taken not to move the skin in near-by regions, this unusual and ill-defined sensation did not result. Throughout the series of experiments the patient reported that, although he perceived the plucking of the hairs, it should be understood that at no time within the area of the diagram did the sensations have the same quality or character as that produced by plucking the hairs on the right arm. How much of this curious sensation difference was due to suggestion and how much to an actual change can not be determined, but it is fair to assume that the patient was ordinarily trustworthy.

After locating as accurately as possible the area in which the hairs were insensitive to traction, I mapped out the area in which the hairs were sensitive to stimulation with cotton wool. When, on normal parts, the hairs were lightly brushed with cotton wool, the sensation was immediately perceived, and an accurate localization was made of the place where the stimulus was given. ^Mien the hairs are stimulated as they lie, we find that many hairs from widely separated regions may bo stinmlatcd. This is pai'ticularly so if the hairs luip})en to be long and overlap each other to any extent. To determine with some precision the presence or absence of this form of sensibility, I carefully lifted the hairs overlapping any special part and stinuilated only those which were immediately beneath those that had been lifted. In carrying on the experiment in this way it was possible to locate quite accurately the extent of this kind of hair sensibility and the error of determining the line of division between sensitive and non-sensitive parts was not more than a few millimeters.

The results of this careful examination were quite unusual and wholly unexpected. T found the hairs to be sensitive to stimulation of cotton wool or of the liffht camel's hair brush in all the areas

Franz, Sensations Following Nerve Division. 221

above the lowest line on the diagram, Fig. 2. It will be noticed, therefore, that the fifteen square centimeters embraced in the horizontal areas E, F and G were sensitive to this form of stimulation but not sensitive to the stimulation of traction.

"\\n3en this large area was shaved, it was found that stimulation of the skin w^ith cotton wool or with a camel's hair brush was accompanied by sensation only in the horizontal subareas A, B and C. In these same areas other forms of epicritic sensibility, e. g., the appreciation of two-ness, were also present.

nodius

E

Fig. 3.

Fig. 3. — Back of forearm near wrist. Horizontal area A, loss of all forms of sensibility. Area B, hairs do not resi)ond to cotton wool or to traction. Area C, hairs react to cotton wool, not to traction with pain sensation, but with only pressurelike feeling. Areas D and E, hairs react to cotton wool and to traction. Area E, brushing hairs and traction on them felt more plainly than in any other areas. Area D and radialwards, area in which epicritic sensibility retained.

This unexpected result on the upper part of the arm led me to a further examination of another section of the arm ten days later. The second axea was on the outer and ulnar side of the forearm beginning about 8 cm. from the fold of the wrist and extending up the arm a distance of 5 cm. This part of the arm is shown in Fig. 3. Area A in the figure is the area insensitive to pressures. Area B is separated for convenience, but experimentally not shai*ply to be distinguished, from area A and is that part of the arm in which pressures usually, but not always, wore felt. The wavy line separates

222 'Journal of Comparative Neurology and Psychology.

the areas in which the hairs do (C) and do not (B) respond to stimulation with cotton wool. The dotted line separates the areas in which plucking the hairs produced pain and no pain. Areas D and E are areas in which the hairs reacted both to cotton wool and to traction. Area C is the area in which plucking the hairs produced a sensation similar to that of pressure or to that of brushing the hairs, but in this area no pain was felt even when the hairs were pulled out with their roots. In area B plucking of the hairs was accompanied by no pain and the hairs did not appear to be sensitive to any form of stimulation. At times in this area, as happened with the area near the elbow, when more than three or four hairs were pulled simultaneously a sensation was obtained. This was poorly localized, but appeared to be in neighboring more normal areas. Here also it was found that the movement of the neighboring parts appeared to be the determining factor in the production of the sensation. The sensation from this form of stimulation was described as similar to that of moving the hairs with cotton wool or with the camel's hair brush, and also partly like that of lightly touching the part with a blunt instrument.

Wben this part of the arm was shaved it was found that cotton wool could be appreciated in areas D and E. In these areas different degrees of temperature could also be appreciated and the subject reported marked sensation differences between cool and cold, and between warm and hot stimuli.

The above facts may be summarized as follows : WHien the ulnar and median nerves are cut, over parts of the hand and arm it is found that the hairs are not sensitive to plucking at a time when lightly brushing the hairs on the same areas is appreciated as a stimulus. The parts in which there is insensibility to plucking the hairs are within the area that, in accordance with Head's differential signs, may be described as possessing protopathic sensibility, but in which the epicritic sensibility has been abolished.

It appears, therefore, that we have in the hairs two forms of sensibility, one for traction and the other for light pressures or

'A full report of the tempera tiire findiugs in this portion of the arm will be found in Section III of this article, beyond.

Franz, Sensations Following N erve Division. 223

movements. The results, I believe, cannot be explained, as von Frey has attempted to explain all of Head's results,'^ as a difference in threshold values ; for in traction we deal with amounts of stimuli much greater (or at least on normal parts they appear much greater) than that of brushing the hairs with cotton wool. Such a two-fold function in the hairs is in ac-cord with the findings of a two-fold nerve supply to the hairs,* although the results on animals have not been confirmed for the common (bodily) hairs of man. So far as I am aware, the only results to be compared with these on the hair sensibility are those on temperature sensations to be reported in the next section and the few results by Head and Sherren on temperature sensations. In one or two cases these authors, it will be remembered, found parts of the skin not sensitive to hot (above 45° C.) and to cold (below 10° C.) objects, but found the same parts when stimulated with temperatures of moderate degree to be sensitive and to give appropriate warm and cool sensations.

III. Temperature Sensations. From examinations of normal individuals it appears that there are special points on the skin that react to stimuli by giving a sensation of heat or cold, but that in the small areas between these temperature points no sensations of hotness or coldness can be evoked by stimuli. On the other hand, when we stimulate the skin with areas of heated or cooled objects rather than with points the sensations of warmth or coolness result. The areal sensations appear to differ from those in which separate spots are stimulated in that only one sensation is obtained, and there is not an apparent mixture of heat and cold from the spots that may be stimulated in the area. For the understanding of these differences in sensation no explanation has been offered that meets with imiversal approval, but the analogy of the rods and cones in the retina has been made. It is

'von Frey. The Distribution of Afferent Nerves in the Slvin. Jour. Arner. Med. Assrt., 1906, vol. 47, pp. 645-648.

'F. Tello. Terminaciones sensitivas en pelos y otros organos. Trah. del. lah. de Invest. Biol, de la Univ. de Madrid (S. Ramon y Cajal), 1906, tomo 4, pp. 49-77.

224 'Journal of Comparative Neurology and Psychology.

said that if the area of the skin which is stimulated has a few spots that ordinarily give an intense sensation of coolness or warmth, the areal stimulation takes this character and that the spots which normally give a weaker sensation help to fill up and to make the areal stimulation continuous. In other words*, it is assumed that when more than one spot is stimulated, the general character of the resultant sensation depends upon the sensation that is most intense in the spots stimulated."

It will be remembered that in their examinations of sensations following nerve injury. Head and his collaborators found certain sensation losses that appear not to confonu with the hypothesis of special nerve endings for warmth and coolness alone. Certain of their results appear unaccountable on the supposition of loss of certain numbers of the fibers that supply the end organs concerned with the sensations of warmth and cold, — supposedly Ruffini's cylinders and Krause's end bulbs, respectively. Their work has cast considerable doubt on the singularity of the sensations of warmth and coolness and it appears from their examination of cases in which nerves have been injured or cut that there are two sets of nerx^es, four different fibers, which convey temperature sensations. The two sets belong respectively to the epicritic and protopathic systems, the former being concerned with medium temperatures, which are appreciated as warmness and coolness, while the latter mediates the sensations which may be spoken of as hotness and coldness.^"

The results of temperature experiments made by me on a patient (H.) in whose arm the median and ulnar nerves were cut confirm, in a general way, those reported by Head." In some particulars, however, differences were found.

Over parts which were insensitive to pressures no sensations of temperature were obtained, even from those temperatures which caused a burning of the skin. H. at one time placed his hand

On this see Titchener. Ea-pcrhiicntal Psi/cliologi/, vol. 1, part 2, pp. 87-91.

"Head and Sherreu. The Consequences of Injury to the Peripheral Nerves in Man. Brain, 1905, vol. 28, pp. 224-228.

"For an account of the patient and for other sensation differences, see Franz : this Joiinuil. vol. 19, pp. 107-124.

Franz, Sensations Follownig Nerve Division. 225

against a steam radiator and produced a burn about 1 cm. in diameter, without appreciating that his hand was in contact with a hot surface. This burn did not heal, as Head has noted in similar cases, as rapidly as burns on noruial parts, and it was over ten weeks before this area of the side of the hand t(K;)k on a healthy appearance. Duriug this time no pain or feeling of temperature could be obtained from this part of the hand. Temperatures as high as I dared use in the experiments, without causing a burn^ were not felt. In the same way low temperatures were not appreciated on this part of the arm and hand. A test tube, the temperature of which had been lowered to ■ — 5° C. was not felt and during cold weather the patient had to depend upon the sensations from the thumb and the radial part of the arm to determine when the hand should be covered.

In the areas of the hand and arm in which protopathic sensibility remained, the extremes of temperature were easily appreciated, although the medium temjjeratures did not call forth a sensation. In the area retaining also the epicritic sense, however, both extreme and medium temperatures were readily appreciated.

In the experiments with my subject, I used heated or cooled test tubes, 12.5 mm. in diameter with hemispherical bottoms. These were filled with water and in each tube a thermometer was inserted so that the temperature could be read directly after or before stimulating any part of the skin. The tubes were placed on the skin and pressed only Avith their own weight. It was found that at no place of stimulation did an area more than 8 mm. in diameter rest upon the skin. For cold sensations the test tubes were cooled by beingplaced in a mixture of ice and salt, and for the lowest temperature the tubes were cooled to — 5° C. For lesser degrees of cold, for temperatures of about 20° C. the tubes were immersed in cooled water. For testing for sensations of warmth and hotness the tubes were immersed in a Avater bath that almost completely covered the test tubes. Irregular orders were folloAved in determining the temperature sensations, and no indication Avas given the subject AA^hat the next kind of stimulation Avas to be. Moreover, the same square centimeter Avas never tested tAvice by the same stimulus in succession. The tulx^s Avere allowed to rest on the skin for only one or two sec

226 'Journal of Coynparative Neurology and Psychology.

onds. Immediately before the test tube was placed on the skin, the signal word Now" was given and after the test tube had been lifted, the subject reported Avhether or not he had felt anything and also the quality of the stimulation. The usual procedure of requesting judgments when no stimulus or when an indifferent stimulus was given was tried to see whether or not the subject guessed. After some preliminary trials, the subject used exclusively the terms "hot," "warm," "cool," "cold," "pressure," and "do not feel anything." The subject also often voluntarily compared the temperature sensations of two or more stimuli in order and these reports gave a clue to certain differences that will be reported later. Throughout the test it was impossible to keep from the subject the knowledge that he was being tested for temperature sensations, and whenever he reported that the stimulus was accompanied by a sensation of pressure only, he was asked to try to determine the character of the temperature. At these times he was able to make a judgment only when a stimulus was repeated, and after the test tube had been left on the skin from 10 to 60 seconds, until in all probability there had been time for radiation of the heat or transmission of the stimulus to more nearly normal parts.

A series of early experiments on the hand showed certain deviations from the results of Head and Sherren, and for this reason, two areas were selected for careful examination, those which had been previously used for the determination of the hair sensibility on the upper part of the forearm near the bend of the elbow, and on the lower part of the forearm near the wrist.^^

In the area near the bend of the elbow, each square centimeter was tested separately, but in irregular order that no suggestion might be given or olitained of the extent of the loss of temperature sensation, and each area was carefully gone over three times with each form of stimulation. The resultant sensations that were reported were, with two exceptions to be mentioned, the same in all three tests, and the uniformity is a striking evidence of the accuracy of the observations. In experiments in this area, no attempt was made

"Figures 2 and 8. itp. 219 and 221. illnstrato tho areas which were carefully tested.

Franz, Sensations Folloiving Nerve Division.

227

to keep the temperature of the test tubes constant beyond that of having them cold, cool, warm or hot to corresponding parts of a normal individnal, but temperature degrees were always noted and records made at the tinw?. In Fig. 4 is shown the area on the upper

1

2

»

e

^

^^

i^

s^

^^

^

^^

s^

^

\\

'^^

//

y

^

Tx

^

A

^

y.

y/,

y,

^

^

//

y

2*-lO*C.

y///y

y

y

^

//

7777 ///y

y

y,'

/y

Vy 7 /7 ///A//

// //

y

y

yy

y/>

y

^

^.

y

yy yy

ia*-22*C.

.■■ f . .i

30- ■40*0. •V5*-60*C.

Fig. 4. Fig. 4. — lUustraliug tlie seusatious accompauying temperature stimuli 011 tlie upper volar part of the forearm. Each of the four parts of the figure represents the same area of skin. The lowest line is the extent of the loss of pressure sensations. The line above area D separates, but not so sharply on the skin, the area of epicritic sensibility (A, B and C), from that of protopathic sensibility (D, E, F and G). Vertical lines, sensations of warmth. Horizontal lines, sensations of hotness. Diagonal lines running from top to the right, sensations of coldness. Diagonal lines running from the top to the left, sensations of coolness.

part of the foreann divided into its component square centimeters with the results from each kind of temperature stimulation.

In only two cases, with very cold stimuli, 2°-10° C, did the three answers differ and in both cases, once each in the two

228 "Journal of Comparative Neurology and Psychology.

squares Ei and E5, the stimulus was perceived as cold. The four parts of the figure represent the sensations from the stimuli which normally would be called cold (2°-10° C), cool (15°-22° C), warm (30°-40° C.) and hot (45°-60° C). The lines running downward from the left indicate sensation 'cold/ the lines running downward from the right indicate the sensation 'cool/ the vertical lines indicate the sensation 'warm/ and the horizontal lines indicate the sensation 'hot.'

With the lowest temperatures all the areas responded with some form of sensation. The areas A, B and C reacted uniformly with the sensation cold, and the square centimeter D5 also reacted in the same manner. Once each, as has previously been mentioned, E2 and E4 gave a feeling of cold, while in each of these squares two other similar stimuli were reported to feel cool. With the next grade of stimulus, cool, there was a wider distribution of cool feeling than there was of cold feeling with the lowest temperatures, the areas comprising A, B, C and 1), and of E and F the square centimeters marked 4 and 5.

The feeling of warmness was obtained over areas A, B and C uniformly with temperatures from 30°-40° C, and in D the subject reported a similar feeling but "only slightly warm." With stimuli from 45°-C)0° C, areas A, B and C reacted always to the stimulus as "hot," while areas D, E, F and G gave the feeling of warmth.

Areas A, B and C are areas in which hot is distinguished from warm and cold is distinguished from cool. Areas E, F and G are areas in which only hot or crtld stimuli are invariably distinguished, although area D5 reacted to cold, areas E2 and E4 reacted once to cold, Dl — 5 and also E4, E5, F4 and F5 reacted to cool stimuli. Area D reacted to warm stinuili with an appropriate reaction, but the intensity of the stimulus appeared to be less than in the neighboring area C. With the hot stimulus areas A, B and C reacted in a normal manner, while the other areas reacted to the samje stimulus with the sensation of warmth. It will be noted that in the area in which the epicritic sensibility remained intact, coolness and coldness and warmness and hotness w^ere always distinguished. This comprised the horizontal sections, A, B and C.

P RANZ, Sensations Folloiving Nerve Division.

229

The area near the wrist gave similar results, although in the experiments of this region I did not map out the area according to squares and have to oflfer only the general results on the horizontal areas. In this case, however, I was careful to keep the temperature of stimulation constant for each area, and each time the stimulus was applied it was of the intensity noted on the diagram. The results of these experiments are shown in Fig. 5. Five dif

-5°C.

lOX,

20X.

Fig. 5. Fig. 5. — Skin area 011 ulnar aspect of forearm near wrist. Area A, no pressure sensations felt. Areas B and C pressures felt and protopatliic sensibilitj' retained. Areas D and E. pressures and touch appreciated and epicritic sensibility intact. For explanations of sensations of temperature, see legend to Fis. 4.

ferent degrees of the stimulus were chosen, — 5°, 10°, 20°, 40° and 60° C. The coldest stimulus felt cold in areas D and E, cool in area C. A temperature of 10° C. felt cold in areas D and E, but did not call forth a sensation of temperature when placed on area C. 20° C. was reported cool in areas D and E, but was indifferent in the other areas. Stimuli of 40° C. were felt warm in areas C, D and E, but indifferent in areas A and B. In area C the subject

230 "Joiirnal oj Comparative Neurology ami Psychology.

noted that this temperature was just warm," "only slightly warm," etc., not so warm a sensation as that given by the stimulus in the areas D and E. Likewise with a stimulus of 60° C. The area in which warm sensations were obtained were B and C, but the sensations from area B were not so intensive as tliose in area C. This temperature was felt in areas D and E always as hot.

The parts of the arm which did not retain their ability to appreciate pressures, area below G in Fig. 4, and area A in Fig. 5, did not respond at any time to any, temperature stimulus. Areas G, F, E and D in Fig. 4 and area B and C in. Fig. 5 may be said to have retained the protopathic sensibility in addition to the deep . sensibility, and in these areas hot stimuli were felt to be only warm, while cold stimuli w^ere felt to 1)0 only cool. In the upper forearm area D-G there was no response, as a rule, to the intermediate degrees of temperature. The remainder of the area A-C responded accurately to all degrees of temperature stimuli. In the area near the wrist, only D and E reacted well to all degrees of temperature, and showed the presence of epicritic sensibility. Areas B and C failed to respond to medium degrees of temperature and their response to the extremes was not well marked.

At different times during the examination of these two arm areas, the following experiment was tried : The test tube, instead of being placed with its end on the skin, was placed so that three to four centimeters of its length extended over the horizontal areas which gave such widely different results. In these experiments the subject described the sensations which were produced and the accounts are in accord with the observations made when the small horizontal areas were stimulated by the end of the test tube. ITear the wrist when cold, ■ — 5° C, w^as used, the subject reported that toward the radius the sensation was of extreme cold, but near the axis of the arm the sensation w^as only cool. In a similar way the sensation from 60° C. was described as hot near the radius and warm near the axis of the wrist. The places from which respectively cool and wann sensations w^ere obtained when cold and hot stimuli were given were pointed to by the subject and they correspond closely to the area C. This area it will be rememibered was found by the previous

Franz, Sensations Folloivtng Nerve Division.

231

experiments to give these sorts of sensations with the extremes of temperature. On the upper part of the forearm simihxr results were obtained. In this case the subject was permitted to keep his eyes open and to mark with his other hand the places where the stimuli changed in quality. The marking of the division between the areas was not clear and distinct, but between two points, about two centimeters apart, the difference in the sensations was reported to be marked. These results are in accord, therefore, with those of stimulation of the individual square centimeters and they strengthen the impression that the sensation difference that was reported in the first series was not due to radiation or conduction. These results are quite unlike the results in normal individuals, for in the latter

Ficx. 6. Fig. 6. — Area insensitive to cold stimuli, compared with the loss of pressure sensations. Area insensitive to cold inclosed within heavy line. Area marked with vertical lines insensitive to pressures.

we find the most intense sensations at the places where the warm or cool object leaves the skin, and the intermediate area appears to be stimulated uniformly with the same temperature.

On the hand, experiments similar to those on the arm were made, but with not so many different temperatures. The results of these experiments are given in Figs. 6 and 7. In Fig. 6 is shown the area of the hand insensitive to cold stimulation. This included the ulnar quarters of the palm and the back of the hand, the whole of the fourth and ring fingers, the first two joints of the second and index fingers and about half of the thumb. For comparison, the area insensitive to pressure with a pencil is illustrated on the same diagram

232 'Journal of Comparative Neurology and Psychology.

by the use of vertical lines. In only a small portion of the third finger does the area of insensitivity to pressure go beyond that of insensitivity to cold. Fig. 7 gives in a similar way the area insensitive to hot stimuli. For compatison, with the areal loss of cold and hot sensations, the area insensitive to light touch is drawn. On the back of the hand and on the index finger the area of insensibility to light touch is greater than that to heat, while on the thumb the area for hot sensation loss is greater than for light touch. With the exception of a very slight portion of the thumb the area insensitive to cold is included within that insensitive to light touch. In a general way, therefore, these results, as has already been

Fig. 7. Fig. 7. — Area insensitive to liot stimuli, compared witli loss of sensations to light touch. Area insensitive to hot stimuli inclosed within heavy lines. Area marl<ed with horizontal lines insensitive to light touches (cotton wool and camel hair brush).

said, confirm the obsen^ations of Head. The facts that do not agree with those of Head are : There are different extents of areas responding to different degrees of temperature ; the area for the appreciation of hot stimuli, for example, not being the same as that for the appreciation of cold stimuli as such ; temperatures of extreme degree are felt like medium temperatures or call forth sensations of warmness or coolness on areas which do not respond to the medium temperatures with a corresponding sensation of warmth or coolness.

The results obtained from the arm may be taken to mean that whenever a cold or hot stimulus was given there was a radiation of effect from the stimulated area to the neighboring areas. This

Franz, Sensations Fol/muing Nerve Division. 233

explauation would also account for the less widely felt sensations from the medium degrees of temperature when the individual square centimeters were separately stimulated. Such an explanation would not account, it seems to me, for the results which were obtained when an extended line was stimulated rather than a small area. It appears more probable that, as Head contends, for temperature sensations we have two sets of nen^es, one of which responds to the extremes of temperatures and the other to the medium temperatures. These two sets of nerves correspond to the fonns of sensibility that are called respectively protopathic and epicritic, but in a different way than that described by Head and his co^-workers. The results from my subject show that, although hot and cold stimuli produce sensations in areas endowed with the protopathic fomi of sensibility, the sensations correspond to those produced in normal areas by stimulation with medium temperatures. This at first sight looks as if we were dealing solely with differences in threshold values, but this hypothesis does not account for another apparently anomalous condition which was found by Head and Sherren, viz., the loss of the ability to sense the extremes of temperatures wdth the retention of the ability to sense medium temperatures.

It seems to me that the temperature results, taken in connection with those which have been described in my articles on the 'Pressurelike' and 'Hair Sensibilities,' show there is an overlapping of the nerves, or rather there is an overlapping of nerve supply, and that the sensation differences which were found in this case are to be explained as due to the presence or absence of certain nerve endings or of nerve fibers. It is j)ossible that the sensations of coolness and warmth are protopathic, w'hile those of coldness and hotness are epicritic. This, it will be observed, is contrary to Head's belief. Received for piiblieation December 7, 1908.

NoTE.^Since the articles on sensations following nerve division were written I received the number of Brain in which Rivers and Head give the results of careful examinations made on Head's arm in which the radial nerve was divided near the elbow (W. H. R. Rivers and Henry Head. A Human Experiment in Nerve Division. Brain, 1908, vol. 31, Part 123, pp. 323-450). Some of tlie new observations reported by Rivers and Head indicate a condition on Head's arm similar to that found by me on H., but some of these

234 'Journal of Comparative Neurology and Psychology.

anomalies are not sufficiently discussed. A full account of tliis recent work, witli criticisms of special parts of it, will be found in a forthcoming number of the Journal of Philosophy, Psychology and Scientific Methods. In this review I shall make a careful comparison of the present work with that of Rivers and Head, and with that of Trotter and Davies {Journal of Physiology, London, 1909, vol. 38, pp. 134-246), which has also appeared since my articles have been in type. At present I wish to call attention to only a few matters reported in these two recent articles.

The recovery of sensibility to light touch is considered by Rivers and Head to be dependent upon the regeneration of the so-called touch spots. These touch spots are, according to these authors, grouped about the hairs and much, if not all, of the sensibility of the hairs depends upon the presence of the nerve endings which make up these touch spots. They say, for example, "Over hair-clad parts these touch spots are strictly associated with the roots of the hairs, they express the sensibility to mechanical stimuli to that part of the hair which lies beneath the surface of the skin. Almost every hair is a delicate tactile sense-organ ; any movement of its tip is transmitted to its root with the increasing power of a lever, setting up tactile impulses." We find, however, they report in one part of their paper that tactile sensations (sensations to light touch) could be evoked only after a period of 365 days following the operation, but in another place we are informed that "No sensations were obtained from the hairs until 86 days after the operation, when there were found four hairs that gave a sensation when they were plucked." It is also noted that 161 days following the operation the hairs on the arm were sensitive to the stimulation of brushing with cotton wool, although the sensibility to light touch did not return for 365 days. Surface indications are that Rivers and Head were dealing with the same form of dissociation of the hair sensibility which has been described by me in the foregoing papers, but that they did not carefully investigate this matter. At any rate, from their account of the sensibility of the hairs we are justified in assuming that the hair sensibility was found by them to be independent of the presence of distinct touch spots and that their observations upon the sensibility of the hairs, casually reported, support the view expressed in the foregoing paper.

According to Rivers and Head all temperature sensations also depend upon sensation spots, and the difi'erences in temperature sensations described by Head and Sherren in a former paper are now reported to be due to the presence of a greater or less number of cold or hot spots in the epicritic and protopathic areas respectively. Head still holds the view that in an area in which there is no epicritic sensibility only the extremes of temperature will be appreciated, while in the area endowed with epicritic sense there is the ability to appreciate the intermediate degrees of temperature. It Is not clear that Head had sensations similar to those of my subject, but we read "It would seem that the number of spots stimulated is of greater Importance than the intensity of cold by which the sensation is evoked." In experiments in which his arm was tested by cold areas of different size Head reported a cool large area to feel colder than an ice cold small area, in the former case there being a number of spots stimulated and in the latter only one.

Franz, Seiisations Following Nerve Division. 235

On account of the fewness of the spots in the protopathic area it is difficult to see how the protopathic skin would give the hot and cold sensations even from the stimulation of comparatively large areas and how the sensations could be of the same intensity as those from normal parts. Head's observations on the temperature sensations are at variance with those reported by me, but it should be remarked that Trotter and Davies were not only unable to confirm Head's observations, but that, in fact, they obtained results similar to those reported here.

The matter of the sensibility to light touch needs hold our attention for only a brief time. The gradual increase in sensitivity from the anesthetic area outwards has also been demonstrated by Trotter and Davies in even a more convincing manner than that given by me. They have shown that the sensibility of individual touch spots differs at different times during the period of recovery, and it is very plain that the sensory disturbances following the section of a nerve are more widespread than Head and his collaborators admit. It is of some interest to note that a criticism of the cotton wool method of testing light touch has also been made by Trotter and Davies and that for similar tests they used a brush.

April 26, 1909.

MODIFIABILITY OF BEHAVIOE IN ITS KELATIONS

TO THE AGE AND SEX OF THE DANCING MOUSE.

ROBERT M. YERKES.

From the Harvard Psychological Laboratory.

With Four Figures. CONTENTS.

PAGE

I. Introductory statements : the dancer as material for the investigation of problems of beliavior 237

II. Relation of age and sex to rapidity of acquisition of a visual discrimination habit 238

III. Sensitiveness to electric stimulus, in its relation to age and sex. . . . 249

IV. Strength of electric stimulus, in its relation to rapidity of habit formation 252

V. Relation of difficultuess of discrimination to rapidity of habitformation at different ages 255

1. Experiments w^ith cardboards in discrimination box 255

2. Experiments with discrimination box in dark-room 259

3. Experiments with one side of discrimination box covered

in varying degrees 263

VI. Relation of age to rapidity of acquisition of labyrinth habits 265

VII. Conclusions and summary 267

I INTRODUCTORY STATEMENTS : THE DANCER AS MATERIAL FOR THE INVESTIGATION OF PROBLEMS OF BEHAVIOR.

The dancing mouse is well adapted, by its abundant and in certain respects peculiar activity, to experiments on behavior. Taking advantage of this fact, 1 have used it extensively as material for the development of methods and the revelation of problems, both physiological and psychological. That the results which have been obtained are typically mammalian I am not prepared to assert. This, however, is, for my immediate purposes, secondary in importance to the methodological values of the work. Animal psychology is urgently in need of exact methods of research. It is an appreciation

The .TorjRNAL of Cojipat;ativi<; Nbiikology and rsvcriOLOGY. — Vol. XIX, No. 3.

238 Journal of Comparative Neurology and Psychology.

of this fact that has shaped my experimental work during the past five years, and that now leads me to offer the following results of my study of the dancer primarily as a contribution to the evolution of method and as an aid to the profitable formulation of problems.

This paper is a direct continuation of the studies in the behavior of the dancer which are described in my book, "The Dancing Mouse."^ Although I have attempted so to write the paper that both methods and results shall be intelligible to those readers who are not familiar with the details of previous publications,^ it has been necessary — in order to keep my account within reasonable space limits — for me to omit everything except the chief points, in connection with methods which I have previously described, and a concise statement of new results. In other words, I have been forced to assume much more knowledge on the part of the reader than I should if this were my first publication on the subject.

Certain problems concerning the relation of age and sex to habit-formation which were proposed in my book, and either left unsolved or only partially solved, are brought nearer to satisfactoi'v solution by the results herein reported, and a multitude of new problems are revealed. To me, however, the investigation presents itself simply as another step toward a realization of the complexity of the phenomena of behavior and of the need for accurate analytic methods.

II. RELATION OF AGE AND SEX TO RAPIDITY OF ACQUISITION OF A VISUAL DISCRIMINATION HABIT.

Can the dancer acquire a given habit with the same rapidity at different ages ? This question was the starting point of a study of plasticity which has already been reported in part.^ Before presenting the results of my experiments I shall very briefly, with the help of figures which are reproduced from an earlier paper, describe the method of work.

'Yerkes, Robert M. The Dancing Mouse : a study in animal behavior. New York, The Macmillan Company, 1907. xxi + 290.

'Yerkes, Robert M. and Dodson, John D. The Relation of Strength of Stimulus to Rapidity of Habit-formation. Jour, of Comp. Islcur. and Psy., vol. 18, p. 459-482, 1908.

•The Dancing Mouse, pp. 270-215.

Yerkes, ModifiahiUty of Behavior

239

The habit whose formation was studied quantitatively, in the case of groups of dancers consisting of five pairs each, for the ages of one month, four months, seven months, and ten months, may be called the white-black discrimination habit. It involved the discrimination

Fig. 1. Fig. 2.

Fig. 1. — Discrimination box. W, electric box with white cardboards ; B, electric box with black cardboards.

Fig. 2. — Ground plan of discrimination box. A, nest-l>ox ; B, entrance chamber ; W W, electric boxes ; L. doorway of left electric box ; R, doorway from right electric box to alley; O, swinging door l>etween alley and A; IC, induction apparatus ; C, electric battery ; K, key in circuit.

of the entrances to two boxes, one of which was white and the other black, and the entering of the white box. Any attempt to enter the black box was punished by an electric shock.

Figures 1 and 2 show the experiment box in perspective and in

240 'Journal of Comparative Neurology and Psychology.

ground plan, respectively. The subject, after being placed in the nest box, A, by the experimenter, was permitted to pass into the entrance chamber, B. Then a piece of cardboard, which was placed between the animal and the opening into A, was slowly moved toward L, E, of Figure 2. Thus the dancer was brought face to face with the tAvo entrances, L and R, of this figure (B and W, i. e., black and white of Figure 1). One of these it would soon attempt to Qnter in order to escape to the nest box, and thus find space for dancing. If it started to enter the black box (and this might be either the box on the left, L, or the box on the right, R, for the white and black cardboards, which were at the entrances and within the boxes, could be transferred readily by the experimenter) it was immediately given a weak electric shock by the closing of the key, K. This usually caused it to retreat from the box and to try the other entrance. In case it entered the black box in spite of the shock, it was not permitted to escape by way of E and to the nest box, but instead was forced to return to B and again make choice of an entrance. This was continued until the white box was chosen, then the animal was allowed to return to A. After an interval of one or two minutes it was given another opportunity to select the right entrance. This was continued until the white box had been chosen ten times. Such a group of ten trials constitutes what we shall refer to as a series. One series was given each individual daily from the beginning of experimentation until the acquisition of a perfect habit of discrimination and choice.

The positions of the white and black cardboards were changed in precisely the same way for each individual according to an order which has already been described.* These shifts in the position of the white box were made in order to prevent the mouse from acquiring the habit of going regularly to the entrance at the left or at the right.

An experiment (test or trial) was recorded as yielding an error of choice if the mouse entered the wrong box far enough to get a shock ; as yielding a correct choice if, without first entering the black box,

Jour. of Comj). 'N cur. and Psy., vol. 18, p. 461, 1908.

Yerkes, Modifahility of Behavior. 241

it entered the white one and passed through to the nest box. In the tables appear the number of errors per series made day after day by the various individuals. At the outset of the experiments each mouse was given two series of what may be called "preference tests." In connection with these tests no electric shock was given and the mouse was permitted to enter and j)ass through either the white or the black box, for it was the sole purpose of the experimenter to discover, by means of these series, any initial preference that the subject might have for either the white or the black box.

A habit of discriminating between the boxes, and of uniformly choosing to enter the white one, was considered perfect when the mouse made no errors in three successive daily series. As a measure of the rapidity of habit-formation we may use the number of tests between the beginning of the first training series (following preference series B) and the end of the series which preceded the three perfect series. This measure of rapidity of learning, which I have named the index of plasticity, proves to be extremely useful for purposes of comparison.

To ascertain age differences in rapidity of white-black habit formation I used groups of individuals which, so far as I could tell, differed from one another constantly only in age. Five males and five females constituted each group, and four such groups were used. During their lives all of the animals were kept under the same conditions. They were paired at the age of twenty-five days, and thereafter a male and a female were kept in a separate cage and were placed in the experiment box for their daily training at the same time and given their tests alternately.

We may now examine the results of the experiments. Table 1 contains records of the number of errors of choice made by each of the individuals of the one-month-old group in each daily series. The numbers at the top of the columns refer to the mice. Even numhers always designate males; odd numbers, females. The two preference series are indicated by the letters A and B. ITo. 210, it will be noted, made six erroneous choices in each of the preference series and also in the first training series ; that is, he attempted to enter the black box instead of the white box six times in ten. In subsequent training ae

242 "Journal of Comparative Neurology and Psychology.

TABLE 1.

Relation of Age to Modifiability of Behavioe White-black Discrimination Habit

Residts for

dancers one month old

Males.

Females.

Series.

210

250

252

254

410

Average.

1 1 215 1 249 j 251

253

415

Average.

A

6

6

2 i 7

6

5.4

8

■ 5 1 6

4

8

6.2

B

6

3

5

6

5

5.0

8 6 5

5

6

6.0

1

6

7

5

5

2

5.0

7

6 6

3

6

5.6

2

4

1

4

4

2

3.0

5

3 4

2

2

3.2

3

3

3

2

5

3

3.2

3

4 ; 2

3

3

3.0

4

5

3

4

1

3

3.2

2 114

3

3

2.6

5

3

3

5

1

1

2.6

114

2

3

2.2

6

2

4

2

1

1.8

2 3 2

1

1.6

7

1

1

1

2

1

1.2

1 11

1

3

1.4

8

1

1

0.4

1 1

1

2

1.0

9

1

1

0.4

1 2

2

1

1.2

10

1

0.2

1

0.2

11

1

1

0.4

12

n

2

0.4

13

' 1

0.2

14

15

16

1

1

TABLE 2.

Relation of Age to Modifiability of Behavior White-black Discrimination Habit

Results for

dancers four months

Id

Males.

Females.

Series.

76

78

114

122

126

Average.

75

77

111

115

117

Average.

A

7

7

3

5

6

5.6

4

8

8

6

5

6.2

B

8

6

4

6

8

6.4

6

5

4

7

5

5.4

1

5

5

5

7

6

5.6

5

5

4

6

6

5.2

2

5

4

5

7

5

5.2

2

2

4

4

2

2.8

3

4

5

4

5

5

4.6

2

5

3

3

3

3.2

4

3

4

5

3

3

3.6

1

1

5

4

a

2.6

5

5

2

4

3

3

3.4

1

3

5

2

2.2

6

3

2

5

1

1

2.4

1

5

2

1.8

/

2

1

4

4

1

2.4

1

2

2

3

2.2

8

5

1

2

4

2

2.8

1

1

0.6

9

1

3

2

2

2

2.0

1

1.2

10

1

2

1

1

1.0

1

1

0.6

11

1

1

3

1

2

1.6

2

0.6

12

1

1

1

0.6

0.2

13

1

1

0.4

1

0.4

14

15

1

0.2

16

17

18

Ykrkks, M Oil ip ability of Behavior.

243

ries the nunilter of errors made by this iii(livi(hial ra])iilly decreased until in the seventh series only one was made. Then followed three perfect series. For this individual, since he acquired a perfect habit as the result of seventy training tests, the index of plasticity is 70.

The tables contain, in addition to the individual results, the average number of errors per series for the males and for the females.

TABLE 3.

Relation of Age to Modifiability of Behavior White-black Discrimination Habit

Results for dancers seven months old

Males.

Fem.vles.

Series.

92

96

98

116

120

Average.

91

93

99

101

109

Average.

A

5

6

5

7

6

5.8

4

4

7

6

6

5.4

B

7

4

7

3

5

5.2

6

6

7

5

7

6.2

1

4

3

5

7

5

4.8

5

6

3

8

5.8

2

4

5

6

5

8

5.6

2

3

7

6

2

4.0

3

4

3

4

3

5

3.8

4

4

4

4

6

4.4

4

7

5

4

5

3

4.8

4

3

3

6

4

4.0

5

3

4

5

4

7

4.6

5

3

5

2

3

3.6

6

3

5

2

4

4

3.6

5

2

4

2

2

3.0

7

3

1

1

4

4

2.6

1

4

4

3

2

2.8

8

6

2

3

2

4

3.4

2

4

2

2

3

2.6

9

2

1

3

5

5

3.2

1

5

1

1

1

1.8

10

5

1

3

4

5

3.6

2

2

3

1.4

11

1

1

1

3

1

1.4

1

1

1

1

0.8

12

2

2

1

4

1

2.0

1

1

2

1

2

1.4

13

2

1

3

2

1.6

2

3

1

1.2

14

4

2

1

3

2

2.4

1

1

0.4

15

2

1

1

0.8

1

1

1

0.6

16

1

2

0.6

1

0.2

17

1

2

0.6

1

0.2

18

1

1

0.4

19

1

0.2

20

1

1

0.4

21

1

1

0.4

22

1

0.2

23

1

0.2

24

1

0.2

25

26

27

Any one who compares this account of my investigation of the relation of age to rapidity of learning with my earlier account will discover that only two pairs of dancers of one month of age for which results were given previously^ have place in the group under discussion. This is due to the fact that I felt it highly desirable to repeat the experiments with one-month individuals in order to make

■'J'lit' I»;iiicin^ .Mouse, jiji. '24'.', -7'.

244 journal of Cojuparative Neurology and PsycJioIogy.

sure that in the interval which had elapsed between the beginning of this portion of my work and its completion no important changes in the plasticity of the race had occurred.^ As a matter of fact this precaution proved unnecessary, for no important differences appeared as the result of the interruption of the investigation.

The condensed results for the four-month individuals appear simi TABLE 4.

Relation of Age to Modifiability of Behavior White-black Discrimination Habit

Results for

dancers ten and twelve months old

Males.

Females.

Series.

90

112*

142

144 5

196 Average.

1

I 1 97* 113* I 119

1

123

141

Average.

A

6

6

5

6 5.6

5 8

7

5

4

5.8

B

5

5

'6

6

5 5.4

5 7

1

6

5

4

5.6

1

4

4

7

4

5 4.8

7 6

5

6

4

5.6

2

6

4

5

4

3

4.4

4 3

7

4

5

4.6

3

7

3

5

7

4

5.2

4 6

8

7

3

5.6

4

3

4

5

5

5

4.4

7 3

5

5

3

4.6

5

5

4

5

4

1

3.8

4 2

3

2

6

3.4

6

. 3

4

1

7

3

3.6

2 2

1

1

5

2.2

7

3

5

3

3

4 i 3.6

2 3

1

7

2.6

8

3

2

2

5

4 i 3.2

2 ' 1

2

1

5

2.2

9

4

3

2

4

1 i 2.8

1 1

1

1

4

1.6

10

3

4

1

1

' 1.8

1 10

1

1

0.8

11

1

1

1

1

! 0.8

1

1

2

0.8

12

1

2

1

1

1 1.0

2

1

2

1.0

13

2

1

1

! 0.8

1

0.2

14

3

1

1

1.0

1

1

0.4

1.5

1

1

1

0.6

1

0.2

16

1

1

0.4

!

1

0.2

17

1

1

0.4

1

0.2

18

1

0.2

19

1

1

0.4

20

1

0.2

21

22

23

Twelve months old. larly in Table 2, and those for the seven-month individuals in Table 3. In the ten-month group (Table 4) I have included the results for three mice which were twelve months old. Although it is not wholly satisfactory to do this, it seemed better than to deal with the three individuals separately. At any rate nothing is concealed by averaging the results for the ton mice, for the individual results are available.

To make comparisons easier, I have brought together in Tal)le 5

"An epidemic \Yliicli destroyed almost all of my mice caused a delay of over a year.

Ykrki-s, Modifiahi/ifv of Behavior.

245

the averages for the males and females of each gi-oup. This table presents also the general averages for each sex. Inspection of these results reveals the following significant facts.

(1) The females exhibit a stronger initial preference for the black box than do the males. Both, how^ever, choose the black box more frequently than the white box, in the preference series. Since

TABLE 5.

Generai> REStTLTS OF THE Stuuy of the Relation of Age to Modifiability

OF Behavior.

EacJi result hi the table is either the avcniiie uiinilter of errors in

ten tests for fire iuillriduals, w the general average

for twenty hnUmrluals.

Male.s.

1 mo.

Females.

Series.

1 mo.

4 mo.

7 mo.

10 mo.

Gen. Av.

4 mo.

7 mo.

10 mo.

Gen. Av.

A

5.4

5.6

5.8

5.6

5.60

6.2

6.2

5.4

5.8

5.90

B

5.0

6.4

5.2

5.4

5.50

6.0

5.4

6.2

5.6

5.80

1

5.0

5.6

4.8

4.8

5.05

5.6

5.2

5.8

5.6

5.50

2

3.0

5.2

5.6

4.4

4.55

3.2

2.8

4.0

4.6

3.65

3

3.2

4.6

3.8

5.2

4.20

3.0

3.2

4.4

5.6

4.05

4

3.2

3.6

4.8

4.4

4.00

2.6

2.6

4.0

4.6

3.45

5

2.6

3.4

4.6

3.8

3.60

2 2

2.2

3.6

3.4

2.85

6

1.8

2.4

3.6

3.6

2.85

i!o

1.8

3.0

2.2

2.15

7

1.2

2.4

2.6

3.6

2.45

1.4

2.2

2.8

2.6

2.25

8

0.4

2.8

3.4

3.2

2.45

1.0

0.6

2.6

2.2

1.60

9

0.4

2.0

3.2

2.8

2.10

1.2

1.2

1.8

1.6

1.45

10

0.2

1.0

3.6

1.8

1.65

0.2

0.6

1.4

0.8

0.75

11

1 .fl

1.4

0.8

0.95

0.4

0.6

0.8

0.8

0.65

12

0.0

2.0

1.0

0.90

0.4

0.2

1.4

1.0

0.75

13

0.4

1.6

0.8

0.70

0.2

0.4

1.2

0.2

0.50

14

2.4

1.0

0.85

0.4

0.4

0.20

15

0.2

0.8

0.6

0.40

0.6

0.2

0.20

16

0.6

0.4

0.25

0.2

0.2

0.10

17

0.6

0.4

0.25

0.2

0.2

0.10

18

0.4

0.2

0.15

19

0.2

0.4

0.15

20

0.4

0.2

0.15

21

0.4

0.10

22

0.2

0.05

23

0.2

0.05

24

0.2

0.05

25

26

27

entire lack of preference would be indicated by an equal distribution of the choices — five for the white and five for the black in each series — the preference for the black, in the case of the males, is .6 in series A and .5 in series B, and in the case of the females it is .9 in series A and .8 in series B.

(2) The females make more errors than the males in the first

246 'Journal of Cof?iparative Neurology and Psychology.

training series, but thereafter they make fewer errors and their training is completed with fewer series than that of the males. In other words, the general averages of Table 5 indicate that a group of twenty female dancers, ranging in age from one month to twelve months, acquired the habit of discriminating between two boxes, whose only considerable difference was in amount of illumination, and of choosing the white box much more quickly than did a comparable group of twenty male dancers. This is especially interesting in view of the fact next to be noted.

(3) The one-month males exhibit a considerably less strong Ijreference for the black box than do the one-month females and, at the same time, they acquire the habit much the more quickly. The reverse is true of the four-month groups : the males exhibit the

TABLE 6. Indicks of Plasticity fob Dancers of Different Ages.

1

] Males.

Females.

Both Sexes.

Age of Individuals.

First correct series.

Total no.

of training

tests.

82 128 192 160

First correct series.

Total no.

of training

tests.

First

correct

series.

Total no.

of training

tests.

74

76

78

106

98

106 106 146 134

75

85

126

113

94

4 months

92

117

7 months

10 months*

146 128

169 147

One male and two females whose ages were twelve months are included. stronger preference for the black to begin with and learn somewhat less rapidly.

(4) The males acquire the white-black habit more quickly at the age of one month than at the ages of four, seven, or ten months. And the females likewise acquire the habit most readily at one month of age.

Several of the above facts are clearer in the light of the results of Table 6, in which are arranged the indices of plasticity for the several groups of dancers, and the number of trials which, in the case of each group, preceded the first correct series of choices. The indices are given in the columns headed "total number of training tests." Judging by these indices we may say that the plasticity of the male dancer, as measured by the particular habit under consideration.

Yerkks, Modifability of Behavior. 247

diminishes from the age of one month to between the seventh and the tenth months. It seems then to increase slightly. Whereas 82 is the index for the one-month individuals, that for the four-month males is 128, and that for the seven-month males 192. For the tenmonth group the index 160 indicates increasing instead of diminishing plasticity.

The results indicate that the plasticity of the females does not change greatly between the first and the fourth months ; that thereafter it decreases for a few months, and then again increases slightly. The indices for the several ages, as they appear in Table 6, are 106, 106, 146, and 131. All of these except the first, indicate a degree of plasticity higher than that of the males.

Figures 3 and 4 represent graphically the principal results of the experiments which have just been described. Figure 3 is based upon the general averages of Table 5. The irregularly broken line is the curve of the learning for the males of the dancer race ; the regularly broken line, for the females of the race. The superiority of the females in the acquisition of this particular white-black discrimination habit is apparent. Figure 4 is based upon the data of the third, fifth, and seventh columns of Table 6. The irregularly broken line may be termed the plasticity curve of the male dancer (for a particular habit) between the ages of one month and ten months. The regularly broken line may similarly be termed the plasticity curve of the female dancer between the same age limits. The solid line, the plasticity curve of the race.

And now we are confronted by the question, Why the age differences in plasticity which are exhibited by our results ? In reply we might say that preference determines rapidity of learning. For we note that the females, apparently because of their strong preference for the black box, make more errors of choice in the first training series than do the males (general averages of Table 5), and that subsequently they very rapidly learn to avoid the black box. It would appear, then, that initial preference for the black box is a favorable condition for habit-formation because it leads to a large number of errors in the early training series and thus gives the animal that experience which enables it to adjust itself to the situation. This

24H 'Journal of Conipnrative Neurology and Psychology.

1

1

I 1

1

i

i ii

1; ii

/;

//

\\ .

/ /'

/■ ./

/

i .

i : /' ;

/ /'

/ ,•

/ ,

/ .■■■■

1

I /

/-"

i 1

! /

' 1

l\

1

\ \

1

\ \

1 1

i /

\ \

1 / / /

\

1 /

■\

1 / i / /

"'■---,

/ /'

/ /

//

--;/"

X /

/

Fig. 4.

Tig. 3. — Curves of leaniin.u; (error curves) for t\Yeiity male and twenty female dancers in white-black visual discrimination habit. Ordinates represent number of errors of choice in series of ten tests (per day). Abscissje represent daily series of tests, beginning with two preference series, A and B, and continuing to the completion of the training. Curve for males — • ; Curve for females

Fig. 4. — Plasticity curves for twenty male, and twenty female dancers, and for the race, in white-black discrimination habit. Ordinates represent indices of plasticity or modifiability (/. c, the number of tests up to the point at which CO no errors of choice were made in three successive series of ten tests each). Abscissjie represent ages in mouths. Curve

for males — - ; Curve for females ; Curve

for race .

Fig.

Yfrkes, Modifiahility of Behavior.

249

interpretation of the facts, however, is contradicted by the following results which comparison of the data of Tables 6 and 7 reveals. The one-month males, although they showed only a slight initial preference (0.2) for the black box, acquired the white-black habit more quickly than did any other group of dancers. On the other hand, the four-month males, while exhibiting a strong initial preference (1.0) for the black box, acquired the habit much less quickly than did the one-month individuals. In view of these facts it is impossible to conclude that preference plays an all-important role as a condition for white-black habit-formation. Evidently we must look elsewhere for the factor or factors upon which the results of the plasticity experiments depend.

TABLE 7.

Gejsteral Results of White-black Preference Tests for Individuals Which Were Used in the Study of the Relation of Age to Modifiability OF Behavior.

The figures in the table represent the number of choices of the black in preference to the white in series of ten tests each.

Males.

Females.

Individuals.

Series A.

Series B.

Average,

Series A.

Series B.

Average.

5.4 5.6 5.8 5.6

5.0 6.4 5.2 5.4

5.2 6.0 5.5 5.5

6.2 6.2

5.4

5.8

6.0 5.4 6.2 5.6

6.1

5.8

5.8

5.7

One male and two females whose ages were twelve months are included. It is quite conceivable that age or sex differences in the value of the electrical stimulus may be responsible for the differences in rate of habit-formation which appear. This possibility was tested experimentally by an examination of (1) the relation of electric sensitiveness to age and sex, and (2) the relation of strength of stimulus to rapidity of habit-formation. Before attempting further to analyze or interpret the results presented in the tables we shall examine the experimental data which enable us to answer the questions, Does the sensitiveness of the dancer to electric stimuli depend upon age and sex, and, Does the strength of the electric stimulus influence the rapidity of habit-formation ?

250 'Journal of Comparative Neurology and Psychology.

III. SENSITIVENESS TO ELECTRIC STIMULUS, IN ITS RELATIONS

TO AGE AND SEX.

The measurements of sensitiveness to electrical stimulation which are now to be presented were made in connection with the study of the relation of age to rate of habit-formation, for the special purpose of throwing light upon the interpretation of the data which have just been considered. Had the experimenter's aim been to make a thorough-going investigation of the limits of sensitiveness in the dancer, other and more accurate methods would have been employed. But as matters stood, it seemed desirable to use for these tests of sensitiveness the method of applying the stimulus that had been used in the plasticity experiments themselves.

In its especially adapted form, this method exhibited the following points of importance. A current from a storage cell was used, in connection with a calibrated Hasler inductorium,'^ as stimulus. The strength of the induced current was regulated by moving the secondary coil. By means of an interrupted circuit device similar to that previously described^ the mouse was permitted to receive this current through its fore feet. ^Miile one observer manipulated the keys of the circuits and regulated the strength of the current, another placed the mouse in position and observed its behavior When it received the shock. Determination was made, by repeated trials, of the lowest stimulus strength to which a definite motor response was given, and of the strength to which only an uncertain response was given. The average of these two results was accepted as the iJireshold value for the individual.

Twenty male and twenty female dancers were tested on two different days. The results in terms of the position of the secondary coil, as they appear in Table 8, indicate: (1) That the males are somewhat more sensitive than the females. This difference, which, according to the calibration curve of the inductorium, is nearly ten per cent., was not evident to the experimenter as he worked with the dancers from day to day in the training tests. (2) There is no indication of change of sensitiveness with increase in age.

'For the use of this inductorium I am inclebted to Dr. E. G. Martin. 'The Dancing Mouse, p. 94.

Yerkes, Modifiahility of Behavior.

251

As it happens tlie averages for the age groups are precisely the same for both the males and the females. This is a surprising result which we could not expect to obtain by the repetition of so small a number of tests. (3) Individual differences in sensitiveness are much more marked, and in all likelihood more important, than either sex or possible age differences.

As there is no reason to suppose, in the light of these rough determinations, that i)ossible changes in sensitiveness which accompany ageing account for any of the results of the jjlasticity experiments, we may ask whether it is at all likely that the sex

TABLE 8. Measurements of Sensitiveness to Electbicax STiMxn:.us.

Sex Differences.

Position of secondary coil, First

Day

Position, Second Day

Average Position

F-!rtrpmp<j I Least sensitive

H^xtremes | ^j^^^ sensitive

Males (averages for 20).

17.88+ cm.

17.72—

17.80*

15.75

19.25

Females (averages for 20).

17.41 cm.

17.25

17.33

16.67

18.00

Age Differences.

Not over four months.

Number of dancers tested 13

Average age 2.5 mos.

Position of secondary coil. First

Day 17.73 cm

Position, Second Day ' 17 .87

Average Position I 17.80

Over four months.

7 8.3 mos.

18.17 cm.

17.44

17.80

Not over four months.

12 2.4 mos.

17.35 cm.

17.30

17.33

Over four months.

8 mos.

17.50 cm.

17.17

17.33

Difference in favor of males .47 cm., or about 10% of the value of the current. and individual differences in sensitiveness furnish the basis for such differences in rapidity of habit-formation as the tables indicate.

Before this question can be answered satisfactorily, we must know what relation strength of electric stimulus bears to rapidity of learning. Does increase in the strengih of the stimulus from the threshold value — or, what for our present purposes amounts to the same thing, increase in sensitiveness — facilitate or retard the process

252 'Journal of Comparative Neurology and Psychology.

of habit formation? As an answer to this question, I offer a summary statement of the results of a special investigation of the relation of strength of stimulus to rapidity of learning in which I was ably assisted by Mr. John D, Dodson. This study, unlike that of sensitiveness, was a thoroughgoing quantitative investigation of the significance of the factor under consideration, and we present our results with confidence that their accuracy, despite many technical difficulties, renders the generalizations which they indicate of imj)ortance not only in connection with our present experiments, but for all work on animal behavior.

IV. STRENGTH OF ELECTRIC STIMULUS, IN ITS RELATION TO RAPIDITY OF HABIT-FORMATION.

Precisely how does increasing or decreasing the strength of the electric stimulus, which the dancer is learning to avoid by associating it with the darker of two boxes, influence the process of learning ? The answer which results obtained with forty dancers enable us to give to this question is exceedingly important in its several aspects.

fl) We have demonstrated that the influence of the stimulus varies with the difiicultness of the visual discrimination which is demanded of the mouse, and that condition of disci-imination must be taken into consideration from the first in foniiiilatiug our answer to the above question.

(2) That when visual discrimination is easy, rapidity of habit-formation increases as strength of stimulus is increased from the threshold to the point of injurious stimulation. In our experiments, the strongest stimulus employed was decidedly disagreeable to the experimenters and caused violent reactions in the mouse. Whether beyond this intensity of stimulation the rate of learning increases, we cannot say from the results of experimentation, but we may say with assurance that it cannot possibl} increase very much, inasmuch as the stimulus would soon become positively harinfnl.

'Yerkes, R. M., and Dodson, J. D. The Relation of Strength of Stimulus to Rapidity of Habit-fonuation. Join: Camp. IVciir. and Psi/.. vol. IS, pp. 459482, 1908.

Yerkes, Modifiahility of Behavior. 253

(3j That when visual discrimination is moderately difficult, rapidity of habit-formation increases as strength of stimulus is increased up to a certain point, and with further increase in the stimulus it rapidly decreases. A moderate strength of stimulus is most favorable for habit-formation under this condition of discrimination.

(4) That when visual discrimination is very difficult, rapidity of habit-formation increases as strength of stimulus is increased for a time, but not nearly so long as in the case of the medium condition of discrimination, and then with further increase in the stimulus it rapidly decreases. A low intensity of stimulus is most favorable for habit-formation under this condition of discrimination.

The law which is indicated by these facts may be formulated thus. As difficuUness of visual discrimination increases that strength of electrical stimulus luhich is most favorable to hahit- formation approaches the threshold. The easier the hahit the stronger that stimidus ivhich most quickly forces its acquisition; the more difficult the hahit the wcalxcr that stimidus ivhich rnost qiiichly forces its acquisition.

From these facts it is evident that the value of a given strength of electric stimulus, for the training of a dancer whose sensitiveness is accurately known, can be stated only if the degree of difficultness of discrimination for the individual also be known. A degree of difference between the white and the black boxes which renders discrimination moderately easy for one dancer may render it (^xtremely easy for another. Male and female, or old and young, or even two individuals of the same sex and age, may differ, both in discriminating ability and in sensitiveness.

This consideration makes apparent the incomparability of the results of the plasticity experiments. Instead of uniformity and simplicity of conditions, we have variability and complexity. It is evident that before a given individual can be used to advantage in any such training experiments as these, or rather before we can interpret the results, we mu,st know accurately the relations of the conditions of experimentation to the individual.

254 Journal of Comparative Neurology and Psychology.

Intensity of the electric stimulus is, then, important in connection with rapidity of hahit-formation. Since, however, no difference in sensitiveness appears to be correlated with age differences we may assume, until we know otherwise, that the age differences in rapidity of learning are not due to the influence of the electric stimulus. But, at the same time, since the males appear to be raore sensitive than the females, it may be that the sex differences in rapidity of learning are in part at least due to the influence of the stimulus. Possibly the particular combination of condition of discrimination and strength of stimulus was more favorable for the one-month males than for the comparable group of females, and possibly also for the females, as a whole, the combination of conditions was more favorable than for the males.

The significance of this suggestion will be clearer in the light of the results of the next section of this paper, for in that we shall have to examine data, which, if I could have foreseen them at the beginning of my work with the dancer, would have altered almost all of my experiments. I do not wish to give the reader the impression that I regard the results of the plasticity experiments as valueless or that I consider this investigation of mine exceptional in comparison with the work of any or all other investigators in this field. On the contrary, I have great respect for both the experimental procedure and the results which it yielded, but I am especially interested in pointing out the complexity of conditions which the investigation has revealed.

Before turning to the topic of the next section, I wish to call attention to the probable significance of the law of habit-formation which I have tentatively formulated above. As I have stated it, this law may not hold for other conditions of habit-formation, or for other animals. Only further investigation along lines which Mr. Dodson and I have followed can decide these questions. Meanwhile, it is evident that the subject is of great importance, for much of our experimental work in animal psychology rests upon the assumption that the stronger the stimulus which conditions a particular act the sooner the animal will learn to perform that act. In the light of our results concerning the relation of strength

Yerkes, Modifability of Behavior. 255

of electric stimulus to rapidity of habit-formation it becomes pertinent to inquire, Is utter hunger as favorable a condition for the discovering of a certain method of obtaining food as moderate hunger ? Is extreme eagerness to escape from confinement as favorable as a moderate desire ? What we really should know before we undertake to study the intelligence of a particular animal is the value for it of the several factors which constitute the chief experimentally controlled conditions of activity. So long as we continue to use external conditions as incentives to habit-formation, without definite knowledge of their values for the individual, we shall work blindly. Food supply — the internal aspect of which is hunger — as a condition of habit-formation, may be studied experimentally; and the same is true of every other so-called motive upon which the experimenter depends. It is high time that we made serious efforts to discover the values of our stimuli instead of slothfully assuming that they will answer our purposes.

That there are a number of important laws of habit-formation to be discovered no student of animal behavior can doubt. These laws, of which the one offered above may serve as an example, should rapidly replace what is too much talked of as "the law of habit-formation."

V. RELATION OP DIFFICULTNESS OF DISCRIMINATION TO RAPIDITY OF HABIT-FORMATION AT DIFFERENT AGES.

Among the important results of the investigation of relation of strength of stimulus to rapidity of learning was the demonstration of the fact that differences in plasticity depend upon the condition of visual discrimination as well as upon the strength of the electric stimulus. What holds with respect to rapidity of acquisition of the white-black discrimination habit in young and old dancers, under conditions which render discrimination difficult, does not necessarily hold under conditions of easy discrimination. This I have demonstrated, and thrown further light upon, by three different methods, the results of which will now be presented in turn.

1. Experinienis with cardboards in discrimination hex furnished

256 'Journal of Comparative Neurology and Psychology.

the first indication of the great importance of condition of discrimination. With other points of method the same as in the plasticity exj^eriments, I so arranged the black and white cardboards of the discrimination box that the amount by which the white box differed in illumination from the black box was very much greater than it had been in the earlier experiments. Whereas, formerly, discrimination had been rather difficult, it was now made

TABLE 9.

Relation of Age to Rapidity of Habit-formatiox Under Conditions of

Difficult and of Easy Discrimination

White-black Discrimination

Dancers 8 or 12 Months Old.

Dancers 1 Month Old.

Difficult

Easy

Difficult

Easy

Discrimination.

Discrimination.

Discrimination.

Discrimination.

Series.

No. 112 No. 113

No. 204

No. 121

No. 292

No. 291

No. 430

No. 432

A

6 8

5

4

7

7

7 1 8

B

5 7

4

5

5 6

8 i 8

1

4 6

6 7

7 8

7 6

2

4 ! 3

4 1

3 6

5 3

3

3

6

4

1 3

2 1

4

4

3

1 1

4 2

2 1

5

4

2

1 2

6

4

2

1 1

7

5

3

1 1

8

2 1

3 1

0* 0* 2

9

0* 0*

10

4 1

1

11

1

2

1

12

2

1

1

13

1

14

3

15

1

16

1

17

j

18

1

19

1

1

At this point condition of discrimination was changed from " easy " to " difficult." easy. The plasticity experiments, it is to be remembered, showed that the young dancers acquired the habit much more rapidly than the old individuals. Just the reverse proved to be true under the conditions of easy discrimination : the old mice learned more quickly than the young individuals.

In order to make the results perfectly conclusive, I carried out series of training experiments at the same time with a pair of

Yerkes, Modifiahility of Behavior. 257

dancers one month old and a pair twelve months old under the conditions of discrimination used in the plasticity investigation, and similarly with two pairs of dancers one of which was one month old and the other eight months, under the conditions which I have just characterized as easy. The results of these experiments, as they are presented in condensed form in Table 9, are striking indeed. As was the case in my first experiments, the old dancers^" acquired their habit, under conditions of difficult discrimination, much less rapidly than did the young individuals. The index of plasticity for the twelve-month mice, Nos. 112 and 113 of the table, is 130; that for the one-month mice, Nos. 292 and 291, is 70. The latter acquired the habit with few more than half as many training tests as were necessary for the former.

When we turn to the results of the experiments made under conditions of easy discrimination, we find that the eight-month mice, Kos. 204 and 121, learned with only 40 tests; whereas, the onemonth individuals, ]!^umbers 430 and 432, required 50 tests.

In Table 10 are presented the results of additional experiments like those just described. Two eight-month dancers, Nos. 136 and 166, acquired the habit on the basis of 40 and 20 tests, respectively. Their index of plasticity is, therefore, 30. The index for two fourmonth mice, which were subjected to the same training, was 80, and that for two eight-month individuals, 75.

The importance of the relation of age to difficultness of visual disci'imination is clearly exhibited by the indices of plasticity for young and old dancers under conditions of easy and difficult dis, crimination in Table 11.

Comparison of the data of Tables 9, 10 and 11 with those of Tables 1 to 6 proves conclusively that the direction of the age differences in plasticity which was revealed by the experiments described early in this paper was determined, in part at least, by the condition of visual discrimination which happened to be chosen for the

^°I shall use the terms of old and young in contrasting two groups of dancers which differed in age by several months. As a matter of fact a dancer at the age of eight, ten, or even twelve months is not, as a rule, obviously senile.

258 Journal of Comparative Neurology and Psychology.

experiments. Had the tests been made with a condition of greater difference in the illumination of the two boxes the results probably would have indicated a slight increase in plasticity with age, instead of a decrease. If then under one condition of training plasticity diminishes as the dancer grows older, and under another condition in connection with the same habit it increases, it is clear that the

TABLE 10.

Relation of Age to Rapidity of Habit-formation Under Conditions of Easy

AND OF Difficult Discrimination. White-black Discrimination.

Discrimination at first easy and later difficult.

Series.

Dancers 8 Months Old.

Dancers 4 Months Old.

Dancers 1 Month Old.

No. 136

No. 166

No. 408

No. 185

No. 416

No. 105

A B

1 2 3 4 5

5

7 6 5

1

6 4

4

1

II

4 4 6 5 3 3 2

1

4 3 5 3 3 1 3

2 T

8 6

7 4 5

4

7

2

8 3 3 4 6 3 2

6

7

2* 2

4 3

2

2 3

1

2 ■

1 1 1 2

1 1

8 9

1* 1

2

1 i

1 1

10

3*

4 6

2 3 1 1 2

1

11 12

0*

1

13 14 15 16 17 18 19 20 21 22 23 24 25

1*

1

1

1*

At this point condition of discrimination was changed from "easy" to "diflBcult." relation of age to rapidity of habit-fonnation is more complex than certain statements made by students of animal behavior would lead one ±0 suppose.

M}^ experiments reveal the presence and importance of a number of variable factors in the white-black discrimination habit; and until we know accurately the values and relations of these several

Yerkes, Modifiability of Behavior.

259

factors it would be rash indeed to make general statements concerning the relation of age to plasticity. We must limit ourselves carefully to particular statements, for what holds of one condition of training may not hold at all of what appears to be a very similar condition.

TABLE 11.

Indices of Plasticity for Dancers of Different Ages, Trained Under

Conditions of Difficult or of Easy Visual Discrimination.

Condition of Discrimination.

No. of dancer.

DIFFICULT.

EASY.

Young dancers

Old dancers

Young dancers

Old dancers

1 month old.

12 months old

1 month old.

8 months old.

112

160

113

100

204

40

121

40

292

60

291

80

430

50

432

50

136

40

166

20

416

60

105

90

Averages.

70

130

62.5

35.0

2. Experiments luitli discrimination hex in darTc-room. The results of the experiments which have just been described suggested to me the idea that ability to acquire the white-black visual discrimination habit depends largely upon two factors: capacity for visual discrimination and associative memory. The facts of plasticity thus far revealed might be accounted for, it would seem, by the assumption that in the young dancer capacity for visual discrimination was either greater at the outset or more readily developed than inj:he case of old individuals, whereas associative memory is more highly developed in the old than in the young mice. This hypothesis I immediately attempted to test experimentally. If it be correct, young mice should develop the capacity to discriminate slight differences in luminosity more quickly than old mice. To test this matter I planned a series of training experiments with the apparatus which I have previously described^ ^ as the Weber's

"The Dancing Mouse, p. 118.

260 'Journal of Comparative Neurology and Psychology.

law apparatus. It is a discriiniiiatioii box iii which the two boxes which have heretofore been referred to as white and black are illuminated by standardized incandescent lamps. There are no cardboards and difference in illumination, as desired, is obtained by shifting the position of the source of light for one of the boxes. This apparatus permits easy and fairly accurate measurements of the absolute and relative illumination of the two boxes, and in this respect it is more satisfactory than the cardboard method. Its chief disadvantage is that it compels experimentation in a dark-room or at least with artificial illumination of the boxes.

In the Weber's law apj^aratus two pairs of dancers were trained systematically until they had been given almost a thousand tests. The individuals represent the age limits of the plasticity experiments. The old ones, Nos. 170 and 95, were ten and twelve months, respectively; the young ones, Xos. 294 and 293, were one month old. Instead of a single series of ten tests per day, all these individuals were given two such series each day.

To start with, all the mice possessed perfectly formed habits of choosing the white box, in the old white-black discrimination apparatus. Experiments in the Weber's law apparatus were begun with the two boxes illuminated the one by 80 hefners, the other by 20 hefners. The difference in luminosity in this case may be stated as three-fourths, since the latter value is only one-fourth the former. I have found it convenient to keep one of these values constant throughout a training experiment and to vary the latter as need dictated. The fixed value, which may then be known as the standard, is indicated in the table by the abbreviation S. The other value, which may be known as the variahle, is indicated by the abbreviation Y.

A habit was considered perfect in this experiment when a dancer succeeded in choosing without error in two successive series. As soon as ability to. discriminate a certain degTce of difference in luminosity had been acquired, the amount of the difference was reduced and the training continued under the more difficult condition of discrimination. We may now examine the results of this experiment as they appear in Table 12.

Yfrkhs, Modi-fxahility of Behavior.

261

At the outset the eouditiou of disci-iniination was fairly easy and the old dancers learned to choose correctly with 110 tests, the young

TABLE 12.

Relation of Discriminating Ability to Age. Experiments with Weber's Law Apparatus.

Dancers 10-12 Mo.

Dancers 1 Mo. Old.

Dancers 10-12 Mo.

Dancers 1 Mo. Old.

Series

No. 170

No. 95

Series

No. 294 No. 293

Series No. 17o[ No. 95

Series No. 294 No. 293

S. 80 h.V. 20 h. Difference three-fourths.

S. 80 h.V. 60 h. Difference one-fourth.

1

6

8

1

9

6

44

3

3

41

4

3

2

4

5

2

8

5

45

1

5

42

2

4

3

7

6

3

3

4

46

4

3

43

4

3

4

5

3

4

6

4

47

3

4

44

3

7

5

4

3

5

4

1

48

6

3

45

8

6

6

1

3

6

5

49

3

4

46

3

7

7

1

3

7

1

50 51

1 4

1

47 48

2 3

6

4

8

1

2

8

5

52

4

5

49

2

5

9

1

1

9

3

1

53

3

4

50

3

6

10

10

3

54

3

2

51

2

3

11

2

11

1

55

56

3 2

3

2

52 53

2 4

3

4

12

2

12

2

57

1

4

54

3

4

13

1

13

1

58

3

5

55

4

3

14

14

1

59

2

5

56

4

2

15

15

60 61

2 4

6 5

57 58

4

5

3 &

16

16

62 63

2 2

4 4

59 60

2 3

3

5

64

3

6

61

6

7

S.

80 h. V.

40 h. I

difference one-half.

65 66

2 5

5

6

62 . 63

3 3

5

4

17

3

1

17 1 2

67

3

4

64

3

7

18

2

18 1:2

68

2

3

65

2

3

19

3

3

19 5 ; 4

69

2

4

66

1

2

20

2

1

20 11

70

2

4

67

4

21

3

21 13

71

3

4

66

4

3

22

1

3

22 1

1

y*>

3

5

69

2

3

23

3

23 3

2

73

3

2

70

4

24

1

1

24 2

5

74

2

4

71

5

2

25

1

2

25

2

5

75

2

3

72

2

3

26

1

2

26

3

76

1

2

73

3

4

27

1

1

27

4

3

77

1

2

74

2

5

28

28

5

S. 8(

) h. V. 26.66 L. Difference one-t

bird.

29

2

1

29

2

78 79

2 1

4 3

75

76

2 2

2

30

1

1

30

2

80

1

5

77

2

3

31

1

31

1

81

2

6

78

1

2

32

1

32

82

1

1

79

1

33

2

2

33

34

1 4

2

83

4

5

80

1

34

1

35

2

2

35

2

84

2

5

81

3

36

2

1

36

85

1

3

82

37

2

2

37

1

86

1

2

83

4

38

3

1

38

1

87

4

2

84

2

39

1

39

88

2

3

85

3

40

1

2 1

40

89

3

5

86

3

41

S. 80 h.V. 32 h. Difference two-fifths.

42

90

1

1

87

1

43

i

91

1

3

88

3

ones, with 95. In view of the results of the previous section we might have expected the young individuals to learn more slowly than

262 'Journal of Comparative Neurology and Psychology.

the old ones. But we must remember that the couditioiis of this experiment are markedly different from those in which cardboards were used to render the two boxes visually distinguishable.

ISText the amount of difference in luminosity was reduced to onehalf, and the experiment continued. Again the young individuals acquired the habit more quickly than the old ones. The index of plasticity for the old is 250, for the young it is 1G5.

With a difference in luminosity of only one-fourth, the training was now continued for several days, but as no one of the four mice succeeded in acquiring a perfect habit it was changed finally to one-third. It is noteworthy that in the thirty-four series (340 tests) that were given to the mice with the difference one-fourth, the old individuals did not succeed in making a correct series, whereas both of the young mice did. With the difference onethird, No. 294 quickly acquired a perfect habit, and No. 293 came very near to doing so, but failed in twelve series. At the conclusion of the twelfth series, neither of the old individuals had learned to choose correctly, with the difference one-third.

Although the results of this experiment are not as convincing as they might be, they do indicate that young dancers can acquire the ability to discriminate slight differences in luminosity more readily than can old individuals. It is conceivable, then, although by no means demonstrated as true, that the young individuals in the plasticity experiments acquired the white-black habit more quickly than the old individuals did because they could discriminate better or acquired discriminating ability more rapidly and not because they acquired an association more readily. In this event, our experiment measures differences in visual discrimination, and in changes which it undergoes with training instead of associative plasticity.

I have already sho\\ai^- that the dancer is capable, as the result of prolonged training, of developing the power to discriminate between boxes which differ from one another in illumination by less than one-tenth. This fact becomes important at this point, for

"The Dancing Mouse, pp. 127, 128.

Yerkes, Modifiahility of Behavior. 263

we are forced to ask, Do the jDlasticity experiments reveal anything except age differences with respect to what might be termed the educability of light vision? With the hope of getting further light on this problem, I carried out additional experiments, with the individuals used in the Weber's law apparatus, by a method whose form and results will now be described.

3. Experiments with one side of discrimination hox covered in varying degrees. For this work the cardboards were removed from the discrimination box which had served for the plasticity experiments, and difference in the illumination of the two boxes was obtained by covering, with a jDiece of black cardboard, the whole or a part of the top of one of the two small boxes. The total inside length of the boxes was 29 cm. I have described the condition of the darker box by giving in terms of a fraction the amount of the top which was covered. Thus 18/29 means that the cardboard covered 18 of the 29 cm., beginning at the entrance and extending toward the rear of the box. Shifting the lighter box (the one to be chosen) from side to side involved merely the moving of the black cardboard from the top of one box to the top of the other.

After the experiments just reported had been completed, mice jSTos. 170, 95, 294, and 293 were given training tests in the discrimination box under the above conditions. Table 13 presents the condition of discrimination as well as the results of the various series of tests. When, as at the outset, the whole of one box was covered, discrimination was extremely easy, because the boxes differed greatly in illumination.

From the first, as the data of Table 13 indicate, the young animals learned more rapidly than did the old ones. We have in these results, therefore, additional support for the belief that discriminating ability is more readily gained by the young dancer.

It may not be out of place to remark here that the simple form of the lighter-darker discrimination apparatus which served for this series of experiments is precisely what should have been used throughout this investigation. It has taken me years to learn that it is not only possible, but also perfectly easy, to devise a condition of experimentation which should be readily and accurately

264 'Journal of Comparative Neurology ami Psychology.

describablc as to the difference of brightness of the two boxes and satisfactory in its results. I cannot too strongly urge, from my present point of view, the avoidance of cardboards as means of testing visual discrimination. The conditions of many of my experiments are practically indescribable so far as absolute value of illumination is concerned, yet, as I now see it, they might perfectly well have been describable with a fair degree of accuracy.

TABLE 1.3.

Relation of Age to Ability to Discriminate on the Basis of Difference

IN Illumination, and to the Capacity for Improvement of

Visual Discrimination.

Dancers 10

-12 Months

Old.

D.VNCERS

1 Month

LD.

Series.

Portion of darker

Portion of darker

box covered by

No. 170.

No. 9.5.

box covered by

No. 294.

No. 293.

card.

card.

1

Whole.

5

5

Whole.

5

3

2

2

6

4

4

3

2

1

1

4

1

5

if

hi

6

1

7

1

1

8

2

9

If

3

10

25

3

5

2

11

M

1

3

1

12

M

2

2

K

4

2

13

5

2

&

2

1

14

5

5

A

4

2

15

2%

5

2

A

2

1

16

5

2

2

f^

1

3

17

4

T

1

18

4

1

59

2

2

19

"5

4

2

fi

1

20

?5

4

1

21

2%

2

3

Having now tested the first of the two important factors in the the acquisition of the white-black discrimination habit, namely, ability to gain visual discriminating power (the educability of white-light vision), we must turn to the second factor and inquire whether associative memory changes with age.

It occurred to me that since the labyrinth habit, as conclusively proved by Professor Watsoii^^ for the white rat, depends more

"Watson, J. B. Kinsestlietic and Orgauic Sensations : Their Role in the Reactions of the White Rat to the Maze. Psychol. Rev. Mon. Siipp., vol. 8, no. 2, 1907. vi -)- 100 pp.

Yfrkks, MoclifahiUlv of BeJiavior. 265

largvly iipoii kiiucstlictic seiisi' data lliau ui>(>ii vision (»r any f»ther special sense, it might serve well to reveal age differences in associative ability. In this connection we may ask, therefore, Do old dancers learn la1\vrinth i)aths more readily than yonng ones?

VI. RELATION OF AGE TO RAriDITY OF ACQUISITION OF LABYRINTH HABITS.

For the labyrinth experiments I selected two mazes which I had i^reviously used for the study of educability in the dancer: they are desiginited as P and C in my Ix^ok.'^ Y) is what I have described as the regular type of maze, and C as the irregular. Training in labyrinth D was given first to each of ten dancers of from one to two months of age, and likewise to the same number of about ten months of age. About a month after the completion of the training in labyrinth D, the same individuals were trained in labyrinth C.

In all cases the experiments were conducted as follows. Two mice, a male and a female from the same cage, were placed in the nest box of the labyrinth together. One at a time they were given first a preliminary test in which they were permitted to find their way from the entrance to the exit of the labyrinth without being disturbed, and then training tests in which they received a slight electric shock each time they made_ an error in the choice of a path. The tests were continued without interruption, first one individual then the other being tested, until each had perfectly learned the path. \ habit was considered as perfect when an individual succeeded in traversing the maze twice in succession without a mistake. Records were kept of the number of errors in choice of a path and of the time consumed in finding the way from entrance to exit.

As tyj)ical series of results I present in Table 14 the time and error records in labyrinth D of No. 416, a six-week dancer, and No. 166, a nine-month individual. The young mouse was slower than the old one in most of the tests, but he acquired a perfect habit

"The Daiiciu.ii Mouse, pp. 210. 222.

266 Journal of Comparative Neurology and Psychology.

no less quickly. I need scarcely state that the time records have little value for our present purposes, and are therefore omitted, except in the case of Table 14.

TABLE 14. Relation of Age to Rapidity of Acquisition op Labyrinth Habits. Typical Series of Results Given by Two Males in Labytiinth D.

Dancer 6 Weeks Old.

Dancer 9 Months Old.

No. 416.

No. 166.

Number of trial.

Time in seconds.

Number of errors.

Time in seconds.

Number of errors.

Preliminary.

280

31

161

19

I

240

60

56

6

2

134

24

25

4

3

53

6

62

8

4

22

143

17

5

141

3

62

8

6

15

57

6

7

22

1

31

3

8

14

15

9

63

6

46

3

10

58

2

29

2

11

15

20

12

r

12

13

TABLE 15. Relation of Age to Rapidity of Acquisition of Labyrinth Habits. Results in the table indicate the number of trials up to the point at which no errors occurred for at least two consecutive trials.

Dancers 1-2 Months Old.

Dancers 10 Months Old.

Number of animal.

Results for Labyrinth D.

Results for Labyrinth C.

Nuiriber of animal.

Results for Labyrinth D.

Results for Labyrinth C.

256 258 396 398 418

5

8

15

16

13

27 22 26 14

7

92

96

98

120

166

15 6 6 9

10

22

6

19

12

Av. for Males.

11.4

19.2

Av. for Males.

9.2

14.7 +

179 181 255 263 395

4 19 10

6 13

16 9

18

7

91 93 97 99 109

19

15

6

8

10

8 13 11 19

7

Av. for Females.

10.4

12.5

Av. for Females.

11.6

11.6

Gen. Av.

10.9

16.2

Gen. Av.

10.4

13.0

The general results of the labyrinth experiments appear in Table 15. Averages are given for the sexes separately, inasmuch as so

Yerkes, Modifiability of Behavior. 267

often heretofore we have discovered sex differences to be of importance for the interpretation of our results. Comparing the young dancers with the okl, we note that the males of one to two months of age acquire both the labyrinth D and the labyrinth C habits considerably less quickly, as measured by the number of tests, than the ten-month individuals. In the case of the two groups of females there is practically no difference in rapidity of learning. The general averages likewise show that the old dancers are somewhat superior to the young in ability to learn these labyrinth paths.

What evidence we have, favors the conclusion that the associative memory of the dancer improves somewhat during the first year of life. Possibly change in the "apperceptive mass" is responsible. It is only fair to admit, in concluding the presentation of experimental data, that I consider this problem unsolved, for I have presented insufficient evidence to convince the critical observer that associative memory improves with age.

VII. CONCLUSIONS AND SUMMARY. This attempt to discover the relation of age and sex of the dancing mouse to plasticity, or rapidity of habit formation, makes possible interesting and important, albeit not altogether favorable, comments upon the methods of the investigation. As the work progressed it became increasingly clear that the use of cardboards as means of producing different degrees of illumination of the two boxes between which the mouse was forced to discriminate was unsatisfactory. Chief among the objections to this method may be mentioned the practical impossibility of keeping the difference in illumination constant throughout even a single series; the impossibility of determining accurately, except by very elaborate and time-consuming methods, either the relative or the absolute illumination of the white and black boxes ;^^ the impossibility of chang "Photometric determinations of the amount of light reflected by the white and the black cardboards which were used throughout these experiments indicate that the white cardboard reflected about 10.5 times as much light as the black cardboard. For a careful measurement of these values of the cardboards I am indebted to Trofessor J. W. Baird.

268 journal of Comparative Neurology and Psychology.

ing the amount of difference in the illumination of the boxes with ease and accuracy. These are only a few of the objections to white, grey, and black cardboards or papers that experience enables me to raise. Naturally I shall neither use them nor recommend their use hereafter in investigations of the visual powers of animals. Later, in connection with a report on '^Methods of studying vision in animals" which is to be made by the committee on standardization of tests of the American Psychological Association, I shall propose a substitute method.

The investigation has shown, I believe, the great importance of choosing conditions of experimentation Avhich may be readily and accurately measured and controlled, and of determining, as a preliminary to any experimental study of habit-formation, the value for the individual animal of the several important factors in the experimental situation. It has further shown that we should Avork with individuals, and with relatively simple and perfectly analyzable situations ; and that no treatment of our results is so likely to hide their real significance as the averaging of groups of observations for different individuals. Evidently the best preparation for an exj^eriment is a thoroughgoing study of the characteristics of each animal to be used in the investigation ; and the best result which an experiment can yield is evidence of the relation of experimentally controlled conditions to the particular traits of an individual animal. Averages are important, but we should not sacrifice individual facts for the purpose of presenting them.

The primary aim of the investigation, it will be remembered, was the discovery of the relation of plasticity to age and sex. The data prove that dancers at the age of one month acquire the whiteblack habit more rapidly than do older individuals, and that the females, on the average, acquire the habit more rapidly than do the males. Of gi-eat importance is the fact (represented by the curves of Fig. 3) that whereas the females make more mistakes of choice at first than do the males they very soon begin to choose with a higher degree of accuracy, and ultimately acquire a perfect habit with considerably fewer training tests than the males.

That these age and sex differences in the form of the habit

Yerkes, Modifahility of Behavior. 269

forinatiou curves are not necessarily indicative of general differences in plasticity is rendered evident by the results of the sections on sensitiveness, strength of stimulus, and difficultness of discrimination. For in the light of these results we are able to name as two important, and to a certain extent independently variable, conditions upon which the acquisition of the visual habit depends, (a) ability to sense the difference in illumination of the two boxes and (b) ability to associate the darker box with the electric shock. Evidently an individual which possesses highly developed white light vision and is capable of distinguishing very slight differences in illumination may, at the same time, possess little ability to associate stimuli. The data of section V indicate that what we have called age and sex differences in plasticity are in all probability, to be referred to differences in visual discriminating ability and in '^associative memory." Admitting that the results of the experiments justify only tentative conclusions, we may say that the young dancer seems to be somewhat superior to the old individual in ability to discriminate on the basis of difference in illumination, whereas the old individual seems to associate stimuli somewhat more rapidly, if anything, than the young dancer.

This suggestion, for it is scarcely more than that, of the way in which a habit may break up into two relatively independent factors is one of the most interesting results of the investigation. It strongly emphasizes the importance of studying the various sense factors separately and of attempting to discover upon what external and internal conditions '^associative memory" depends.

To sum up the results of the investigation point by point, it appears that —

1. The dancer at one month of age acquires a particular whiteblack visual discrimination habit more rapidly than do older individuals. From the first until the seventh month there is a steady and marked decrease in rapidity of habit-formation; from the seventh to the tenth month the direction of change is reversed. These statements hold for both sexes.

2. Young males acquire the habit more quickly than 3'oung females, but between the ages of four and ten months (at least) the females acquire the habit the more quickly.

270 Journal of Comparative Neurology and Psychology.

3. Curves of learning for tlie sexes indicate that the female makes more mistakes early in the training tests than does the male, but that this condition soon gives place to greater accuracy of choice on the part of the female.

4. Initial preference for the white or the black box does not seem to be a very important determinant of the rate of habit-formation.

5. Tests of sensitiveness indicate that the male dancer is somewhat more sensitive to electric stimuli than the female. There are no evidences of changes in sensitiveness with change in age.

C. The strcng-th of the electric stimulus which is used as an incentive for habit-formation is extremely important as a determinant of rate of habit-formation. For a given animal and condition of visual discrimination there is a certain strength of stimulus which is most favorable for the acquisition of the habit (the optimal stimulus),

7. It is extremely important that experimenters discover the optimal stimulus for habit-formation.

8. For the dancer the following law appears to hold in connection with the particular habit under consideration and for the electric stimulus. As difficuUness of visual discrimination increases ihat strength of electric stimulus which is most favorable (the optimal) to habit- formation approaches the threshold. The easier the habit the stronger that stimulus which most quickly forces its acquisition; the more difficult the habit the weaker the stimulus which most quickly forces its acquisition.

9. A given diiference in the illumination of the boxes which are to be discriminated cannot be detected with equal ease by old and young dancers. When the difference in illumination is slight the young individuals detect it more readily than the old mice; when the difference is great, the old individuals apparently detect it as readily as do the young mice.

10. The capacity for "associative memory" is greater, if anything, in the dancer of ten months of age than in the one-month individual. This is indicated by certain results of the visual discrimination experiments and by the results of labyrinth tests.

Yerkes, Modifiability of Behavior. 271

11. The results of this investigation indicate, then, that the acquisition of a visual discrimination habit depends upon two independently variable (within limits) capacities in the dancer: (1) the power to detect differences in illumination or to gain this power (educability of white-light vision), and (2) the power to associate the darker box with the electric shock (associative memory). The former of these capacities seems to be greater in the young than in the old dancer;, the latter seems to be somewhat greater in the old than in the young individual.

12. Should the statements just made hold true for animals generally, it is evidently important that the senses be trained early in life and that the development of associative memory be furthered later. Investigation of the problems suggested by our results should yield important practical data for the science of education.

THE EEACTIONS OF THE DOGFISH TO CHEMICAL

STIMULI.

BY

RALPH EDWAIU) SHELDON.

Contribiitioii from the Woods Hole Lahomtonj of the United Htutes B II lean of Fisheries*

With Tiikee Figures.

P.\GB

Introductory 273

Conditions of experinioutatioii 27(>

Reactions obtained 278

Sensitiveness to cliemical stimuli.

Experimental results 281

Analysis of results 281

Operations.

Destruction of the spinal cord 288

Transection of the spinal cord 289

Innervation and reactions of the nostrils 291

Chemical sensation as a sense quality 294

Conclusions 29-1

Summary 297

Bibliography 299

Figures 308

Introductory.

The smooth dogfish, Mustelus eaiiis (Mitchell), was the subject of experiment in an endeavor to find ont the sensitiveness of the general body sni-face to chemical stimnli and the extent to which the nerves of general sensation share in the reactions called forth through stimulation of the mouth and nostrils. In the investigation of these problems certain accessory points, such as the spinal animal and the innervation of the olfactory capsules, are taken up. The substances

Published by permission of George ISI. Bowers, U. S. Commissioner of Fisheries.

The Journal of Comparative Neurology and Psychology. — Vol. XIX, No. 3.

274 'Journal of Comparative Neurology and Psychology.

used as stimuli were those known to affect the gustatory, and to a less extent the olfactory sense, in higher forms.

The work was carried out at the laboratory of the U. S. Bureau of Fisheries at Woods Hole. I wish to express to the Director, Dr. F. B. Sumner, my appreciation of his assistance in furnishing me with every facility necessary for the successful prosecution of the work. The subject was originally taken up in 1907 under the direction of Dr. G. H. Parker in the Zoological Laboratory at Harvard University. I desire to tender to him my thanks for many helpful suggestions made both at that time and at the Woods Hole laboratory.

During the last few years much has been done, particularly among the invertebrates, on the reactions of animals to different kinds of stimuli. Part of this work has been concerned with chemical stimulation, the character of which is shown by the work of Pearl ('03), Bell ('06) and Jennings ('04 and '06). So far as the vertebrates are concerned, work on chemical stimulation has dealt almost exclusively with their two chief organs of chemical sense, smell and taste. A serious attempt has been made, however, to determine for the organs and their functions a physico-chemical basis. Haycraft ('87) was one of the first to attempt seriously to deal with taste from a strictly chemical standpoint. He was followed shortly by Corin ('88). The most important work of this character appeared about a decade ago from several sources simultaneously. Overton ('97) considered osmosis, while Kahlenberg ('98, '00), Kichards ('98, '00), Kastle ('98) and Hober and Kiesow ('98) have taken up critically the physical and chemical characters of substances which stimulate the gustatory apparatus in man, together with the chemistry of taste itself. Still more recently Herlitzka ('07) and particularly Sternberg, in a series of papers published from 1898 to 1906, have made a detailed study of the chemical basis of sweet, sour, salty, bitter, metallic, electrical and alkaline tastes.

Many other writers have made a physiological study of the action of this same series of chemical substances on the gustatory apparatus of man. This includes the work of Kiesow ('94b), Haycroft ('00a), Hanig ('01), ISTagel ('05) and Lemberger ('08), together with numerous others, practically all of whom, however, consider in this con

Sheldon, Reactions to Chemical Stimuli. 275

iiGction only the taste buds and associated nerves. Other authors have argued that, in addition to these structures, the nerves of general sensation take part in the sense of taste in man. Such a view is supported by the work of Caraerer ("70), von Vintschgau ('79b), von Anrep ('80), Adducco and Mosso ('86), Hooper ('87), Berthold ('88), Oehrwall ('91), Shore ('92), Kiesow ('94a), Vinci ('97, '99), Fontane ('02), Ferrari ('04) and Herlitzka ('07).

Extensive work has also been done on the sense of smell in man from the physiological, and to a less extent the chemical viewpoint, as may be seen by consulting bibliographies such as that given by Zwaardemaker ('95) and Bawden ('01). There will be no general consideration of the subject here, as it does not bear directly on the problem at hand. It is to be noted, however, that in connection with the olfactory organ as well as the gustatory, the free nerve endings take part in the reactions secured. Physiological evidence is noted by Haycraft ('00b), while the presence of such terminations has been demonstrated by a number of writers from Grassi and Castronovo in 1889 to Eead (1908).

In spite of the evidence presented by these authors, outside the single work of Grlitzner ('94), little has been done on mammals toward a study of the reactions of the free nerve termini generally to chemical stimuli. This has been due partly to preconceived ideas on chemical sense and partly to the feeling that nothing is to be gained by a general study of the chemical sense among the vertebrates, — even to the extent of including smell and taste under the same category. Such is the view of Zwaardemaker, who says ('03), "Bei den Wirbelthieren jedoch sind Geruchs- und Geschmackssinn in vieler Hinsicht so grundverschieden, dass es meines Erachtens keine Empfehlung verdient, sie zusammen zu behandeln."

On the chemical senses of the lower vertebrates little has been done. Bateson ('90a, '90b) discussed the senses of smell, taste, and touch in several fishes. The work is of little value in the present connection. In 1894 Kagel published his great monograph on smell and taste. Nagel repeatedly uses the term chemical sense, always meaning, however, the combined organs of taste and smell and not a general chemical sense. He stimulated selachians, teleosts, and am

2/6 Jonriinl of Comparative Neurology and Psychology.

pliibia by means of solutions of ditferout clieinicals. The selachians were very sensitive to weak stimuli, reacting to dilute solutions of vanillin all over the body, but to quinine only about the head. Barbus failed to react to salty, sweet, bitter, or sour substances on the general body surface, while Gasterosteus reacted to quinine only about the head. With Cobitus and Gobius he obtained reactions with meat juice and sugar solutions. Lo])hius was sensitive to chemical stimuli over the entire skin. Triton, the only amphibian tested, reacted on stimulation of the head only. It is evident from later work, particularly that of Hcrrick ('():>, '03c) and Parker ('07, '08a, '08b), that Nagel's i-esults did not permit the drawing of sound conclusions, partly because of the su1)stances used as stimuli, and partly because he failed to ditferentiate Ix'tween fishes with taste Imds on the outer surface and those lacking such structures with their visceral sensory innervation. Herrick ('02) was concerned almost exclusively with the sense of taste, in a narroAv sense, that is, the reactions to sapid solutions through stimulation of the taste buds. He performed a few experiments of a general chemical nature, insufficient, however, to permit any conclusions. Almost the only work on vertebrates which takes up the reactions of the nerves of general sensation to chemical stimuli is that of Parker ('07, '08a, '08) on Amphioxus and Ameiurus. His results will l)e reviewed later.

Considering that a general chemical sense is proliably more primitive in phylogeny than taste and smell and that a careful study of such a general sense may do much to make clearer the development of these two senses, as well as their physiology, it is strange that so little has been done on this topic. This is especially true in the lower vertebrates, where, in many cases, it is difficult to separate the reactions due to stimulation of the organs of taste and smell from those due to the nerves of general sensation.

COXDITIOXS oi' Exi'EKi:\rp:xTATiox.

The substances used in this work were hydrochloric, nitric, and

sulphuric acids for acid stimuli ; sodium, ammonium, and lithium

chlorides for saline stimuli; sodium hydroxide for alkaline; cane

sugar, dextrose, saccharine, and its carbonate for sweet; and quinine

Shkldon, Rcaitioii.s to Chetmcdl Stimuli. ZJJ

lijdrocliloride, picric acid, aiuiiioiiiuin and .sodium picrates for bitter. All were made up in distilled water on the basis of the gram-molecular solution. The inorganic acids were prepared as normal solutions, titrated against an alkali of known strength for accuracy. The other solutions were made up by weight, the concentration first used as a test depending partly on the solubility of the chemical used. The chlorides were i)repared as 5n solutions, the sugars 3n, sodium hydroxide as n, saccharine n/0, quinine hydrochloride n/10, picric acid and its salts n/15. In the experimental work all of these solutions were gradually diluted until the liuiit of reaction was reached. Sufficient time was giyen between tests at ditferent degrees of concentration and with ditferent substances to eliminate after-eifect.

A large number of dogfishes were used in the exj)eriments in order to rule out individual yariation, ]\rost of the normal fishes used were those caught in the fish traps and placed shortly in tanks about a meter and a half long, two-thirds of a meter wide and a third of a meter deep. A current of sea water was kept constantly running through the tanks. After a few days in these tanks the fishes could be handled with little difficulty. For most of the Avork, indiyidual adult dogfishes were removed and placed in a smaller trough about eighty cm. long, thirty cm. wide and fifteen cm. deep, through which a strong current of sea water was flowing. iVfter a little handling the animals would lie quietly in this trough either on the dorsum or venter, submitting to a certain amount of manipulation. In cases where it was necessary to have part of the animal out of water or where the fish was very unruly it was fastened to a frame.

Application of the Siimulus. — The solutions were applied by means of a pipette and were, in most cases, ejected slowly with the tip of the pij^ette about two millimeters from the skin of the fish. In such cases the fish was completely covered with water. Where it was essential that the region stimulated should be out of water absorbent cotton saturated with the solution was usmilly applied. Occasionally, however, the solution was ejected directly against the skin and the time of the tactile reaction taken, after which the slower chemical reaction could usually be identified. In stimulation of the mouth or nasal capsules a metal guard, closely fitted to the snout, was placed

278 'Journal of Co7nparative Neurology and Psychology.

between the two sets of apertures f)reventing diffusion of the stimulus from one set to the other.

Regions Tested. — ^With each solution used, approximately fifty places on the body were tested at each concentration used, and the time of reaction always recorded with a stop watch. These regions included the mouth, nostrils, spiracles, anus, claspers, and selected places on the fins, dorsal, lateral, and ventral surfaces of the fish. For the location of these points see figures 1 and 2.

Reactions Obtained. The reactions obtained varied according to the part of the body stimulated, as follows: Stimulation of the mouth or spiracles is followed by one or more violent gulps, accompanied, of course, by a quick ejection of water through the branchial openings. A more rapid respiration for a greater or less length of time, depending on the stimulus, follows this. This is the only reaction secured by chemical stimulation of the mouth and spiracles, and it is secured by stimulus of no other region. When the nostrils are stimulated by any of the substances used, the reaction is likewise very characteristic. It consists essentially of a very quick jerk of the head. This reaction is likewise secured by stimulation of no other region. In the case of the paired fins the characteristic reaction is a quick movement of the fins, usually of a vibratory type. Often, particularly if the stimulus is weak, the first reaction is a turning or movement of the whole fin toward the stimulus, occasionally away from it, followed usually by vibration of the fin. With the median fins the reactions are very similar. The more usual reaction, however, begins with the movement of the fin toward the stimulus. Often the small caudal finlets of the median dorsals and the anal fin will react by a rapid vibration, even though that part of the fin is not stimulated. If this reaction occurs, it usually begins by a quick movement of the finlet toward the stimulus, more rarely away from it. When the finlet of either the anal or second dorsal fin takes j^art in the reaction, the other does also, so that the action of the two is simultaneous and in the same direction. When the caudal fin is stimulated, the reaction consists in a rather slow sidewise movement of the tail either toward, or away from, the

Sheldon, Reactions i(j Cluvinrnl Sfi/niili. 279

sl.iniulus. This is evidently the beginning of n, swimming movement. Stimnhition of the anns results usually in bending over ventrad of the pelvic fins. Occasionally the fins react alternately in an attempt to turn the body over. If the. stimulus is strong or long continued, these reactions are followed by a lateral squirming of this part of the fish culminating in the swimming away of the animal. Stimulation of the claspers results in a (juick lateral movement and vibration of the structures. The head responds by a rather slow movement away from the solution. In general, stimulation of the dorsal, lateral, and ventral surfaces, other than those already mentioned, results in a movement of the fish which is very evidently a part of the general swimming movement. In fact, stimulus of almost any region of the fins or body, if persisted in, will transfer the local reaction to one which forms jiart of the swimming movement of the animal. This is shown especially in the case of stimulation of the fins or lateral surfaces. If the caudal, second dorsal, or anal fin is stimulated and the reaction is toward the stimulus, for instance, there will often be a movement of the first dorsal fin but in the opposite direction. The same relation holds true if the first dorsal fin is stimulated. Often a reaction of all the fins is secured. For example, there will be a movement toward the stimulus by the caudal, anal, and second dorsal fins, a movement away from the stimulated side by the first dorsal fin, an u])ward movement of the jiaired fins on the side stimulated and a downward movement on the opposite side. This reaction was first pointed out to me by Dr. Parker as a response secured by tactile stimulation of the same regions. Such reactions are unquestionably part of the general swimming movements of the fish, as may be seen by observing the animal in an aquarium. As caused by chemical stimuli, they are evidently of a kind to preserve the fish from injury, enabling it to remove itself from an injurious environment. Very similar reactions are secured by stimulation of the sides of the body and tail. In general it can be said that a slight stimulus calls forth a local response, while a stronger or longer-continued stimulus almost invariably results either in a new reaction which is part of the swimming movement, or else in a gradual change of the local reaction into such a part of the swimming movement. The former occurs where

28o Journal of Comparative Neurology and Psychology.

the local reaction differs decidedly from the swimming movements, as in the case of stimulation of the moutli, while the latter holds in cases where the two are similar, as in the reactions due to stimulation of the tins. Certain interesting special cases are to he noted. If the dorsum or side of the fish near the small iinlets of the dorsal or anal fins be stimnlated, a, quick movement of the finlets toward the side stimulated, usually followed by vibration of the finlet, occurs. This may be a reaction to remove an irritant, as is noted in the case of the frog when a droj) of acetic acid is placed on the skin, or it may be i^art of the swinuning movement, as seems more probable. Evidence against the former interpretation is ottered by resnlts which Parker obtained by tactile stimulation. He found that tactile stimulation of the dorsum near the fiidet of the second dorsal fin caused this reaction ; but he also found that if he now stimulated a point between the finlet and the mid-dorsal line on the same side the finlet continued to wipe the skin, but ventrad of the point now stimulated. It would, therefore, appear that the reaction is called forth by stimulation of any part of the side in this region and is not a local response to remove an irritant. It might be argued, however, that the power of localization is not well developed in this form. The strongest evidence in favor of the second interpretation is that when the skin beside both dorsal finlets is stimulated on the same side at the same time one turns to one side and one to another, as is the case when the aninml is swinuning. This reaction of the finlets was one of the most delicate found. Reactions could be secured by stinuilation of the skin beside the second dorsal finlet when all the remainder of the body was insensitive.

If the claspers are turned to one side and the venter underneath stimulated, a quick vibration of the claspers over this point follows. This is i^robably })art of the general swimming movement also, although its constancy and accuracy suggest the wiping reaction. In the case of the pectorals, however, when the ventral surface between them is stimulated there follows a quick scissors-like action of the two fins over the point stimulated. If one fin is held, the reaction takes place with the other alone. This reaction is not a part of the general swimming movement, is very consistent and accurate, and apparently is of the same character as the wiping reaction of the frog. It is

Shf.LDOK, Rrnctioiis to CliriiiKnl Sfinnili. 28 1

]irol)al)ly })ur])osofiil only in the sense tliat; sncli a reaction is of a general i)reserval ive cliarader such as is discussed by Sherrington ('()G), This reaction often occurs, also, when tlie ])ectoral itself is stimulated, ])artienlarly on its median margin.

SeNSITIV' KNKSS 'i'O CuKMldAl. StIiMUI.I.

Expcrlmcuial Besulis. — The least stimulus which will canse a reaction, the comparative sensitiveness of dilferont parts of the body, and the time of reaction for the different snbstances nsed are shown in the tables. The data were obtained under the following conditions. Several aninuils were always used for the tests and the figures given are based on results obtained from two or more individuals. From three to five tests were made at each point stimulated, with each solution used, and at each different degree of concentration of that solution. When individuals were used as controls, however, fewer tests were made if these demonstrated that the reactions were in conformity with those first obtained. Before the solutions were applied, both distilled and sea water were used to make certain that no reaction would result from their use, exclusive of the test solution. So much variation in the reaction time between different individauls wascioted that both upper and lower limit in seconds are stated for each point tested. These limits differ considerably in many cases, yet it is easy to see that there is a general difference in the reaction time for different regions of the body and for different degrees of concentration of the solutions. For all tests on the dorsal or lateral surfaces, the fish lay on the venter ; while for experiments on the ventral surface, it lay on the dorsal or dorso-lateral aspect. About ten of the points stimulated are omitted from the tables.

Analysis of Results. — It will be noted that the same reactions are secured by the use of any of the inorganic acids used as stimuli, that is, the reactions are due to the hydrogen ions. The reactions to the acids in the more concentrated solutions are very strong and definite. In nearly every part of the body they take place as quickly as mechanical conditions will permit, that is to say, almost instantaneously. With the decrease in concentration the reaction time becomes a trifle longer. Practically the entire body is sensitive to n/20 acid, the head

282 yourimJ of Coinparative Neurology and Psychology.

Pectorals.

Pectorals.

Pelvics

Pelvics.

First

Subs.

Base.

Margin.

Base.

Margin.

Mth.

Spir.

No.«t.

Anus.

Cone.

Dors.

Vent.

Dors.

Vent.

Dors.

Vent.

Dors.

Vent.

Base.

1

2

3

4

5

6

7

8

9

10

11

12

13

HCl,

HNOi,

H..S()4,

n/1

1-3

2-3

2-3

1-4

2-4

2-4

2 3

1-3

1-4

1-3

1-3

1-2

1-4

11/ 10

2-3

2-4

2-7

2-5

2-5

2-0

2 (i

1-6

3-5

.3 5

3 5

2-5

10-14

n/20

2-6

2-4

2-7

2-8

2-5

2-0

2-6

1-6

3-5

3-8

3-7

2

n/40

2-6

2-4

3-0

3-0

3-8

3-8

n/50

2-6

2-5

8*0

87-0

n/75

2-4

4-0

2 5

n/1 00

NaOH,

n/1

n/.5

2-3

2-5

n/10

3-4

5-7

4-10

10-14

5-8

5-7

4-8

.3-12

7-13

3-7

3-9

9-13

n/20

5-9

4-13

6 9

3-8

8-15

6-7

.3 14

12 19

11-0

n/.3()

5-0

12*18

9-12

5-12

11/40

11/50

n/60

n/70

Li 01,

NH4CI

5n

5-0

7-9

2-8

4-10

25-0

10-30

7-32

10-17

8-14

14-20

7-10

11-16

9-29

2n

10-15

5-10

7-18

10-18

8-17

19

11-15

7-15

11-22

27-29

n/1

Quin.

h.chl.

n/10

2-3

3-4

5-0

n/20

2-4

2-3

n/30

2-3

6-7

n/40

3-0

Pier.

acid.

1

n/15

1-2

2-5

7-8

8-10 10-14

7-12

7-14

10-14

10-21

4-11

8-12

15-37

n/30

3-4

2-3

1-4

22-24

X

6-19

%■

14-58

t

11-14

X

18-0

X

n/60

1-4

5-0

2-14

n/1 20

?

?

?

Am. and

Sod.

pier.

n/1. 5

2*10

10*20

n/30

Sacch.

n/6

^

X

X

X

-X

X

X

X

X

X

X

X

X

Carb.

of

Sacch.

n/6

Cane

sug. and

dext.

3n

Sheldon, Reactions to Cheimcal Stnmili.

283

Do us

XL,.

Second Dorsal.

Anal.

Caudal

H

Snout.

Dors, tip. 14

Finl. 15

Base. 16

Dors, tip. 17

Finl. 18

Base. 19

Vent, tip. 20

Finl. 21

Base. 22

j Vent.

marg.

23

Tip. 24

Dors. 25

Vent. 26

Dors. 27

Vent. 28

1-4

3 6

3-8

1-3

4-5

4-8

4-0

2 4 3-5

3-8

(1

2 4

6-8 6-0

1)

13 5-6 5-9

13 2-6 3-8

1-3 6-8 10-12

2 3

2-4

4-10

1-3

4-7

4-11

6-0

2-4 5-7

1-4 3-7 3 12

4-8

3-4

5-7

5-8 1

1

4-5

5-0

5

7-8 7-0 7-0

i

i

4-14 8-13

14-18 10-0

7-14

4-12

8-10

5-7 8-13

7-9

8-18

6-14 12-13

9-16 6-10

8-14 5-9

6-0

1 5-0

5-0 7-0

5

12-36 6-36

12-20 12-0

12-19 13-25

9-19 19

8-16 6-32

8-21 7-14

6-24 18-20

6-15 6-19

11-23

15-22

8-25 10-21

4-0

7-13

7-0

7-0

15-37 t

12-18 t

9-27

t

2-10

X

7-13

t

14-22 t

11-15

t

6-11

X

12-16 X

8-22

X

9-35

14

42

1

X

X

X

X

X

X

X

X

X

X

X

X

X

X ,

X

1

1

1

284 'Journal of Comparative Neurology and Psychology.

Subs, and

Dorsal Aspect

L

ATERAL

Aspect.

Ventral Aspect.

Ceph.

ID

Betw.

2D

Vent.

Mid

Anal.

Tail.

Betw.

Mid

Clas

Dors.

Caud.

Cone.

of ID

Finl.

Dors.

Finl.

of ID

dle.

Finl.

pect.

dle.

pers.

of cl.

of cl.

29

30

31

32

33

34

35

36

37

38

39

40

41

HCl,

HNO3

H.SO,

'

2-3

1-4

1-3

1-2

1-3

2

n/1

1-4

1-4

3-6

1-2

1-4

2-3

1-2

2-7

3-6

3-5

2-4

2-6

n/10

3-5

2-3

4-8

2-4

3-7

2-7

1-2

2-6

3-5

3-8

2-7

2-8

n/20

3-5

3-4

4-8

2 6

3-4

2 6

n/40

4-0

17*0

9*78

n/50

18*22

18*0

n/75

18*30

18*30

n/100

NaOH

n/1

n/5

5-15

5

n/10

6-9

4-10

6-9

6-11

8-17

8-17

5-8

5-8

4-0

5-15

10-14

n/20

9

4-13

n/30

n/40

n/50

n/60

n/70

LiCl,

NHiCI

5n

12-21

5-24

7-14

15

7-26

5-12

7-27

13-27

5-14

7-8

20-37

2n

25-0

7-0

25-0

5-32

10-15

11-0

6-16

10-22

8-18

n/1

Quin.

h.chl.

n/10

n/20

n/30

n/40

Pier.

acid.

n/1 5

6-35

5-8

10-12

5-8

2-10

0-10

7-10

8-11

3-10

12-15

n/30

20-40

20

20

20

t

23 29

t

9-27

t

12-42

22-45

n/60

n/1 20

Am. and

Sod.

pier.

n/15

10*20

0

10*20

10*20

15*30

15*20

n/30

Sacch.

n/6

X

X

X

X

X

X

X

X

X

X

X

X

X

Carb.

of

Sacch.

n/6

Cane

"

sug. and

dext.

3n

°

Sheldon, Reactions to Cliomcdl Sitimili. 285

SIGNS AND ABBREVIATIONS USED IN THE TABLES.

Numbers at the heads of columns refer to the points stimulated as shown on Figs. 1 and 2. Other numliers indicate the reaction time in seconds.

— is equivalent to the word to.

signities (hat a reaction wiis secured with the region stimulated out of water, according to some of the methods already descrilied. The reaction time in such cases was rarely taken. The exceptions are indicated by the rei>lacement of — by *.

I refers to cases in wliicli tlie reaction was secured with the region tested under water, the reaction time not being taken. This sign usually signifies that the reaction was very weak and incomplete.

X is used in one case where the reaction was very strong and definite, but where the reaction time was not taken.

indicates that no reaction could be secured with the fish, either under or out of water. When used after another number it signifies that no reactions were secured in two or three of the five tests made.

Where blank spaces occur no tests were made.

Am. and Sod. trier., ammonium and sodium picrates; Betiv., between; Cane siig. and dext., cane sugar and dextrose; Carh. of Saeeli., the carbonate of benzylsidphonic amide formed by the neutralization of saccharine by sodium carbonate; Caud.. caudal or caudad ; Cepli.. cephalad; cl., claspers ; eone.. concentration; Dors., dorsal or dorsals; flnl., finlet, caudal prolongation of anal and dorsal fins ; niarg., margin ; 3Ith., mouth ; Nost., nostrils ; peet., pectorals; Pier. aeid. picric acid, trinitrophenol ; Quin. li.ehl. quinine hydrochloride; Spir.. spiracle; Suhs., substances used as stimuli; Vent., ventral or ventrad; ID, first dorsal fin; 2D, second dorsal fin.

iK'ifig the only part at all iiisonsitive. At ii/40 the body surface g'Cficrally reacts, although the reactions are less definite, particularly on stimulation of ninch of the dorsal surface. At n/50 the inouth, s])iracle, anus, nostrils, fins, claspers, and side are still sensitive, as is also the dorsal surface beside the finlets. Stimulation of the mouth, spiracle, claspers and skin beside the finlets by n/75 still calls forth a reaction. At n/100 no reaction could be obtained, although the fish in many cases seemed to perceive the stimulus. In all this work many observations were made which indicate that the dogfish actually feels stimuli to which it does not react to an appreciable extent. One probably comes to a point in decreasing the concentration of the solutions where the stimulus is perceived yet not sufficiently strong to cause action of any kind on the part of the animal. All of the reactions secured at n/40 or less were weak, although usiuilly definite. The reactions to acids in general are characterized by their quicktiess

286 "Journal of Comparative Neurology and Psychology.

and definiteness. There are rarely premonitory symptoms of auy kind before the reaction takes place, even thongh the reaction time is long. In snmmary, it is evident that the dogfish is sensitive to acids of a solution of n/75, both in the mouth and spiracle in which are found taste buds^ and also on the outer body surface when no such structures are found.

J^aOII is the only hydroxide given in the table, although experiments were made with KOH which indicate that essentially similar reactions would be secured by its use. To NaOH the dogfish reacts quickly and definitely in the stronger solutions, although not quite so quickly as to acids. It will be noted likewise that the animal is not sensitive generally to so dilute solutions. Reactions are secured from the general body surface, however, to a solution of n/70. It is of special importance to note that the mouth and spiracle are almost insensitive to alkalis except in very strong concentration. Some slight chemical reactions take place between the hydroxide and the sea water, as was the case with the acids used also. This probably renders the solutions less powerful than they would otlierAvise be.

To salts the reactions are slower than to acids and alkalis. The responses to both lithium and ammonium chlorides are practically the same. IvTo reactions to sodium chloride could, however, be obtained. This is due, doubtless, to its })resence in such quantities in the sea water. The reactions to the chlorides are usually preceded by premonitory symptoms a few seconds before the definite reaction occurs. These consist of a local or general uneasiness. The reactions are also often prolonged, continuing for a few seconds after the stimulus is removed. It will be noted here, as in the case of the alkali, that the general body surface is more sensitive than the mouth. Definite reactions are secured from the former to a 2n solution, while the fish shows a sensitiveness to a normal solution. The mouth and spiracle, however, react very weakly to solutions as strong as 5n, and almost never to a lesser degree of concentration. The general lack of effect of the solutions of a weaker gi-ade than iioniial is ])r()bably due to the fact that the salts of the sea water make it about a 5/8 normal solution. This inter])retation is supi)orted l)y the results when NaCU was used nnd ;dso l)y the tests made by Parker ('08b) on the fresh

Sheldon, Reactions to Chemical Stimuli. 287

water catfisli (Ameiurus), which is quite sensitive to salts. The reaetious secured by the use of salts are not due to osmosis, as sugar solutions of equal osmotic strength have no effect.

Reactions to quinine hydrochloride take place oidy in the mouth and spiracles, that is, reactions to quinine take place only on stimulation of surfaces hearing taste buds, as Parker found for the cattish. The dogfish is extremely sensitive to picric acid. In strong solution it is extremely distasteful and the aninnd responds vigorously. The response is slow, however, and there are usually premonitory symptoms before the reaction, as noted for salts. Reactions are also often })rolonged after the stimulus is removed. The mouth and body surface are sensitive to n/(30 unquestionably, w^hile an apparent uneasiness of the fish seems to indicate that it feels in the mouth, spiracles and nostrils a still greater degree of dilution. Picric acid was used by Parker ('07, '08b) as a bitter stimulus. It was with this idea in mind that the substance was used on the dogfish. It might be argued, nevertheless, that inasmuch as aquatic forms are very sensitive to acid stimuli, the reaction in this case is due to the acid radical in the trinitrophenol rather than to the base which gives to us the bitter taste of picric acid. To test this point neutral ammonium and sodium ])icrates w^ere used. Both of these are as bitter to the human taste as is picric acid. From the tables it will be seen that the fish is by no means as sensitive to these as it is to the picric acid, although weak reactions can be secured by the use of strong solutions. The results show that it can not be assumed that the stimulus of picric acid rests entirely or even largely with the base, but that the acid radical is ])robably responsible almost entirely for its influence on fishes. It is evident, however, that the base does possess a stimulating power, as slight reactions were secured to the neutral ])icrates. It can be stated, therefore, that the dogfish reacts to substances which give us a bitter taste. To stimuli of this character the mouth is more sensitive than is the remainder of the body. On the whole, the fish seems less sensitive to l)itter substances than to the other kinds of stimuli used. Outside of |)i('i'ic acid the animal showed little distaste for the solutions used, even ihongli reactions were secured.

]^o reaction al all could be obtained to sugars. This holds true

288 "Journal of Comparative Neurology and Psychology.

for all aquatic vertebrates, due probably to the fact that sweet substances are substantially unknown to aquatic life. To saccharine or benzylsulphonic amide very quick and definite reactions were always secured. Thinking that this reaction was probably due to the acid radical, the saccharine was neutralized by sodium carbonate, the resulting product l)eing as sweet to man as is the saccharine. No reactions at all wei-e then secured, proving that those first obtained were due to the acid radical.

Comparing different regions of the body as to sensitiveness, it will be noted that the head is least sensitive, Avhile the mouth, nostrils, paired fins, particularly the pelvic, the anal and dorsal, especially the second dorsal, are more sensitive. Areas of skin closely associated with the dorsal and anal finlets are included with these fins.

OrERATIONS.

Some operations were next performed to find out what part of the nervous system takes part in these responses. After operations no fishes were subjected to experimentation until at least twenty-four hours had elapsed. I^arcotization was accomi:)lished by a mixture of ether and water from the effects of which the animals did not appear to suffer in any way.

Destruciion of the Spinal Cord. — In this experiment the tail was cut off", the caudal artery and vein plugged with cotton and the cord entirely destroyed as far cephalad as the cephalic margin of the first dorsal fin by means of a small steel wire. This method was suggested to me by Dr. Parker, who had used it with much success. By this operation the peripheral innervation of all of the caudal part and middle of the body is destroyed, except that from the lateral line nerve. The individuals subjected to this operation lie in the water perfectly motionless so fai- as the caudal part of the body is eoucerned, occasionally trying feebly to swim by meaus of the })ectoral fins. The skin caudally gradually turned white as is the case with dead dogfish. Such fishes ]i\'ed, however, for some weeks. As was to be expected, no reactions could be obtained chemically either by stimulation of the general body surface or the lateral line. Parker ('05) showed that the function of the lateral line is to respond to slow mass movements

Shkldon, Rractious to Clu'iiiirnl Sfiiiuili. 289

(»!" Wiilci-. The lu'ud is sensitive to elieiiiical sliinuli after the operation as before.

Scdioti of lite Spinal Cord. — The cut was made about two centimeters eeplialad of the first dorsal fin. After the operation the fishes lived for weeks, differing' from the normal individuals in their ordinary actions only in the fact that the caudal part of the body kept up a constant swinnning. The fish would be propelled to the side of the tank, where it would remain for hours keeping up a vigorous swimming movement of all the body caudad of the cut. This was never observed to cease until the death of the animal. Bethe ('99) noted this same swimming after section of the cord in the dogfish, but found that there were intervals of cessation. When this fish was tested with the solutions used on the normal fish, it was found that the caudal part of the body was more sensitive than before. The reactions take place from a fraction of a second to a few seconds quicker than before, the reactions seemed more positive and definite and some of the fishes studied reacted to slightly weaker solution. The reactions of the head were practically normal. After most of the operations many of the fishes failed to react to quite so weak a solution as before, due, probably, to the shock of the operation or to diminished vitality. The reactions of the spinal fish are almost never long continued after stimulation with salts or picric acid, contrary to the condition in the normal animal.

Several points are involved in these experiments. These are, how does the spinal fish difter from the normal and why ? Is the observed difference due to a lack of inhibition by the brain, the use of the new paths for the refiexes, or stimulation of the cut ends at the cephalic end of the cord? Taking up the first: Danilewski ('92) on Aniphioxus found that "voluntary" movements ceased with removal of the rostral part, the only action recorded taking place in response to definite stimuli. Tie coiududed that the centers for movement lie in tlie rostral part of the animal. Steiner ('88) worked on the lamprey, selachians, ganoids, and teleosts. In the lamprey he found no voluntary movement of the caudal part after section^ while the other fishes acted about as usual after section of the; cord except for the dead weight of the head. Steiner believed that the centers for equilibrium

290 'Journal of Coinparative Neurology and Psychology.

and voluntary iiiotioii lie in the cord. Loel) ("1>1 ) tViuud that a dogfisli with the cord cut was no hunger able to orient itself, even though it could swim oti" readily. Bethe ('00) observed no difficulty in the swininiing movements of the s|)inal dogfish. Pfiiiger ('5o) and Bickel ('97) took up the study of the spinal eel. The former found that the eel had no difficulty in nniintaining its equilibrium and could swim readily, while the latter says that the spinal eel is incapable of remaining in the normal ])osition and can swim only forward. The normal individual can swim in both directions. The spinal. frog has been studied with great care. Pfiiiger ('53) found that the spinal frog still retained its muscle tonus and could spring; Mendelssohn ('82, '83, '85) found the refiexes good after section of the cord; Steiner ('85) observed no change in the reactions; Schrader ('87) obtained similar results. Moore and Oertel ('!)!)) found, however, that the reflexes were increased, although fatigue occurred quickly, Babak ('03, '05, '07) found that section of the cord in the larval frog resulted in no loss of ])ower in the spinal part. In regard to mammals, ])articularly man, there is a mass of experimental and clinical evidence on the results of the separation of the spinal cord from the brain. This is presented mainly by Kosenthal ('73, '84) ; Bastian ('90) ; Burns ('93, 'OG) ; Rosenthal and Mendelssohn ('97) ; Senator ('98) ; Brauer ('99) ; Moore and Oertel ('99) ; Walton ('02) ; Walton and Paul ('06), Sherringtou ('97, '00b, '05, 'OOa, '0Gb); and Sherrington and Laslett ('03 ). Tn mannnals, as observed by most investigators, the loss of reflex i)ower is very great, often practically total after a break in the functional continuity of the cord and l)rain. It is evident, therefore, as is ('m])hasized by Sherrington, Walton, and Walton and Paul, that in the ascent of the animal scale the cord loses, its power as an automatic center and becomes less and less capable of responding to stimuli.

In the fishes undei- obscrxation in these ('X])('riiiiciits no lack of the power of equilibrium was ol)served. This may easily bo due to the fact that the innervation of the ])ectoral fins was left intact. This ])robably accounts foi- Ihc dilferent resnlts ol)laiiic(| l)y kocb and Bickel.

The greater activity and sensitiveness of the caudal jtart of the

Shhi.DON, Rcoriioiis to Chrinirdl St/t/iuh. 291

fish after section of t\\c cord nvo duo, in all ])rol)al)ility, to a removal of inhibition fi-oni ihc higher cenlci's, hy the nse of slioi'ter paths, or to hoth. Pik(^ ('*'^j heliex'es that the second is the triu; explanation, lie considers that norniallj the long paths by way of the brain are made nse of. After section, pathways through the cord must serve, giving a shorter arc which requires less time for the transmission of the imi)nlses. This is sn])])orted by the conclnsions of jNIoore and Oertel who suggest, however, that the higher centers nuiy have a regulatory rather than an inhibitory influence over the cord. Walton and Paul think that the cord is particularly for active, instantaneous and violent reflexes. It is prol)al>le that use of sliorter ])aths by the reflexes has much to do with the resnlt secured, but the evidence presented in the literature is strong in su])port of the view that the l)rain possesses an influence over the cord of a regnlatorv ty])(>, at least.

Wiping movements of the fins and finlets occur as in the normal aninnd. This is in line with the results in the spinal frog and can not, in either case, be considered purposeful, as Howell ('05) and Sherrington ('06b) clearly show. Such movements in a spinal animal are simply such as are of a general protective nature for the normal animal or else form a part of the habitual action of the organism.

'No spinal shock was observed. This is as would be expected. Steiner, Loeb, Bickel, and Bethe do not mention it for fishes. Moore and Oertel show that it is slight in the frog and passes off quickly. In mammals, as is well known from the work of Bastion ('90), Burns ('93), Sherrington ('1)7, '00b), Senator ('98), Babak ('07) and Pike ('08) the inflnence of shock is greatest. These citations fall in line with the statement made above to the effect that the power of the cord as an automatic center decreases as one ascends the animal scale. Probably in the dogfish many of the reactions of the caudal part of the animal take place normally through the cord, so'that section of it brings about little change.

Tlie Innervation and Bear I ions of ilie Xosfrih. — It will be noted from the chart that the nostrils ai-e very sensiti\'e to the solutions used as stimuli. The reactions obtained were also very decided. The question arises as to whether these are due to the olfactory nerves or the nerves of general sensation. Accordingly, different nerves were

292 ^ournnl of Cojjiparntivc Neurology mid Psychology.

out in order to clear up this point. The only nerve hitherto known to 8U23ply the nasal nuicosa in the dogfish is the olfactory. It has long been known that the ti-igeniinus nerve snpplies this region of the head for tactile sensation, l)nt none of its fibers have been traced to the olfactory eaj)snle.

Aichel ('95) fonnd ti'igeniinal fibers in the olfactory mncoiis membrane of embrj'onic telcosts, bnt was uncertain as to their derivation. Jagodowski ('01) observed the same fibers. Later, Sheldon ('08) demonstrated the existence of fibers, derived from the trigeminal nerve, in the mucous membrane of the adult cai-p. Eetzius ('92) in the frog and Rnbaschin {'iVZ) in the chick obtained similar results. In the lower mammals and man such an iiniervation has been demonstrated by Grassi and Castronovo ('SO), Eamun y Cajal ('89, '93), van Gehuchten ('90), von Brunn ('92), von Lenhossek ('92), Retzius ('92), Disse ('94) and Read ('08).

In the dogfish the trigeminal nerve is divided into four rami. These are the ophthalmicus superficialis to the dorsum of the snout, closely associated with the ophthalmicus superficialis facialis of the lateralis system ; the ophthalmicus profundus to the side of the head and snout; the maxillaris to the upper jaw and the mandibularis to the lower. These latter two are bound together for some distance from the brain in close association also with the buccalis. All of these trigeminal rami are general sensory with the exception of the mandibularis which is coml)ined general sensory and motor (Strong, '03). The relations of all these rami except the ophthalmicus superfacialis are shown in fig. 3.

The olfactory crura were first cut in an endeavor to destroy the sense of smell. At first the section was made from the dorsum, but this operation was usually fatal. Such fishes would either die inside of twenty-four hours or else react very feebly. Finally the method of Lyon ('00) was adopted. The fish was etherized, the mouth pried open as far as possible and the olfactory crura reached -by means of an incision in the roof of the mouth. The maxillaris and mandibularis were also reached in the same way (see fig. 3). Before the cnt was made the mucosa was reflected as shown in the figure, then a piece of cartilage was removed, care being taken not to cut any of the

Shki.don, Rrtiit/oiis to (^liruiicnl Stiiim/i. 293

blood vessels of the roof of the mouth, as this leads to serious blcediug. The loeatiou of the vessels to be avoided is shown in the dissection. The fishes seemed to suffer no ill effect from this o})eration and would live for some weeks thereafter. The profundus and superficialis nerves were cut by means of ;in incision into the orl)it at the caudal margin of the eye l)alb

After section of the olfactory crura, stimulation of the nostrils causes the same reactions as were obtained for normal fishes. It is, therefore, evident that these reactions are not due to stimulation of the olfactory nerve. When the four rami of the trigeminal are cut and the olfactory left intact, no reactions are secured. This shows clearly that the reactions obtained are due to the nerves of general sensation. The associated nerves of the lateralis system have been proved to be insensitive to chemical stimulation. The following nerves were next cut without destroying the reactions ; the ophthalmicus sujDerficialis, the ophthalmicus profundus and the oph. sup. and oph. prof, together. When the nuixillaris-niandibularis trunk is cut, however, all reactions cease. When normal fishes are taken, this trunk cut on one side and not on the other, stimulation of the nostril on the operated side calls forth no responses, while the nostril on the normal side is as sensitive as before. As the mandibularis nerve goes to the lower jaw excdusively, it is evident that the sensitiveness of the nostrils to the chemical stimuli used is due to the maxillaris nerve. These experiments also show that the nostrils in selachians are innervated by the trigeminal nerve, as is the case in most other vertebrates.

Some odorous substances were also used in the nostrils although without results so far as the sense of smell is concerned. The substances used were the oils of cloves, pennyroyal, thyme and aniline oil. A few cc. of the oil were placed in a 200 cc. flask of distilled water and shaken violently. After a day or so most of the oil would collect while the remainder remained in suspension in the water in the form of small drops. This water was drawn off and called a saturated solution. ISTormal fishes reacted very quickly when this solution is applied to the nostrils. When the olfactory crura are cut the nostrils remain as sensitive as before. To clove oil, for instance, both normal and operated fishes are sensitive to a solution of about 1/100

294 Journnl of Couipnrntivc Neurology nud PsvcJioJogy.

'saturated. These results sliow that here as in man the trigeminal ner\T' shares in what we nsnallj call the sense of smell.

(^iiEZsrifAi. Skxsatiox as a Sense Quality.

Experiments were next performed to ascertain if these reactions to chemicals are due to the stimulation of nerve endings distinct from those excited by tactile stimuli. First a given region of the body was fatigued for tactile response. One method used was to keep a steady stream of water from a ]>ipette playing on some sensitive portion of the skin. After fi\'e minntes, or such a matter, the fish will no longer respond. Chemical reactions can always be secured thereafter in a little more than the ordinary length of time. When the experiment was reversed, however, the results differ. A given region may be fatigued in from five to ten minutes for any kind of chemical stimulus, acids, for example. It will then react to salts, for instance, but not to tactile stimuli for some minutes, if care is taken not to get outside the area fatigued. A blunt point was often used also to fatigue the tactile organs, with the same result. Parker (Y)8b) found in Amjihioxus that when the animal was fatigued for mechanical stimuli it still reacted to chemical stimuli, and vice versa.

Further a 2 per cent, solution of cocaine on absorbent cotton was ap]died to the skin and the region stimulated every few minutes for both tactile and chemical purposes. TTuder these conditions tactile response disappears in from ten to twenty minutes. Tiesponses to chemical stimuli can be ol)tained thereafter, although the reaction is slower than usual. Finally, response to chemicals disappears, the first to cease l)eing that to bitter substances, as is the case in mammals in the mouth, after Addncco and Mosso ('00) and Fontane ('02).

(ViXCEFSIOXS.

The results obtained show that the reactions of- the dogfish to substances which we call sour, alkaline, salty, and bitter are obtained by stimulation of the nerves of general sensation. One is not justified in saying that the taste buds and taste nerves (the facialis, glossopharyngeus and vagus) do not take part in these reactions in the mouth. It is certain, however, that such nerves do not

Shfldon, Reactions to Chemical Stimuli. 295

sliaro in the responses secured bj stimulation of the nostrils and body surface generally. It has been shown above that the reactions secured from the former are due to the nerves of general sensation. No gustatory nerves, such as Herrick ('01, '03b, '03c) described for the siluroids, innervate the trunk of the dogfish and there is no evidence that visceral sensory nerves such as Herrick ('99) found in Menidia reach the skin of the trunk in this form. It can be said, therefore, that in the dogfish the reactions secured from all parts of the body, excepting the mouth and ]n-obably largely from that, are due to stimulation of the nerves of general sensation. It is appai-ent, however, that the taste buds and nerves are concerned with the responses to bitter substances. These conclusions are supported by Parker's work on Ameiurus, where he secured reactions to this same series of chemical stimuli after section of the recurrent branch of the facialis. His work shows that in the catfish the free nervetermini react to chemical stimuli, even of a very weak character.

Griitzner ('94) found that the nerves of general sensation in man are very sensitive to chemical stimuli, experimenting on himself and the frog with acid, alkaline, salty solutions, etc. This is the only work dealing with this chemical sense in man. Haycraft ('OOli) comments on the well-known fact that the sensitiveness of the nasal mucous membrane of man to ammonia and chlorine is due to these same nerves. Camerer ('70) obtained taste reactions from papilla-free regions of the mouth. A little later von Vintschgau ('79b) stated that the nerves of general sensation take part in the sense of taste in man. Shore ('92) followed with the argument that salty and sour tastes are largely due to the general sensory nerves. This is based mainly on the fact that by the use of gymnemic acid he rendered the tongue insensitive to bitter and sweet substances without the loss of acid, salty, or tactile sensation.

He also found that the perception of acids goes fairly well in hand with tactile sensation. In view of the results following the use of picric acid and saccharine on the dogfish, it is interesting to note that Shore obtained no reactions to these substances in man aftei' the use of gymnema. This shows that in man picric nd

296 Journal of Comparative Neurology and Psychology.

and saccharine stimnlate only through their hitter and sweet properties and not, as is largely the case in the dogfish, by means of the acid radical. Kiesow ('94a) found that he could separate the tactile from the gustatory sense in the reactions to acids and salts. In 1903a he added that some tastes are mixed with pressure. Ilerlitzka in 1907 insisted that metallic taste is partly tactile, while in 1908 he concluded that it is entirely due to smell and touch. These citations show that even in man the nerves of general sensation possess the ]iower of reacting to chemical stimuli to a far greater extent than is usually considered to be the case. In addition to the special sensitiveness of certain portions of the outer skin to chemical stimuli the nerves share in what we are accustomed to call the sense of smell and jdav a large part in the sense of taste. They i)rol)ably have much to do with the reactions to acid, alkaline, and salty substances, at least, although such work as that of Kiesow ('98) shows that the taste buds and nerves certainly take some part in these responses. Kiesow found that individual papilla? react to acid and salty tastes. The work on the dogfish is in line with such an interpretation, as here the body surface, innervated by the nerves of general sensation, is especially sensitive to acid, alkaline, 'and salty sulistances, while bitter stimuli affect chiefly the mouth.

We now come to the question as to whether or not these chemical reactions are due to a sense quality distinct from the general tactile sense. It is generally assumed that there are several distinct tastes: sour, salty, bitter and sweet, at least. The work of Kiesow ('98) on the different papilla3 of the tongue, and that of von Anrep ('80), Hooper ('87), Berthold ('88), Oehrwall ('91), Kiesow ('94a); and Vinci ('97, '99) with cocaine, gymnema and allied substances support the view that the different taste qualities can be separated from one another. Practically all authorities are agreed in believing that there are likewise in man several cutaneous sense qualities. It is not essential for the point at issue whether one follow the dicta of the Elix, Goldscheider, von Frey school or arrive at the conclusions expressed by Xagel ('05) ; or, on the other hand, assent to views of Head, Rivers, Sherren and

Sheldon, Reactions to Chemical Stnniili. 297

Tlionipsoii ill their rccout work. So far as the hnver vertehrates are coneerued little work has heeii done on the siihjeet. Von Anrep ('80) found that tactile reflexes disappeared before chemical, following the use of cocaine on the skin of the frog, while Parker in his work on Ainphioxus and Auieiiirns showed that such a view is fully supported ])}' the results obtained. It may, therefore, be said that workers are agreed in assigning to the gustatory and general cutaneous senses several sense qualities and that this work falls in line in stating that, in the dogfish, the results secured by the use of cocaine and by fatigue suggest that a separate mechanism exists for the reactions to chemical as distinguished from general tactile stimuli.

From the evidence given we may say, with little hesitation, that there exists a general chemical sense from the protozoa to man. This general sense is the only chemical sense present in the lower invertebrates; as we ascend in phylogeny, however, there are probably developed from it, as Ilerrick ('08) suggests, smell and taste. The undifferentiated sense remains as the sensitiveness to chemical stimuli exhibited by the nerves of general sensation in the lower vertebrates and man.

The association of chemical sensibility with the nerves of genseral sensation need not in the least militate against the functional analysis of the nervous system as elaborated by Herrick ('O^a) and Johnston ('02, 'O(i), as it has never been assumed by them that sensitiveness to chemical stimuli must be confined to the nerves and organs of smell and taste.

Summary,

1. The smooth dogfish, Mustelus canis (Mitch.), is sensitive to chemical stimuli over the entire body surface, mouth and nostrils, responding by reactions which are characteristic for the dift'erent regions stimulated.

2. All parts of the body are very sensitive to acids and alkalis in very dilute solution, less sensitive to salts and l)itter substances, and do not react at all to sugars.

3. CJertain ])arts of the general body surface are more sensitive

29^^ Jojtriial of Comparative Neurology and Psychology.

than is the mouth to salts and alkalis. The outer skin and the mouth are equally sensitive to acids, while the mouth is more sensitive to bitter substances.

4. The most sensitive portions of the body ai-e the mouth, nostrils, anus and iins, while the head is the least sensitive to chemical stimuli.

5. When the spinal cord is destroyed, no reactions are obtained from the caudal part of the body, showing that the lateral line nerves have nothing to do with chemical sensation.

G. When the -cord is severed from the brain, the caudal ])art of the animal is more sensitive to chemical stimuli than before. There is no spinal shock.

Y. Section of the olfactory crura and different rami of the trigeminus nerve show that the extreme sensitiveness of the nostrils to the stimuli used is due to the ramus maxillaris trigemini, a nerve of general sensation, rather than to the olfactory nerve.

8. Parts of the body fatigued for tactile response always react to chemical stimuli, but when any given region is fatigued for a given chemical it rarely responds to tactile stimuli, although it usually reacts to other kinds of chemical stimuli.

9. When cocaine is applied to the skin, tactile response disappears before chemical. Among the different chemical sense qualities, the sensitiveness to bitter disappears first.

10. This sensitiveness to chemical stimuli is due almost exclusively to the nerves of general sensation, not at all to the olfactory and very little to the gustatory nerves.

11. Evidence presented suggests that this sense is a true sense quality, with a nervous mechanism distinct from that serving general tactile sensation.

12. A true chemical sense is found not only in the invertebrates, but also in all vertebrate groups from the lancelet to man.

Hull Laborator\' of Anatomy.

The ITiiiversity of Chicago,

Feb. 4, 1909.

Sheldon, Reactions to Chemical St/niuli. 299

LITERATURE CITED.

All papers cited have been consulted excepting those marked with an asterisk (*). The subject matter of articles so indicated was ascertained through reviews.

Adducco and Mosso. ISSG. Richercbe sopra lo tisiol del gusto, (.lioni dcUa R. Accad. de Med., No. 1-2.

AiciiEL. Otto.

181)5. Kurze MittlieiUuig iibei' den histologischen Ban der Riechschleimliaut enibryoualer Teleostier. Sitzungshcr. d. Gcscllsch. /. Moit)hoI. It. Physiol in Miiiichen, Bd. 11, lift. 2 u. 3, 19 Nov.,

S. 73-78.

Ankep, B. von.

1880. Ueber die physiologische Wirkung des Cocain. Pfliir/cr's Arch. f. Physiol., Bd. 21, S. 38-77.

1903. I'eber die Entwicklung der locomotorischen Coordiuationsthatigkeit ini Riickenmarke des Frosches. Pfliiger's Arch. /. Physiol., Bd. 93, S. 134-1G2.

190.J. Teller die Beziehung des centraleu Nervensystems zu den Gestaltungsvorgaiigen der Metamorphose des Frosches. Pfliigcr's Arch. f. Physiol.. Bd. 109, S. 78-82.

1!K»7. Ueber die Shockwirkungen nach den Durchtrennungen des Zentrulnervensystems und ihre Beziehung zur outogenetischen Entwicklung. Ccntbl. f. Physiol, Bd. 21, Nr. 1, S. 9-11.

Barker, Lewellys F.

1896. A case of circumscribed unilateral, and elective sensory paralysis. Jour. E.rp. Med., vol. 1, pp. 348-300.

Bastion, H. C. ^

1890. On the symptomatology of total transverse lesions of the spinal cord, with special reference to the condition of the various reflexes. Brit. Med. Jour., vol. 1, Mar. 1, pp. 480-481.

Bateson, W.

1890a. The sense organs and perceptions of fishes, with remarks on the

supply of bait. Jour. Marine Biol. Assoc, London, N. S.,

vol. 1, pp. 22.5-256, pi. 20. 1890b. The sense of touch in the rockling. Ibid., p. 214.

Bawden, H. H.

1901. A bibliography of the literature on the organ and sense of smell. Jour. Corny. Nenr., vol. 2, pp. i-xi.

Bell, James C.

1906. The reactions of crayfish to chemical stimuli. Jour, of Comp. Neurol, and Psychol, No. 4, vol. 16, pp. 299-326.

300 'Journal of Comparative Neurology and Psychology.

Berth OLD, B. 18SS. Rev. Cciitralh. f. d. iiicd. ll'/.s-.s., S. 400-461.

Bethe, Albeecht.

1899. Die Locomotion des Haifiscbes (Scyllimn) niul ibre Beziebungen zii den einzelnen Geliirntbeilen nnd zuni Labyrintb. I'fliigcr's Arch. f. I'hi/sioL, Bd. 7(5, S. 470-493. Text Fig. 2.

1897. Beitnige zur BiicUeuniarkspbysioIogie des Aales. I'fliKjci's Arch, f. I'lnj.sioL. Bd. (iS. S. 110-119.

Blix, Magnus.

1884. Experinientelle Beitriige znr Bfisung der Frage iiber die speci tiscbe Energie der Hantnerven. ZcitscJir. f. Biol., Bd. 20, S. 141-156. Taf. 1.

1885. Exi)erinientelle Beitriige znr Losung der Frage iiber die specifisclie

Energie der Hantnerven. IbuL, Bd. 21. S. 145-160. Taf. 2.

Brauer, L.

1899. Die Beziebnngen der Biickenniarks(pu'rl;isi()n zu den nnterlialb

derselben sidi abwickehiden Selmeureflexen. Miiiicli. Mcd^

Wochciisclir.. Jalirg. 46. lift. 1. Nr. 2. 10 Jan.. S. 41-42.

Brunn. a. von.

1892. Beitriige znr niikroskopisdien Anatoniie der nienscblidien Nasen bi'Uile. ScJniUzc'.s Arch. f. mikr. Aiiat., Bd. 39. S. 632-6.j2. Taf. 29-30, Fig. 10-11.

Bruns. L.

1893. I'eber einen Fall totaler tranniatiselier ZerstTirung des liiidven niarkes an der Grenze zwisdien Hals- nnd DorsaJniark. Ein Beitrag znr Frage voiu Verlialten der Liilininng nnd der lleflexe speciell der I'atellarretiexe l»ei liocbsitzenden totalen Qnersdinittsliisionen des liiickenniarkes. Arch. f. Psychiatric, Bd. 25, 8. 759-830, Taf. 15-16. 1896. Klinisdie nnd patbologisdianatoniisdie Beitriige znr Cbirnrgie der Riidienniarkstnnioren. Ibid.. Bd. 28. S. 96-168, Taf. 4-5.

Camerer, W 1870. I'eber die Abiiiingigkeit des (iesdiniaekssinnes von iler gereizten Stelle der MnntUir.ble. Zcitschr. f. Biol.. Bd. 6, S. 440-452. Corin, Joseph.

1888. Action des acides snr le gont. Arch, dc bioJ., T. 8, pp. 121138. Danilewski. B.

1892. Znr I'liysiologie des Centralnervensystenis von Aniplnoxus. Pfliiv^i's Arch. f. Phiisiol.. Bd. .52, S. 393-400.

DlSSE, J. 1894. I'eber Epitlielknos])en in der Regio olfactoria der Siinger. Xach richt. der k. Gcscllsch. d. Wiss. zu Gottingcii.

Sheldon, Reactiojis to Chemical Stitmdi. 301

Fano, G.

1!)();{. Colli riliul ion a IT'tiulo di's rrllext's siiinaux. .1/(7/. Hah tic liloL, Tom. oO. pp. S;"i-12S, 1 pi. Fkrraki. p.

1004. Coiiuneiit se niodifie la seiisibilite gnstative pour les tres

petite doses des anesthesiqiies locaux. Arch. Itul. dc Biol.,

Tom. 42, i)p. 411-420. FoNTANE, Arthur.

l*i)02. Feber die Wirkuiig des P]iieain B auf die Gescbmacksorgaiie.

Zclt.sclir. /. rsi/clHtl. II. I'lii/siol. (1. Siini.. Bd. 2.S, S. 253-200. Frey, M. von.

lNi)4. lU'itriige zur Pbysiologie des Scbmerzsimis. Bcr. iihcr (lie If/ liuiidl. dcr ley]. >S'ttc7/.s. Gcsellscli. der Wisscrisch. ,ch Lcip-hj.

Muth. i)liusii<cltc Chis.><c, Bd. 40, S. 185-100. 1S04. BeitrUge zur Pbysiologie des Scbmerzsiniis. Zweite Mittli. Ibid.,

S. 283-200. 1805. Beitviige zur Siimespbysiologie der Ilaut. Dritte Mittb. Ihid.,

Bd. 47, S. 100-184.

GOLUSCHEIUER, ALFRED.

1885. Ueber Wiirme-. Kiilte-. uiid Druckpuukte. Aivli. f. Aiiut. u. PhiixioJ.. I'hii.siol. Ahtli.. S. 340-345. Gkassi, Battista and Castronovo, a.

1880. Beitriige ziir Keimtiiis des (ieriulisorgaus des Hnndes. t^clniltzc's Arch. f. iiiikr. Anal.. Bd. 34, h^. 385-300. Taf. 21. Grutzner, I*.

1804. I'eber die cbemiscbe Keizniig seiisibler Nerveii. J'fliif/cr'i< Arch. J. I'hi/.sioL. Bd. .58, S. 00-104, Taf. 1.

Hanig. D. 1'.

1001. Zur I'sycbopbysik des Gescbmaeksiuues. WiiudV.s I'hiloi^. Sliidicii, Bd. 17. S. 570-023. 7 Fig. im Text. IIaycraft. John B.

1887. Tbe uature of Ibe olgeetive cause of sensation. I't. 2, taste.

Brain, vol 10. pp. 14.5-1('3. 1000a. Tbe sense of taste. HcJiiifcr's Tcrl-hook of I'hii.siol. . vol. 2, pp.

1237-1245. 1000b. Tbe sense of smell. IJtid.. pp. 1240-12.58. Head, Henry; Rivers, AV. H. K., and 8uerren, .James.

1005. Tbe afferent nervous system from a new aspect. Brain, vol. 28,

pp. 00-115, pi. 1. Head, Henry, and 8herren. .Tames.

1!MJ5. Tbe consequences of injur.v to tbe peripberal nerves in man. U>id., pp. 110-338, Figs. 44. Head, Henry, and Thompson, Theodore.

1000. Tbe grouping of afferent impulses witbin tbe spinal cord. IJtid.. vol. 20. pp. .537-742, figs. 83, ]»ls. 1-3.

302 "Joiinial of Comparative Neurology and Psychology.

Heklitzka, a.

1907. Ueber deu "metalliscben Gescbiuack" iind den Geschmack der

metalliscbeu loneu. VII Interuat. Kougress zu Heidelberg. Eev. Zcntralhl. /. I'lnjs., Bd. 21, S, 504-505. 190S. Sill "sapore metallico," snlla sensazioiie astiingeiite e snl sapore dei sail. Arch, di FisioL, T. 4, Fasc. 2, pp. 217.

Hekkick, C. Judson.

1S99. Tbe cranial and first spinal nerves of Menidia ; a contribution

upon tbe nerve components of the bony fishes. Jour. Comp.

Neur., vol. 9, nos. 3 and 4, pp. 153-455, pis. 14-20. 1901. The cranial nerves and cutaneous sense organs of the North

American siluroid fishes. Ihid., vol. 11, no. 3, pp. 177-249,

pis. 14-17. I9(»2. The organ and sense of taste in fishes. Bull, of the U. S. Fish

Commission for 1902; also Bull, of the Scientific Lab. of

Denison Univ., vol. 12, 1903. article 6, pp. 39-96. 1903a. The doctrine of nerve components and some of its applications. Jour, of Comp. Nevr., vol. 13, no. 4, pp. 301-312. 19031). On the morphological and physiological classification of the

cutaneous sense organs of fishes. Am. Kat.. vol. 37, no. 4.37,

pp. 313-318. 19()3c; On the phylogeiiy and morphological position of the terminal

buds of fishes. Jour. Comp. Ncur., vol. 13, no. 2. pp. 121-1.38.

1908. On the phylogenetic differentiation of the organs of smell and

taste. Jour. Comp. ISicurol. and PkiicIioI., vol. 18, no. 2, pp. 157-166.

IIoiiEU. KUDOLF, AND KlESOW, F.

1898. I'eber den Geschmack von Salzen und Laugen. Zcitschr. f. Plnjsilc. Chcmic, Bd. 27, S. OOl-OKJ.

Hooper, David.

1887. Examination of the leaves of the Gynniema sylvestre. Nature, vol. 35, pp. 5(j5-567.

Howell. W. H.

1905. A text-book of physiology. Phihidelpliia. W. B. Sauudcrs and Co..

1905.

Jagodowski, K. r.

1901. Zur Frage nach der Endigung des Geruchsnerven bei den Knochenfischen. Anat. An:;., Bd. 19, S. 257-267, 10 Ab.

Jennings, H. S.

1904. Contributions to the study of the behavior of lower organisms. Pub. 16, Carnegie Inst, of Washington.

1906. Behavior of lower organisms. 14 -+- 2 -f 366 pp., Neiv York,

1906.

Sheldon, Reactions to CUicmical Si/niiili. 303

Johnston. J. B.

1902. All attempt to define tlie primitive fuuctioiuil divisions of the

central nervous system. Jour. Coinp. Neurol., vol. 12, pp.

87-10(>, pis. 1-4. 1!)0C>. The nervous system of vertelirales. I'liihidclpliia. V,M>.

Kahlenuekg, L.

18'JS. The actions of solutions on the sense of taste. Bull. Univ. of Wise, Science series, vol. 2, no. 1, pp. 1-31.

rJOOa. The relation of the taste of add salts to their degree of dissociation, -four, riit/s. Cliciii.. vol. 4. pp. 33-37.

liJOUb. The relation of the (aste of acid salts to their degree of dissociation II. Ibid., i)p. .">33-ri37.

Kastle. J. II.

1S9S. On the taste and allinity of acids. Avi. Vlicin. Jour., vol. 20, pp.

466-471.

KiEsow, Friedrich.

1894a. Ueber die Wirkung des t'ocain und der (iyinneniasaure auf die

Schleimhaut der Zunge und des Mundraums. Wumirs I'liilos.

Stud., Bd. 9, S. .")10-r)27. 1894b. Beitriige 7air physiologischen Psycliologie des Gesehmackssinnes.

IMd., Bd. 10, S. 329-368, 523-561. 1800. Beitriige zur physiologischen Psychologic des Gesehmackssinnes.

IV. Ibid.. Bd. 12, S. 464-473. 1898. Experiences gustatives sur diverses papilles isolement ex citees. Arcti. Ital. dc Biol.. T. 30. pi>. 399-425. 1903a. Ein Beitrag zur Frage iiach den Iteaktionszeiten der Geschmacks empfindungen. Zcitsdir. f. J'si/cliol. u. Plii/siol. der. Sinu..

Bd. 33, S. 453-401. 1903b. Zur Psychophysiologie der Mundlir,li]e uebst Beohachtungen iiber

Funktionen des Tast- und Schmerzapparates und einigen

Bemerkungen uber die wahrscheinlichen Tastorgane der Zun genspitze und des Eii>penrots. Ibid.. S. 424-443. 1. Fig.

KiESOw, F.. AND Haiin, R.

1901. Beohachtungen iiber die Emptindlichkeit dor hinteron Teile des Mundraumes fiir Tast-, Schmerz-, Temi).-. und Geschmacksreize. Ibid., Bd. 26, S. 383-417.

Lemberger. Frieda.

1908. I'sychophyslsche Untersuchungen iiber den Geschmack von Zucker und Saccharin (Saccharose und Kristallose). Pfldgcr's Arcli. f. PlnjKioL, Bd. 123, S. 293-311, 1 text fig.

liENHOSSEK, M. VON.

1892. Die Nervenurspriinge und -Endigungen iin .Tacobson'schen Organ des Kaninchens. Audi. Auz., Bd. 7, Nr. 19 u. 20, S. 628635. Mit 1 Ab.

304 'Journal of Comparative Neurology and Psychology.

LoEB, Jacques.

1891. Ueber Geotruiiismiis Ijei Tliiereii. PpiKjcr's Arch. f. Plnisiol.. lid. lit. S. 175-18U, 1 Holzsch. Lyon. E. P.

1900. Coinpeiisatory motions in fishes. .1;//. Joai: of I'ln/sioL. vol. 4,

no. 2. PI I. 77-82.

Mendelssohn. Moritz.

1S82. Untersuolum.nen iilier Itettexe. N//.:. her. d. ki/l. iirciiss. Akad.

(1. Wis.scnsrU. z. BcrJin. Ilalfbd. 2. S. 897-900. I880. Untersuclinnjien iilter Iteflexe. Zweite Mitth. Ihld.. Ilalflid. 1.

S. 123-1 2.J. 1885. Untersncbnnfien liber Itefle.ve. Dritte Mittli. IbUl.. llaltbd. 1.

8. 107-111. ^looRE. Ben.jamin. and Oertel. IIOUST.

1899. A comparative stndy of reliex action after complete section

of the spinal cord in tlie cervical or npper dorsal reiLjion. Am.

Jour. I'lnjsUih. vol. o. no. 1. pp. 4.3-.)2. MULLER. .7.

1842. Elements of physiolo.uy. Trans, by W. Ikily. vol. 11, 1842. \agel. AVilujald a.

ISDl. A'ergleichend physiolo.t^isclie nnd anatomisdie T'ntersucliunj^en iiber

den (Jeruclis- und (icschmackssinn nnd ilire Organe. mit ein leitenden Betrachtungen ans der allgemeinen vergleichenden

Sinnes])liysiologie. IlihJothrva Zoolofnca. Bd. 7. Hft. 18, S. vii,

207. 7 Taf. 1890. I'eber die Wirknny des Cblorsanren Kali anf den (4eschmackssinn.

ZcitHchr. f. PKiichoJ. 11. /'hi/siol. <J. Sinn.. Bd. 10. S. 23.5-239. l!M»."a. Der Geschmackssinn. Nagel. J/oiKjbiich tier PhiisioJoiiic des

Mcnachcn. Bd. III. Physiologie der Sinne. S. (i21-(;4<>. 3905b. Die Lelire von den spezifisclien Sinnesenergien. Jhid.. S. 1-15. Oehrwall. Hjalmar.

1891. I'ntersnchnngen fiber den (Jesclimackssinn. !<l,<iii<l. Arcli. f. I'll iisiol..

Bd. 2. S. l-(;9. Taf. 1.

1901. Die Modalitiits- nnd (jiialitiltsbegriffe in der Simiespliysiologie nnd

deren Bedeutnng. ]J)iil.. I'.d. 11, S. 245-272.

OvERTON, Ernst. 1897. I'eber die osmotlschen Eigenschaften der Zelle in ihrer Bedentung fiir die Toxikologie und Pharmakologie mit besonderer Beriicksichtigung der Ammoniake und Alkaloide. Zeitschr. f. jtlniKik. Choii.. Bd. 22. S. 189.

Parker. (J. II.

1905. The function of the lateral line organ in (islies. Hull, of flic

U. 8. Bureau of Fisheries, vol. 24, 1904. pj). 1S3-207. 1907. The interrelation of sensory stimulations in Amphioxus. Science.

X. S.. vol. 25. no. n45. ])p. 724-725.

Sheldon, Rcactujus to Clunmcal Stiimdi. 305

Paukek, G. H.

lOOSa. The sense of taste in fishes. IhUJ.. vol. 27. no. (iOO, pp. 453. lOOSb. The sensory reactions of Anipliioxus. Proc. of Aincr. Acad, of Arts and ScioKX'n. vol. 43. no. 10, pp. 413-4.55. 1*E.\RL, Raymond.

1903. The movements and reactions of fresh water IManarians. A study in animal behavior. Quart. -Jour, of Micr. »S'e., N. S., vol. 40, pt. 4, pp. .5U!)-714.

I'flugi:k, W. 1853. Die sensor. Functionen d. Kiickenm. Berlin, 1853. L'lKE. F. I-I.

1908. Spinal shock ; a preliminary note. Science, N. S.. vol. 28. no. 727, pp. 808-809.

'•'liASION Y CaJAL.

1889. Nuevas applicaciones del nietodo de coloration de (Jolgi. Term.

del nervio olfat. en la mucosa nasal. Barcetona. 1893. Xuevo concepto de la histologia de los centros nerviosos. 189;>.

MrcJioic')^ Arcli., 1890.

liEAU, Effie a.

1908. A contrilnition to the knowledge of the olfactory apparatus in dog, cat and man. Am. Jour, of Anut., vol. 8, no. 1, pp. 17-47, pis. 17, Fig. 1.

liETZlUS, GUSTAF.

1892. Kleinere .MKllieilungen von deni Gebiete der Nervenhistologie. III. Znr Kenntniss der Nervenendigungen in der IJiechsdileimhaut. Bi(jl. UiitersKchuiKjeit, Bd. 4, Neue Folge, S. 02-04, Taif. 18.

ItlCIIAUDS, T. W.

1898. 'Jlie relation of llie taste of acids to (heir degree of dissociation.

.i//(. Cliciii. -loin-., vol. 20, pp. 121-12(>. 1900. The relation of the tastes of acids to their degree of dissociation. II. Jour, of riiijs. rliciii.. vol. 4, pp. 207-211. KoLLET. Alexander.

1899. Beitriige zur I'liysiologie des (ieruchs, des (Jeschmacks, der Ilaut sinne und der Sinne im allgemeinen. I'ftii(jer's Arcli. f. Plii/niot., Bd' 74, pp. 3s;!-4(i5, 1 Text-fig.

Rosenthal, J.

187.3. Studien iil)er Refiexe. Moiiatsbcr. der l,(/t. preiis.s. Akad. d. ll'/.s- sciiscJi. zii Berlin. S. lt»4-l(>7. 1884. T'eber Itefiexe. Biol. Ventralhl.. 15d. 4, S. 247-2.50. Rosenthal. J., and Mendelssohn, M.

1897. I'eber die Leitungslialnien der Refiexe im Riickenmark und den Ort der Reflexiibertragung. KciiroL Centralhl, Jahrg. l(i, S. 97S-9S5.

3o6 'Journal of Comparative Neurology and Psychology.

RUBASCHKIN, W.

1903. Teber die Bezieliuugen lies Nerviis trigeaiinus zur Rieehschloiui baut. Atuit. Anz., Bd. 22, S. 407-435, 4 Ah.

.ScHKAUKK, Max E. G.

1887. Zur Physiologie des Froscbgeliinis. (N'orliiutige Mittbeilung. ) Pfliigcr's Arch. f. I'liiisioL. Bd. 41. .S. 7r)-<)0.

Senatok, II.

1898. Zwei Fiille von Querscbniltsei-kriiiikiiiig des Ilalsiiiarkes. Zcitscln: f. kliiiisclic Mcdiciii. r.d. :j.'.. S. l-:r.. Taf. 1-3.

Shelijon. K. Vj.

1908. Tbe iiai-liciiialion of iiii-dullak'd lilicrs in lliu hiiiei-valiou of Ibe olfactory mucous nieuilinine (»f tisbes. t^ciencc, N. .S., vol. 27, no. 702, pp. 91.'5-91(i.

Sheeuen, James.

1908. Injuries of nerves and tlieir trealnient. Atwr Yurie. 1908, pp. 1-130, pis. 20, Figs. 2.").

Sherrington. Charles S.

1897. Experiments in t'xaminalion of tbe iicriplHTal dislril)Ulion of tbe

fibers of I lie i)ost('rior roots of some spinal nerves. I't. 2.

PhU. Trans, of Roij. i<oc. of London. Series B, vol. 190, pp.

4.5-186, pis. 3-0. 1900a. Cutaneous sensations. Schiifci-'s Tcrt-hool: of I'liiisiology. vol. 2.

pp. 920-1001. 1900b. Tbe spinal cord. Ihid.. jip. 78.3-883.

1904. Qualitative difference of si)lnal reflex corresponding wi(b (piali tative difference of cutaneous stinullus. .Jour, of I'lii/sioL,

vol. 30, pp. 39-40. 190.5. Ueber das Zusammenwirken der Biickenmarksreflexe und das Prin zip der gemeinsamen Strecke.- Ergchnissc dcr Phy.^iol., 4 ,S.

797-850, ab. 1 u. 2. 19ona. Tbe integrative action of tbe nervous system. l<lcw York,

1900. 190i)b. Observations on Ibe scratcli reflex in tbe si>inal dog. -Jour, of

Physiol., vol. .34. p]). 1-50, 27 figs.

Sherrington, C. S., and Laslett, E. E.

1903. Observations on some sfiinal reflexes and tbe interconnection of spinal segments. Ibid., vol. 29, pp. 58-96, 45 figs.

■Shore, L. E.

1892. Contribution to our knowledge of taste sensations. Ibid., vol. 13, pp. 191-217.

Steinach, Eugen.

1896. I'eber die electromotoriscben Erscbeiuungeu an Ilautsinnesnerven bei adaquater Beizung. Pfliiger's Arch. f. Physiol., Bd. 03, S. 495-520.

Shfi.don, Rractinns to ChnniraJ Stiuudi. 307

Steiner. J.

1S85. Dit; Funclionen des Centraluerveusysteuis nnd ihre Pliylogenese. Erste Abth. Untersucbnngen iiber die I'liysiologie des Froscliliirns, S. i-vi, 1-143, 32 fig. Braunschircin. 1888. Die Functioneu des Centraliierveiisystenis und ihre riiylogeiiese. Zweite Abtli. Die Fische. S. i-xil, 1-127, 28 fig. lirauiiscli weig. Sternberg, Wilhelm.

1898. Beziebnngen zwiselion dem clieinisclien Baa der siiss scbmeelvenden

Substanzen nnd ilirer Eigensdiaft zn s<'lnneclven. Arch. f. Anat. u. Physiol., Physiol. Ablh.. S. 451-483. 1808. (Jesclmiaclv nnd Cbemismns. Vcrhandl. ilcr pJii/fs. Gcsclhvh. zii Berlin, 9 Dezemljer, 1898. Ibid., 1899, S. 3(J7-:!71.

1899. (Jesebuiadv nnd Cbeniisnuis. Zcitschr. f. /'xi/cliol. 11. I'lnisiol. dcr

Sinn., Bd. 20, S. 385-407. 19<)2. P.eitriige znr Physiologie des siissen Gescbmaclies. Vcrhandl. dcr

Physiol. Gesellsch. cit Berlin, Jabrg. 1902-03, 5 Dez.. 1902,

Arch. f. Anat. u. Physiol, Physiol. Ahth., 1903, S. .538-543. 1902. Gescbmadcseiupfindnng eines Anenceplinlns. Zeitschr. f. Psychol.

u. Physiol, d. Sinn., Bd. 27, S. 77-79.

1902. ITeber das wirlvsanie Princip in den siiss sclnneclvenden Yerbind migen, das dem siissen Geselnnaclv zn Grnnde liegt, das sogenannte dnldgene Princip. Vcrhandl. dcr yhysiol. Gesellsch. zu Berlin, .Tabrg. 1902-03, 24 Oct.. 1902. Arch. f. Anat. 11. rhysiol.. Physiol. Ahth., 1903, S. 190-199.

1903. TT(>ber das siissende Princip. Ihid., S. 113-119.

19<)4. li-rtiiinlidies mid Tatsilcblidies ans dei' I'hysiologie des siissen

(Jescbmacl^es. Zcitschr. f. P.sycJml. u. I'lii/siol. d. Sliw., Bd. 38,

S. 259-304.

1'.M)4. Le principe dn gout dans le second groupe des corps sncres. Arch, internat. dc fhiinmu-odyiKiinie ct dc Thcrapie. Tom. 13,

Fasc. 1 et 2.

1904. Znr Pliysiologie des siissen (iescbmacks. Zcitschr. f. Psychol. 11.

Physiol, d. Sinn., Bd. 35, S. 81-131..

1905. Der salzige Gesclimaclv nnd der Gescbniacli der Salze. Arch. f. Anat. t(. Physiol, Physiol Ahth., 5/G, S. 483. 19f)6. Geselnnaclv nnd Gerucb. Pbysiologiscbe Untersncbnngen iiber den

Gescbniadvssinn. Berlin, 190(5. S. i-viii, 1-149, 5 fig.

S'i'UONG. Oliver S.

19fl3. Tlie cranial nerves of Sipialns acanthias. Science, N. S., vol. 17, ♦ no. 424, pp. 254-2.55.

Thxtnberg, Torsten.

1905. Pliysiologie der Druck-, Temperatm'- uiid Scbnierzempflndnngen.

Nagel. Handhiich dcr J'liysioloyic des Mcnschen, Bd. 3, I'liysiologie der Siniie, S. 047-7.33.

3o8 'Journal of Com par alive Neurology ajicl Psychology.

Van Gehuchten, A.

181X>. Contributions a I'etude de la nuiqutnisc olfactivc i-lu>z les nianiniiferes. La Cclliilc. A. (!, Fasf. 2, ]ii>. •.VX^-M)'.). Fi.trs. 1-14 (1 pi). Vinci, Gaetano.

1807. TToi)er das Eueain R. VircJioir's Arch. f. pfitJi. Aunt. ii. Plii/fiiol.

11. f. l-liii. M ('(].. P.d. 1-10. Folse 14, P.d. 0, S. 217-2.^1. Taf. S. l.SOO. Snr rKucaine P.. ( henzoTl-trans-vinildi.-icr'lonallvaininc ). Aicli. itnl tie hioj.. Tom. :>1. jip. .'^'>2-.'>(>.

^'INTSCIIOAI'. M. VON.

lS71>a. Peitriisf' zuv Physiolofrio dcs (ieschniafkssinnos. I Thcil. P/Jiif/rr's

ArcJi. f. nnjxiol.. P,d. 10. S. 2.^>(!-2.";.^. 1S701). Pciti'iif^t' znr Pliysi()l()j,'i(' dcs Ocscliin.-ickssinnes. II Theil.

Klcktrisclic Kci/am^i; dcr Ziiii^c. Urn}.. I5d. 2(). S. Sl-114. Walton, (J. \j.

1!M)2. The localization of the vertex nieclianism. .hnir. of Xcrroiis (iiid

Mental 7)/.STf/.sr. vol. 20. ])i). ."',.'^,7-;14.'k Walton, (i. L.. and Pail. W. E.

I'.Mk;. 4'lie cereltral I'lenient in the reflexes, and its relation to tlie si)inal

element. Iltid.. vol. ?>H, pp. (!Sl-(i01.

ZWAARnEMAKER, II.

ISO."). Die Physioloirie des (iei-ndis. Lriiizif/. ISO.". S. i-vi. l-n24, Fis. 1-2S.

liH).",. (Jesclnnack. Erticliiiissr d. Plnishilon'ic. Zweiter .T.-ilu'^'., IF .Vlitli., S. »iOO-72r..

Fig. 1. — Mustelus canis, the smooth dogfish; lateral aspect. X l/o- The numbers indicate the resi<nis stimulated, as ,s;;iven in the tables.

Fig. 2. — Same individual, ventral asjiect. X 1/3. It will be noted that this is an innnature male and that the clasiiers are. therefore, oidy i)artially develo]ted. Thus, in the adult male, rei^ion 41 is caudad of the tips of the pelvic fins, instead of between them as shown in the figure.

Shfldon, Reactions to Chcinical Stimuli. 309

310 "Journal of Conipnj'ntivr Neurology and PsychoJoe^\.

Fig. 3. — Dissection of the roof of the mouth of an adult dogfish. X 1%. On the left of the figure are shown the incisions necessary for tlie transection of the olfactory crura and the niaxillaris-niandibularis V trunk in the living fish, while on the right are seen the relations of the various nerves and blood vessels of tlie roof of the mouth, as revealed by the dissection of a dead specimen.

1. Incision through the cartilage for section of the olfactory crura (C) ; 2, incision made in cutting the maxillaris-mandibulai'is trunk (3) ; 3, maxillaris-mandibularis Y. trunk, includes also the buccalis VII; 4, external carotid artery which should be avoided in cutting the nerve; 5, everted nuicous membrane of the roof of the mouth ; A. R., anterior rectus muscle ; B.. olfactory bulb ; hue. VII. ramus buccalis facialis ; C. olfactory crus ; V. VII., ramus from the trigeniino-facial complex to the truncus hyomandibularis. probably general cutaneous; cxt. car. it., external carotid artery; G., Gasserian ganglion; g., geniculate ganglion; 11., cerebral hemisphere; hyoid. a., hyoidean artery; hyomand. VII, truncus hyomandibularis facialis; int. car. a., internal carotid artery; /. O., inferior oblique muscle; /. R.. inferior rectus muscle; Maud. ]'., ramus mandibularis trigemini ; Max. V., ranms maxillaris trigemini; Olf. cap., olfactory capsule; Optic, optic nerve; pal. VII, ramus palatinus facialis; pretr. VII, ramus pretrematicus facialis; Prof, v., ramus ophthalmicus profundus trigemini.

To perform the operations the mouth is pried open, the mucosa is retiected and the incisions made directly, care'being taken to avoid the carotid arteries, both internal and external, and the hyoidean. It is necessary to reflect the mucosa, otherwise the location of the blood vessels cannot be determined.

ShI'-.i.ijon, Reactions to (Ihcmical Stinmli.

I I

KdtharLafc>iai.n°1

TIJE WORK OF J. VON UEXKUELL 0¥ THE PHYSIOLOGY OF MOVEMENTS AND BEHAVIOR.

11. S. JENNINUS.

Bir.LiOGUAPiiic List. (In the discussion, vofcrence is made to the anthor's papers by the use of their iinmbcrs in the followinii; list.)

1. Ueber secmuliire Zuckung. Zcitfivhr. f. BioL, vol. 28, pp. 540-549. 1891

2. I'hysiologisclie Untersucluiiigen an Eledone moscliata. Zcitschr. f. Biol.,

vol. 28, pp. 550-50G. 1891.

3. I'bysiologisclie Uutersnchuiigen au Eledone moscliata. II. Die Reflexe

des Annes. Zeitschr. /. BioL, vol. 30, pp. 179-183. 1893.

4. Pliysiologisclie Ilntersucliungen an Eledone moscbata. III. Fortpflanz nngsgescbwlndigkeit der Erregung in den Nerven. Zeitschr. f. Biol., vol. 30, pp. 317-327. 1893.

5. Ueber paradoxe Zncknng. Zeitschr. f. Biol, vol. 30, pp. 184-186. 1893. G. Znr Metbodik der meclianiscben Nervenreiznng. Zcitsclir. f. Biol., vol. 31,

l)p. 148-1G7. 1894.

7. I'bysiologiscbe Untersnclmngen an Eledone inoscbata. IV. Znr Analyse

der Fnnctionen des Centralnervensystenis. Zeitschr. f. Biol., vol. 31, pp. 584-609. 1894.

8. Ueber Erscliiitternng und Entlastnng des Nerven. Zeitschr. f. Biol., vol.

32, pp. 438-445. 1895.

9. Yergleifbend sinnespbysiologiscbe Untersnclmngen. I. Ueber die Nabr ungsanfnabnie des Katzenbais. Zeitschr. f. Biol., vol. 32, pp. 548-56G.

10. ITeber die Fnnction der Poli'scben Blasen am Kauapparat der regnliiren

Seeigel. Mitthcil. d. Zool. Stat. Neapcl, vol. 12, pp. 463-476. 1896.

11. Znr Mnskel- und Nervenpbysiologie von Sipnncnlus nndus. Zcitschr. f.

Biol, vol. 33, pp. 1-27. 1896.

12. Ueber Reflexe bei den Seeigeln. Zcitschr. f. Biol, vol. 34, pp. 298-318.

1897.

13. Vergleicbend sinnespbysiologiscbe Untersnclmngen. II. Der Scbatten als

Reiz fiir Centrostepbanns longispinns. Zcitschr. f. Biol, vol. 34, pp. 319-339. 1897.

14. Ueber die Bedingnngen fiir das Eintreteu der secundJlren Zncknng.

Zeitschr. f. Biol, vol. 35, pp. ls.3-191. 1897.

15. Entgegnnng auf den Augriff des Herrn Prof. Hubert Lndwig. Zool. Anz.,

vol. 20, pp. 36-38. 1897.

The .ToriiXAL of Co:\irAT;.VTivE NErnoLOGv Axn rsvcnoLoov. — Vol. XTX, No. 3.

314 'Journal of Coniparative Neurology and Psychology.

10. Die riiysiolosie der Pedicel la rien. Zcitxcln: f. li'iol., vol. :'>T. \\\\. ol.'4-lu;!. ISOO.

17. Dei- Neui-oldiiet. Zcitxilir. /. liiol.. vol. :!S, ].]). 2'.>1-l'!)'.». lS!t<.».

18. (Beer. Th., Betiie. A., uiul v. rEXKi'rx. J.) ^'()l•sclll!ige zn eiiiev ohjelv tivierendeii Nonienlvlatnr in der I'liysiologie des Nervensystenis. Jiiol. Cciitralhl., vol. 19, pp. r>17-.")21. ]S!>1). Also in Cnitidlhl. f. I'hiisioL. vol. 13. pp. 137-141. 18110. 1!). Die Pliysiolo.uie des St^fcielstacliels. Zcitsclir. f. Biol., vol. .'>!), jip. 73-112. 1000.

20. Die Wirkun.y; von IJclit nnd Sclnitten auf die Seeigel. Zc'iischr. f. Bio].,

vol. 40. pp. 447-47l>. 1000.

21. Ueber die Erriclitnng eines zoologischeu Arlieitsplatzes in Dar es Salaam.

Zool. Anz.. vol. 23. pp. 570-583. 1000.

22. T'elier die Stellnng der vei'gleiohenden Physiologie zur Ilypotliese der

Tierseele. Biol. Ccntralhl.. vol. 2(». pp. 407-502. 1000.

23. Die Scinvininilieweginigen von lUiizostoma pnlnio. Mittluil. a. d. Zool.

Stat. XcdiKl. vol. 14. pp. (,2(>-(i2(). 1001.

24. Psycliologie nnd P>iologie in ihrer Stellnng znr Tierseele. Aslier und

Spiro's En/chiiis.^c der J'lijisiolonic. 1 Jalirg.. 2 Abtli.. pp. 212-238. (Rel>rinted witli tlie title "Im Kanipf nni die Tierseele.")

25. Stndien iilier den Tonus. I. Die biologische Bauplan von Sipnnculus

nndns. Zritxchr. f. Biol., vol. 44. ])p. 2(iO-344. 1003. 2(;. Die ersten T'rsacben des Illiythnius in der Tierreilie. Asher nnd Spiro's l-Jri/chiiissc ilcr J'ln/siologic. 3 Jalirg.. 2 Abtli.. 11 pages. 1004.

27. Stndien iilier den Tonns. II. Die I'.ewegnngen der Sclilangensterne.

Zcitscin: f. Biol., vol. 40. pp. 1-37. 1004.

28. Stndien iiber den Tonns. III. Die P.lntegel. Zcitscin: f. Biol., vol. 40,

pp. 372-402. 1005. 20. T.eitfaden in das Stndimn der experinientellen Biologie der Wassertiere. 1.30 lip. Wiesbaden, 100.5.

30. Stndien iiber den Tonus. IV. Die Herzigel. Zcitscin: f. Biol., vol. 40. pp.

307-332. 1007.

31. Stndien iiber den Tonus. V. Die I.ibellen. Zcitscin: f. Biol., vol. 50, pp.

108-202.

32. Die A'erdiclitnng der Muskeln. Zciitnilhl. f. I'lnisiol.. vol. 22. ]»p. :>4-37.

1008.

33. Die neuen Fragen in der experinientellen Biologie. Riristti di >^cieiKa

"Scientia." vol. 4. No. 7. 17 pp.

So:\[E CllARACTEKISTIC DiCTA.

"We may qtiietlv throw overboard the entire sum of our knowledge thus far acquired, so far as it comes from experiments on the frog's leg. The nerve-muscle prei)aration has misled us in almost every point" (30, p. 331).

Jennings, Uexki'dl on Physiology of Behavior. 315

"Furtluu'iiiorc all ol('c'tro-})hysi()]()iiic;il ('xpcrimciits 011 muscles have shown ihcnisclvcs to bo bioloiiiciiUy woi'tlilcss, so that we may here pass them o\('r in silence"' {'1\K \>. i^lS).

The fate of hioloiiv in Euro})e, in s])it(' of the eifoi'ts of excellent workers, seems to me to he sealed. """ * "' One need not he a jirophet to ])i'edi('t that in a few years hioloiiy will he an American science (21), preface).

The question of the function of tho ucrNous system in the aninud body has aroused a strife between two sciences that must end with the annihilation of one of the two tMuubatants, — and the champions of both sides are determined to carry the cond)at to the end. "" "' * While the comparative psychologists debated concerninsi' the amount of sensation, memory, reiiection, that one should attribute to these animals, there arose in the growing science of comparative physiology an enemy to the death of all comparative psychology" (24, pp. 212, 213).

Before objective investigation the sensations, the memory, and thoughts of animals disappeared like Huttering forms of vapor. The iron chain of objective changes, which began with the stimulation of the sense organ and finished with the movement of the muscle, was welded together in the middle. Xowhere remained a smallest spot for the psyche of the aninuil. Dasing itself on these incontestable facts, comparative ])hysiologv pronounced the ])sychological conclusions mere sujicrstitions and denied coui])aral i\e psyehology the right to call itself a science" (24, ]). 21.')).

"We stand on the eve of a scientific bankiMiptcy, whose consequences are as yet incalculable. Darwinism is I0 be stricken from the list of scienfitie theories" (.'].'], p. 3).

Concerning the oi-igin of species we know, after fifty years of unparalleled effoi-t aii<l inxcstigation, only the one thing, that it does not take place as Dai'win ll'o:ight it did. A josilive enricdnnenf of our knowledge has not resulted. The wh<ile enormous intellectual labor Avas in vain" (.')1, ]). KiS).

"When in biology one has freed himself from the idea of development, — an idea which has at last been hunted to liiial death, — so that one is again in position to look u])on eacdi aninud as a unity

316 'Journal of Comparative Neurology and Psychology.

closed within itself, instead of as the last eliancc product of an ancestral series that has Ix-en speculated together, then form and color gain a new interest and a heightened brilliance" (30, p. 318).

"Driesch succeeded in proving that the germ cell does not possess a trace of machine-like structure, but consists of throughout equivalent parts. With that fell the dogma that the organism is only a machine. Even if life, occurs in the fully organized creature in a machinedike way, the organization of a structureless germ into a complicated structure is a power sui generis, which is found only in living things and stands without analogy" (33, p. 9).

It is not to be denied that the vitalists are the victors all along the line. After having put an end to Darwinism, they have seized u})on the entire field of the production of animal form, and now threaten the last positions of their opponents" (33, p. 14).

So there persists in the outer world of objects an unresolvable contradictoriness— " (29, p. 129).

TxTnODUCTORY CirARACTKRTZATTO:V.

J. von Uexkull, of Heidelberg, has long been one of the most active and original workers in the physiology of behavior ; his work will be found of the highest interest to all seriously concerned with these matters. It undertakes for the lower animals much the same sort of analysis that Sherrington gives us for higher ones, though with many features that are in the highest degree original. Further, an examination of his work has hardly less interest as a study in scientific ideals and method than for the concrete results attained. That he has been led to radical and iconoclastic views, and that he is not afraid to express these views with decision and picturesqueness will be evident from the quotations given above. The pessimistic impression given by the quotations taken together will not escape the reader ; this pessimism appears rather an unintentional result than as characteristic of the animated and niib'tant si)irit of our author. The sweeping and ]ierliaps ill-found<Ml character of some of the propositions advanced should not prejudice the reader against the accuracy of the author's work in his own field; this would be a serious mistake.

Jknnings, UexkuU on P/tysiology of Behavior. 317

Investigations. Von llcxkiill began as a student of the nerve-muscle preparation of the frog (1). He quickly determined to carry the study of the questions involved to the lower animals ; this plan, carried out largely at the ISTaples Station, led to fundamental results. The work first undertaken was a study of the reflexes of the cuttle-fish (3, 3, 4, 7). This was followed by a study of certain sensory problems on the skate (9), by work on the muscle and nerve physiology of the worm Sipunculus (11, 25), and by an extensive series of thorough and fundamental papers on the nerve-muscle and sensory physiology of the sea urchin (10, 12, 13, 15, 16, 19, 20). Through these studies the author had developed the outlines of an original system of nervemuscle physiology, having at its basis the concept of tonus. This system he developed farther in the series of Studies on Tonus (25, 27, 28, 30, 31), — a series dealing with various invertebrates, which is still in progress, and which we hope may count many numbers besides the five that have appeared. Arising in connection with his systematic series of investigations there have come from his pen many incidental contributions, including notes on special points in nerve-muscle physiology (1, 5, 6, 8, 14, lY, 32), studies of rhythm (23, 26), and discussions of fundamental scientific questions (18, 22, 24, 33). In 1905 a brief textbook (29) was published, giving an outline of the views to which he had come, with directions for practical work.

Characteristic of v. Uexkiill is the intellectual working over of results at the time they are reached, so as to give a graphic and comprehensible scheme of the way processes occur. This appears in the very first studies, on the cuttle-fish (2, 3, 4, 7) ; they show the author's characteristic abhorrence of everything vague, and particularly his objection to psychic explanations in physiology. They contain much important detail on the physiology of muscle and nerve. The paper on the skate (9) is largely a polemic against the use of terms implying consciousness in the nerve physiology of lower animals, directed mainly against W. Wagel. The account of the skate is intended to illustrate the purely objective treatment of the facts, and is perhaps not in ifself a sti'ong example of the value of

318 'Journal of Comparative Neurology and Psychology.

the method. In this polemic, as in later work, is seen the positive, definite character of the anthor's thonght, with no appreciation for shadings, transitions or compromises; sharpness of concept and of distinctions, black or white, is demanded everywhere.

The Work ojs^ the Sea Ukcjlin.

It is in the series of papers on the sea urchin (10, 1:^, Vo, 10, 19, •20) that the author "strikes his pace" and l)rings out clearly the important phenomena which form the basis for his peculiar system of concepts. The plan of investigation is to work out the structure of Ihe muscular, nervous and skeletal systems and the way they act together to produce the characteristic behavior of the animal ; to work out "the biological plan" of the sea urchin. The paper on Reflexes in the Sea Urchin (12) gives a general sketch of this biological plan and the main reflexes which make up the behavior ; the structure of the simi)le nervous system, the reflexes of the pedicellarite, the spines, the tube-feet, the teeth, are briefly described. Then followed in 1899 a thorough detailed paper on the physiology of the pedicellarise (10), a paper which must form the point of departure for all further work on these organs. The history of our knowledge of the pedicellari*, their structure, the ditt'erent sorts, their uses and particularly the laws of their movements, are developed in great detail. In 1900 came a similarly thorough, and perhaps still more important, study of the movements of the spines (19). Special studies were made of the reactions of the spines to light and shadow (1-3, 20) ; for the purpose of obtaining species favorable for this work the author made a tri]) to the cast coast of Africa; the outtit for this trip is described in a separate paper (21).

In general, v. Uexkiill found that the various organs of the sea urchin, — each spine, each pedicellaria, — can perform a niiiii1)er of different acts or reflexes, and that these reflexes occur to a large degree independently of each other, so that they may occnr when the organ in question is isolated on a small piece of the shell. Thus each organ is a "reflex ])erson. Yet all the difl'eriiig reflexes of these various "■pci-soiis" are of such a character that they woi'k together in a sysli'iuatic way to perfoi-m the necessary fiiuctioiis of the animal.

JenninGvS, UcxkilU on Physiology of Behavior. 319

They carry it about, toward food and away from danger, keep the sea iirehin clean, capture prey, combat against enemies, and all together do the work of life in a competent way. This is brought about without any regulation by a "higher center," merely through the essentially in(lc])ciident activities of the various ])arts. Therefore the sea urchin is characterized as a '"republic of reflexes." The ap]>arent unity of action is due merely to the way the different actions of the different parts fit into one another, by a sort of preestablished harmony. "It is not that the action is unified, but that the movements are ordered ; that is, the setting off of the different reflexes is not the result of a common central impulse, but the separate reflex arcs are so constructed and so fitted together that the synchronous but independent setting off of the reflexes, as a result of an external stimulus, produces a definite general action of the animal, just as happens in animals in which a common center })roduces the actions" (16, p. 390). "When a dog runs, the animal moves its feet. When the sea urchin runs, the feet move the animal" (11), p. 7-3).

But the author finds that the separate reflex-])ersons are not absolutely independent ; impulses may pass from one to the other, in such a way as to produce a unified action of all. But this is due merely to a set of nerve nets having a special arrangement; it is a matter of interconnections, not of regulation from a ^'higher center." The question may be raised whether this distinction docs not depend on an undefined and mystical use of the term higher center. If the higher center, in accordance with the illuminating ideas of Loeb, is after all essentially a place for complex interconnections, then the difference between the sea ui'cdiin and higher animals is only one of degree.

Von Uexkiill finds that the separate reflexes of the various organs arc not absolutely stereotyped, but the action of each part is changed at times in certain ways, often depending <>\\ the action of neighboring parts. We have then in the sea urchin an opportunity of studying coordination in its most elementary condition, and thus perhaps of determining the fundamental nature of the phenomena involved. This aninnd is a sort of model, which can be taken to pieces Avitliont serious ultci'ation of the action ni the [)arts; yet in

320 'Journal of Comparative Neurology and Psychology.

the aninial as a whole the parts do influence one another. The author therefore took in hand a thorough study of the changes in the reflexes and the way they influence each other, undertaking to formulate and define precisely the underlying phenomena. It is here that we find the origin of the most important concepts in von ITexklill's highly original formulation of behavior. It will be worth while therefore to examine carefully a sample of the author's method of analysis; for this purpose we select certain features in the physiology of the spines.

Ty1'1CAT> EXA.MI'LK OF MeTUOD OF ANALYSIS.

It will be recalled that the rounded shell of the sea urchin is covered with long spines. Each spine is a tapering calcareous rod, with a concavity at its base, by which it articulates with a hemisi)herical elevation of the shell. The spine is held in position by two circles of muscles radiating from the circumference of its base to the shell. These muscles, and particularly the inner circle, are steadily pulling upon the spine, thus holding it stiffly in position. This pulling takes place without external stimulus ; it is due to a certain amount of tension which forms the normal condition of the muscles and continues without any such repeated contractions or tremors as are called tetanus. That is, the muscles have a certain normal tunus. This tonus becomes the central concept in v. Uexklill's formulation of the physiology of movement. The iinier layer of muscles is devoted chiefly to maintaining by its constant tension a certain position of the spine ; it is an example of one of the two great tyjies of muscular action, — the Sperrung" or tension, as distinguished from actual contraction, involving shortening. The outer circle of muscles is more active in its changes; they shorten quickly and readily, thus moving the spine in various ways. They exemplify tlie other great type of imiscidar iiction, — ^'Vcrl-iirziinrj," — shortening or contraction. Tension and contraction v. T^exkiill shows occur quite independently of each other, ar.d this independence, with all its theoretical and practical consequences, conu^s to play a very great part in the later development of v. ITexkiill's views; li(^ finds it throughout animals (see 33). The neglect of the fact that we have

Jknnings, Uexkiill on Physiology of Behavior. 321

here two eiitirely (lili'crcnt fuiK'tioiis has, the author believes, led all nerve-muscle physiology into false paths.

The first reflex shown by the spines is as follows: When a certain spot on the body is moderately stimulated, the surrounding spines bend toward it. The muscles of the side of the spine next the point stimulated contract; that is, their tonus is increased. Thus the points of the spines are directed, for example, toward an approaching enemy.

But if the stimulus is very intense, the reaction just described is reversed ; the spines bend away from the point stimulated. This result is ])roduced by a decrease in the tcjnus of the muscles on the side of the spine facing the j^oint stimulated. Thus from the same spot on the body opposite effects may be produced, depending on the strength of the stimulus. This phenomenon is called by v. Uexkiill reversal of the reflex ("Eeflexumkehr") ; it is observed in other organs of the sea urchin and other animals, nnder various conditions. The author holds it to be due to some sort of ap})aratus in the ganglion cells; an apparatus that he calls the "tonus switch" ("Tonusschalter"). This reversal is well seen in the spines when a strong chemical stimulus afl'ects the body. Now, a further consequence of such a powerful stimulus is seen. After such a stimulus, even a weak stimulus, which formerly caused the spines to bend toward the spot stimulated, now causes them to bend away. So the same stimulus on the same spot may cause two diflerent reactions, depending on what stimulus has preceded it. This phenomenon, very common in animals, v. Uexkiill calls the "switching" of the tonus ("Schaltung") ; he holds it due to the same apparatus as the reversal of the reflex.

The two reactions of the spines serve, under natural conditions, certain functions. The bending toward a stimulated point serves for defense; the bending away under a strong stimulus, particularly a chemical one, preserves the spines from injury, while giving opportunity for the action of certain large ])oisonous pedicellariffi, which tiow bend their envenomed jaws towai'd tlie region attacked and seize whatever is there present.

Certain other facts in the physiology of the s])ines are of extreme importance. A steady tension, not violent, exercised on the muscles

322 'Journal of Comparative Neurology and Psychology.

of the spines causes them to lose their tonus ; they become limp. This effect of tension on tonus is common among animals. If then a spine is pressed steadily to one side by the fingers, or by the weight of the animal's body, the muscles on the side pressed lose their tonus. The spine, therefore, becomes loosely movable in certain directions, but not in others. On the other hand, a sudden violent increase of tension, or a mechanical jar, increases the tonus, so that the spines stand out firmly.

Now, the loss of tonus, caused in the way just described, is conducted, doubtless by the nervous network, to the neighboring spines. This conduction occurs in such a way that it is the muscles of corresponding sides of the neighl)oring spines that lose their tonus. (This involves complicated conditions in the nervous net ; v. Uexkiill holds that it shows the existence of many independent nets.) Hence when a spine is pressed toward one side, the neighboring spines likewise bend in the same direction. This v. Uexkiill calls the chaining of the reflexes ('^Reflexverkettung"). It shows itself (an important fact) most readily when the spine is bent toward the mouth; the other spines also bend toward the mouth.

These facts have the following result. When a spot on the body is strongly stimulated, so that the spines bend away from it, the disturbance is not limited to those in the innnediatc neighborhood. The S})incs in bending away press upon the surrounding s])ines, tending to bend them down. They are more easily bent toward' the mouth than elsewhere, so a new set of spines bend over in that direction. They again press on the next spines, bending them in tnrn toward the mouth. Thus a sort of wave })asses toward the mouth from the point stimulated, the spines bending in turn far over toward the mouth, then back again. The entire phenomenon v. Uexkiill calls the wandering of the center of excitation (Wandcrnng des Erregungsniitt('l])nnkts ).

Another most important fact shows itself. Muscles that are not in tonus are much more easily stimulated to contraction than those which are in tonus. When the muscles lia\'e their usual strong tonus, it re(piii'es a jiowerfnl stimulus to cause tlieni to contract further. But muscles which have lost their tonus as a result of

Jennings, UcxkiiU on Physiology of Behavior. 323

stt'uJy tension (as (k'scrilicd alinvc), contract readily in response to even a weak stiumlus, tending thns to bend tlie spine toward that side on wliicli there has been tension.

From this a nnndnu- of jieenliar facts resnlt. As we have seen, a moderate stimidns at a certain point tends to cause the spines to bend toward that point. If, as a result of pressure, the muscles that face the point stimulated have lost their tonus, they respond readily; the spine at once bends toward the side stimulated. P>ut if the spines have been pn^ssed over in the opposite direction, so that their muscles facing the point of stinmlation are in strong tonus, no effect is produced; the spines retain their ])osition. Hence, when a spot on the body is stimulated, certain s])ines will respond while others will not, depending on the previous tonus of their muscles. This phenomenon v. Uexkiill calls "Ivlinkung" ; those spines which are in such a condition or position that they can respond to the stimulus are said to be "eingekliukt" ; those which are not are ausgeklinkt. These expressions may perhaps be translated by "in circuit" and "out of circuit," — comparing the spines with instru ments in an electric circuit. This condition of affairs has great importance for the functioning of the s])ines in locomotion and elsewhere, and parallel conditions are found in other organisms.

A similar analysis is given by the author for the pedicellariis, tube-feet, teeth, etc.

Thus by a close and thorough study v. Uexkiill has been able to analyze and formulate a number of what have been called vaguely the varying "physiological states" of organs or organisms; such analysis is needed for all cases. By making use of the concepts of Reflexumkehr, Reflexverkettung, Wanderung des Erregungsmittelpunkts, Schaltung, Klinkung, and by observing the changes in tonus and the rules for its increase and decrease, one can explain some of the most important features in the behavior of the sea urchin under natural conditions; locomotion, negative reactions to various stimuli, defence from enemies, capture of food, etc. It is, of course, no disparagement of the value of this analysis that it does not exhaust the matter for the sea urchin. Thus, when the animal is turned on its back, its spines move in ways that would not be expected

324 'Journal of Comparative Neurology and Psychology.

from the pliysiological analysis based on their other movements (19, p. 105) ; if they did the sea urchin would not regain its normal position. In the starfish the method of action may he changed by the formation of habits, and this is doubtless true also for the sea urchin. Thus any formulation that is complete must provide also for the laws of change of behavior; for its regulatory features. Possibly no complete formulation can ever be reached, but the most direct way to approach it is by such analysis as v. irexkilll gives.

Later Investigations. We have given this account of the spines as a type of v. UexkiilFs methods of analysis ; by following carefully such a concrete case the reader will get a better idea of the nature and justification of his work than by any systematic survey of the concepts to which he finally comes. Let us now follow further the development of these concepts. As we have seen, the central concept is that of tonus, and the laws of the changes of tonus are the chief object of research. To research on this matter, to studying the properties of tonus in various organisms, and to devising schemata which shall help us to understand how it acts, and hence how behavior takes place, have lieen devoted the later researches of v. Uexkiill. He has thus far analyzed from this point of view, besides the sea urchin, the worm Sipunculus (25), the brittle-star (27), the leech (28), the heartshaped sea urchin (30), and the dragon fly (31). In the latest contrilmtion, on the dragon fly, v. Uexkiill attempts to make provision for a modification of the machinery of behavior through the experiences of the organisms. It would manifestly l)e impossible to resume here these researches, filled as they are with minute and technical detail.

V. IlEXKiTLT/s Syste:\i of Concepts. A view of V. Uexkiill's system of concepts can be gotten most directly from his Guide to Experimental Biology" (20). But here one does not see the development of the ideas ; the actual grounds that have given origin one after another to the jjccnliar concepts, so that they are likely to seem on first introduction Inzarre and artificial, having little similarity to anything dealt with in orthodox physiology.

Jfnnjngs, UexhiUl on Physiology of Behavior. 325

Tlu' fundamental concept is tonus. Just what arc we to understand by tliis ? V. Uexkiill at first defines it merely as the sum of those manifestations of the life of the cell that produce effects on external things (as distinguished from the internal energy used in metabolism, etc.) (19, p. 78). As his work develops, he finds need for a more precise idea of tonus. It is defined as a "form of energy-' which has the property of flowing in certain ways (20, p. 4Y4). The concept of tonus gradually becomes more and more definite. For purposes of handling and imaging it with ease, it becomes convenient to think of tonus as a fluid, which flows through a set of tubes (the nerves). This fluid becomes at last identified with Bethe's "Fibrillensiiure," — an actual chemical, visible under the microscope (27, p. 31). V>\\\ this identification is not held to uniformly.

This fluid tonus is contained in a system of tubes, the nerves. "The structure of the nervous system may then be conceived as an aggregate of peculiar vessels united one with another, which interchange and equalize each other's contents with relation both to pressure and quantity" (25, p. 305). From the nerves the tonus either j)asses into the muscles, or causes in them the production of a fluid with similar properties, giving rise to either tension ("Sperrung") or contraction ("Vcrkiirzung). In dealing with tonus, eith(n' in the nerves or the muscles, we must distinguish its quantity from its pressure; these may varj' independently, so that any given quantity may have high or low pressure. On the quanlitij of tonus depends the contraction of muscles; on the pressure, the tension of muscles.

There are certain general laws iv r the movements of tonus. In simj)le uerve nets it always flows into muscles that are extended (causing them to contract again). This is attributed to a change in the capacity of the muscles ; extended muscles have greater capacity than contracted ones, so in extending they suck, as it were, the tonus out of the nerves. This property gives a remarkable degree of self-regulation to the action of the nerves and muscles.

In most animals, further, the tonus shows a marked tendency to flow toward a -certain part of the body, — usually the anterior end, —

326 'Journal of Comparative Neurology and Psychology.

so that this part rospoiids when any })ai't of the Ixxly is stimiihitcih This part to which tonus flows as water flows into a valley is denominated, with poetic feeling, the vale of tonus ("Tonustal," 25, p. 310; 29, -[). 50). After the tonus has flowed into certain muscles, it is possible (in some cases at least) to capture and hold it there, by cutting the ner^'es leading to the muscles (Tonusfang/" 25, p. 302) ; the muscles then remain contracted.

During rest the fluid tonus is gradually used u}) and disa})pcars; at stimulation it is newly manufactured. There exist, however, reservoirs of tonus in the nervous system, so that the lost tonus of the muscles can be replaced without new manufacture.

The nervous system then contains, besides a system of communicating tul)es, reservoirs of tonus; at the same time it is an elaborate apparatus for controlling the distribution of tonus. Each muscle has somewhere in the nervous system an organ which is its "representative" (25, p. 303; 29, p. 44). The ofHce of this representative is to see that the tonus ])ressure in the muscle remains sufficient to cause the tension of the muscle to correspond to the weight which it has to bear. When the pressure in the muscle is insufficient, this acts on the representative (through the nerve) causing it to increase the pressure, until this raises the tension so as to sup])ort the weight. The increased ])ressure is produced by the fact that the representative uses up a certain quantiiy of tonus to increase the pressure of what is left.

The author's further development of this system consists in working out in detail the structure and action of this system of tubes, reservoirs and other nuu'hinery, by which the distribution of tonus is controlled. ]\rain tubes, feeders, reservoirs, valves, etc., are devised and represented by diagrams, till we finally get figures which resemble the plan for a dye-w'orks or a flour mill (see for example the schema for Sipunculus, 25, Tafel 6).

The method of ])resentation is in general the ideal construction of an a]iparatus which could produce the results shown by the organisms. In this construction no attempt is made to represent apparatus that actually exists in the organism ; it is merely a figure or illustration ; "a mere schema in accordance with which one can group

Jf,nninc;s, Uexki'iU on Physiology of Behavior. 32/

the experimentally foimd facts in a convenient way" (25, p. 287). The schema of indirect investigation is not a theory at all, but merely a sign language by means of which it is possible to at once express new results in a graphic ('anschaulich') way" (27, p. 31). All emphasis is laid on making the illustration thus "anschaulich" ; that is, of such a character that one can see through it" ; see how it would work as a machine works. The author makes extensive use of this "fictitious schema" (25, p. 291), basing long discussions for the greater part of entire papers on its properties. Perhaps nowhere else in biology has a figure of speech, as it were, been worked out in such tremendous detail, through a long series of papers.

Regarded thus as a figure or illustration, the author appears very successful in constructing apparatus that would produce results similar in their complication and regulatory character to the processes observed in organisms. This has necessarily been done, of course, by attributing new characteristics to the various components wlien required. The tonus is sometimes given the characters of a definite material fluid, and much pains is taken to account for the entire quantity ; again, it ma_y be produced or disappear as required ; sometimes it is considered a form of energy; at times it shows the properties of electricity in producing effects by induction (31, p. 195) ; at times we are informed that the figure of a fluid quite fails (25, p. 21f3). When the author attempts to show how his complicated machinery may become modified in a way corresponding to the production of what are called psychologically memory images (31), clearness has to be given up, and the entire figure becomes unconvincing.

As to the value of this figurative and artificial method of presenting the results of work, opinions will, of course, differ. The point can be best discussed in connection with a review of the guiding principles and scientific ideals in the author's work ; to this we now turn. It is peculiarly true in the work of v, Uexklill that the author's concrete results cannot be understood without an appreciation of the principles that have guided him. This will lead us to a consideration of his general and theoretical papers.

328 'JourtinJ of Comparative Neurology and Psychology.

Theoretical Views and Gitidixg Principles.

Perhaps the main characteristic shown throughout v. ITexkull's work is the abhorrence of anything vague, ill-defined or mystical. In his early papers he sets forth clearly the ideal of scientific work as . the discovery and presentation of what is verifiable or demonstrable. We have to do only with processes that can be objectivel}' demonstrated, and to write the history of these processes in an animal from the moment of stimulation to the resulting reaction" (9, p. 559). This led him at once into a polemic against authors that used psychic explanations in work on animal behavior (see 7, p. G08 ; 9, etc.). The circle was soon widened, and in 1902 v. Uexkiill declares that a war of extermination has arisen between comparative physiology and comparative psychology, a war that spells annihilation for one of the combatants and "both are determined to carry the fight to the end" (24). His fundamental jwint is, of course, the fact that there is no way of observing or verifying the existence of psychic phenomena in animals, so that they cannot form a part of a strictly verifiable science; and a further postulate is that all objective processes can and should be fully presented and accounted for without bringing in anything from outside. To substitute psychological interpretations for certain steps of objective experimental analysis is vicious and destructive of consistent science. V. ITexkiill's polemic j^apers take extreme positions and are written with much picturesqueness of statement; they are of great value for rousing to a realization of the difficulties those who need such a spur. Apparently, however, all the valuable results that would be reached by utterly destroying the unhappy comparative psychologists would be equally well attained by keeping carefully separate the two fields of work. If the experimenter never substitutes a psychological explanation for a physiological one, he may also be interested, as a separate problem, in the development of mind, without injury to his objective scientific work.

This same demand for objective verifiable results, without admixture of anything else, has led v. Uexkiill to take a i)art, with Beer, Bethe, and others, in trying to establish a purely objective nomenclature for the processes occurring in the movements of animals (18;

Jennings, Uexkilll on Physiology of Behavior. 329

see also 20). This noinciiclature has philosophical value and has ' been used by a few authors, though its employment is by no means common. There is difficulty, as with all ideal system of nomenclature, devised before investigation is complete, in the fact that its use often implies a precise knowledge of the nature of the phenomena, when such knowledge does not exist. To give precisely the correct name to a process inqdies that we know fully the nature of the process.

The author's abhorrence of the vague later jjocomes still more accentuated iu the (IciiiaiKl (which we have noticed above in our account of his investigations) that work shall be presented always iu a way that is "anschaulich" ; that is, in such a way that one can see just how the processes would occur, as one sees how a machine works from knowing its structure. It is extraordinary to what an extent the author makes the attainment of this "Anschaulichkeit" the chief object of biological science; he declares it plumply to be the "most essential character of all" ("die allerwesentlichste") for the science of biology (31, p. 184). "Biology is in its essence 'Anschauung' " (33, p. 10).^ "Only the anscJiaidich structural diagram, not proving, but sli airing the unified working together of different factors, is adequate to the requirement of bringing the life processes together into an intelligible unity without omitting life itself" (31, p. 185). It is only by grasping fully the fact that "Anschaulichkeit" is the author's ideal, that one can understand many of the peculiarities of his work.

The first far-reaching consequence of this ideal arises from the fact that that which is "anschaulich" is not always that which is verifiable. The author is therefore sometimes compelled to a choice between the two, and in his later papers he at such times deliberately chooses the "Anschaulichkeit" in preference to verifiableness. He thus falls into contradiction with his own earlier requirement that we shall deal only with what is objectively demonstrable ("objectiv nachwcisl)ar," !), p. 5,59), as well as with the procedure of other in 'Tlie word intuition, hy wbicli "Aiisrhaiiiinf/" is commonly translated, certainly fails to carry to most minds the same graphic idea as the German word, so that I do not employ it.

330 'Journal of Comparative Neurology and Psychology.

vestigators to whom it is more important that scieiitilic propositions shall be verifiable than that thej shall be anscliaulicli. Let us here look in a general way at the contrast between the results reached by making Anschanliehkeit" the ideal, and those which flow from making verifiablcness the ideal.

For many investigators the object of science is to prepare a system of verifiable propositions, in order that we may know what to depend on in our conduct ; to know what is true in order to do what is right," as Huxley put it. Verifialjle projjositions are propositions that say "Under such and such conditions you will find such and such things to occur or exist."' Now, if one supj)lies the conditions set forth, and does nut find the })rcdicted things to occur or exist, the ])roposition is not verifiable, and many would therefore hold that it should be stricken from science. A large proportion of the propositions concerning machinedike structures in organisms, given by v. IJexkiill, do not even profess to l)c verifiable. One of the main objects of investigation is to find out what particular kinds of machines are i)resent in aninuils, and how these actually present machines have arisen and lioW they are changing. This object is incompatible with the mere assumption of fictitious machines, for the first result of investigation with this object in ^'iew is to cancel these fictitious machines. This would indeed leave our science for a time less shar})ly formulated, but "at a certain stage in the development of a science a degree of vagueness is what best consists with fertility,"^ for reference of the phenon^ena to complete fictitious machines tends to cut oft" search for the real ones. When we have found out what really occurs in organisms and what machines actually exist there, then our knowledge will be as "anschaulich" as the facts warrant, no more, no less.

To the present reviewer it seems that, even for practical purposes, the author has overestimated the value of a rather gross "Anschaulichkeit." . The bringing in of machine-like strncturcs, — tubes, valves, etc., — that confessedly do not exist, seems rather to confuse

"Or, ])ut in a form which holds wliatovcr one's theories, "when yon liave such and such experiences, you will ha\e such and such other experiences." 'Vrm. James, Psychology, I'reface.

ji'.NNiNGS, Uexkiill on Physiology (jj Behavior. 33 1

than to aid the mind. It is nut possible to conclude directly from the properties of the assumed machines as to what physiological pro])erties one will find, for the parallelism is far from complete, so one must try to keep the system of machinery separate from the system of })hysiological facts ; there are two systems to grasp in place of one. The reader finds it difficult if not impossible to disentangle statements which the author wishes to present as verifiable facts, from statements which are a mere necessity for carrying out the figurative schonui. V. Uexkiill shows all through his work an astounding facility in concluding as to the structures that must be present, from the functions which he sees performed. The reader wonders whether these structures are held to actually exist, or whether they are part of the fictitious schema. The reviewer finds that for his own use it becomes necessary in reading v. Uexkilll's work to ask Now, what did the author here actually observe and demonstrate?" It then becomes neces.sary to transform or almost re-write a paper before the verifiable results can be disentangled from the figurative presentation. I believe that this condition of affairs has ])revented the work of v. Uexkiill from exercising the great influence that it deserves from its im})ortance. JSTothing would be more helpful to most readers than for the author, after putting his results together in the figurative language of his peculiar system as he has done in his Guide (29), to give us a new compendium of his experiments and results, making the test of admission that which is verifiable, or at least that which the author believes will be found verifial)]e. This would not involve, of course, the omission of his important concepts of "'Schaltung,'" "Klinkung" and the like, for these are names for experimentally verifiable processes and conditions ; nor would it involve the omission of general laws, as verifiable statements tliat apply to whole classes of objects ; nor Avonld it exclude hy})othtses, presented as such, for these are propositions which the author believes will be found verifiable on further investigation. It would involve simply the omission of what the autlioi- himself recognizes as fictitious. I l)clieve it would be found that nothing of value had been lost; that the author's important work would stand out with a clearness not before attained. Further, in the technical accounts of his

332 'Journal of Cotnparafivc Neurology and Psychology.

investigations, it would be extremely helpful if the author would at least segregate carefully his verifiable, experimental, results from his fictitious schema, if he finds that he cannot bring himself to totally abandon the latter.

This demand for Anschaulichkeit" rather than verifiableness in a scientific account is what has led to an apparent opposition between V. Uexkiill's work and that of some others. Such is the case, I judge, with the dilferences between his work and my own. He presents his work in an "anschaulich" form that is confessedly not verifiable, while I have tried to present strictly what is verifiable, whether immediately anschaulich or not. The results are bound to be different in the two cases. If my work should be presented by the aid of "anschaulich" fictions, or if v. Uexkiill should present his own results without these fictions, the two accounts would show a most gratifying agreement; this is especially true now that v. Uexkiill has included, in his last Study (31) attempts to show how his machines could be modified by the influences which act on the organism. I have never argued against the existence of machinelike arrangements in organisms. JMy point was merely that these machines are not fixed and final, but that they are continually changed by the environment and by the action of the organism itself.^ Personally I believe that even these changes occur in an essentially machine-like way.

The demand for "Anschaulichkeit" at all costs is apparently what has led the author to certain extreme views ; to his sc])arating and contrasting biology and physiology; and to his tendency to fall into vital

Iu his recent paper on Nciv Questions in Ej-i>vriincntul BioJoijn (33) v. Uexkiill, iu presenting a graphic pictnre of my exposition // carried to a logical crtremc, has attribnted to nie extreme views whicli I liave never lield. He says that I "denied tlie existence of tlie I'eflex ; denied tlie existence of any struetio-e in the central nervons system." Tliis statement I am sure is given as part or an "ansclianlich" fictitious schema, not as a statement of verifial)le fact ; I have made no sucli denial. Again lie quotes me as saying that "the organism is onlij something happening." when what I said is that "The organism is something happening." The difference is like the difference between black and white. I was trying to insist upon certain facts that had been commonly left out of account. — not trying to substitute these facts for everything else known.

Jennings, Uexkull o)i Physiolugy of Behavior. ^^^

ism at certain juiictures. Having abandoned (in favor of the constrnction of fictitious machines) the requirement of finding out what are the real forces at work in organisms, of finding out Avhat machines actually do exist (as determined by the test of verification), the author finds himself in oi)p(>sition to physiology, which searches precisely for the real (verifiable) forces, imiterials and machines of organisms. To escape this oj)position, v. Uexkiill renounces physiology and all its works ; renounces finding out the causes of things, and calls himself a biologist only ; biology he nuiintains has an entirely different purpose from })hysiology. "We distinguish two sciences of aninuitc nature; Physiology, which arranges her nniterials according to causality ; Biology, which arranges her materials according to purposiveness (Zweckmassigkeit)" (29, Vorwort). The purpose of biology is to work out the plan according to Avhich the body is made up and acts (33, pp. 10, 11, etc.). The inafcrials — the actual chemical and physical substances, properties and forces — used in realizing the plan, do not concern biology, but form the field of i^hysiology (28, p. 370). Hence the biologist may content himself with schemata which r('])roduce what the organism does, even though the organism and the schenni are operated by different forces acting on different materials in different arrangements. Thus "'When I for example lay out the })lan of structure of a worm, and in so doing use any convenient jdiysical schema, it doesn't occur to me at all to touch ui)on a physical problem. One may always think of any other force as at work in the same object. I am not concerned with that. I seek only for a fitting expression in order to luake the })lan of the animal anscliaidlcli" (28, p. 377). The biologist need not concern himself with causal questions; with i)hysiology. "It is therefore not to be complained of if we biologists, who are asking about the functions of animals, look with much coolness at the end problems of ])hysiology" (28, p. 377).

denouncing then a causal study for biology, and holding that "Anschaulichkeit" or the demonstration of the production of processes in a machine-like way is the '""most essential of alT' things in l)iological explanations, v. Uexkull naturally gets into serious ditficulties when he confronts processes which he is unable to present as

334 'Journal of Cofuparative Neurology and Psychology.

"anscbaulicli" Ly "searching about for a satisfactory mechanical scheme of structure" (31, p. 188). Such he feels that he finds in all develo]3mental processes, both in development from the egg, and in the development of new features in movement and the organs of movement. It is greatly to be regretted that we must give up the hope of finding an anschaulicli structural schema for animal development. But there is no structure that could explain (veranschaulichen) its own production" (31, p. 185). Since, then, it is impossible to bring development under the <^)nly point of view which seems to V. Uexkiill to give a satisfactory explanation, he finds it necessary to take refuge in vitalism. He is, however, under no illusions as to vitalism's being an explanation ; it is a mere renunciation ; "when we therefore give over the production of form to vitalism, this giving over involves a renunciation of all real understanding in this science" (31, p. 187). In his latest jjajjor v. UexkiUl counts himself, if I understand him correctly, as a vitalist so far as develo})mental processes go, but as a "machinalist" so far as the functioning of developed organs is concerned (33, ]). 1-4).

If in place of nuiking "Anschaulichkeit" the end to be reached, one takes verifiableness as his aim, a very different set of views will be reached in biology. There are many fields of exact science in which such "Anschaulichkeit" as v. Uexkiill demands is not re(piii-ed. To understand how water is produced from ojiygen and hydrogen, most chemists do not construct a fictitious machine on the plan of a flour mill or a dynamo. They merely accept the fact as a datum, in connection with other similar facts. V. Uexkiill himself mentions a inimber of fields of science which are not "anschaulicli" in character (33, p. 16), so that it seems extraordinary to found vitalism on the basis that biology is similar to other sciences in this respect ! The only condition that science requires in oi-der that accepted principles of explanation shall apply is this ; that differences in resulting conditions shall always be found to l)e i)receded by differences in foregoing conditions, so that nothing shall happen undetermined. But why oxygen and hydrogen in the proportion of one to two should give the properties of water rather than those of alcohol we do not know any more than we know why

Jknnings, Uexkull on Physiology of Behavior. 335

in biology one combination produces a sea urchin, another a starfish. Throughout both chemistry and biology we find unpredictable results produced by new combinations. The repeated changes shown by the development of an organism seem, as to intelligibility, quite on a par with a series of transformations due to recombinations of chemicals. If in either field the same combination under the same conditions should sometimes produce one result, sometimes another, then indeed science would l)e in distress, and if Inology were the field in which this occurred, then the biologist might ])erhaps grasp at vitalism as a drowning man grasps at a straw. Our cpiotation from V. Uexkiill (given above, }). 310), in which he holds that Driesch has shown that the germ cell does not possess a trace of machine-like structure, but consists of throughout equivalent parts" and that it is "structureless," perhaps implies that he conceives this distressing condition to have been reached. But those who have spent years in working with the astoundingly complex machinelike structures and processes in the chromatin of the germ cell, and have considered the demonstrative evidence brought forward by Boveri, Wilson, Herbst and many others as to the distinctive functions of these various parts in development, will find the statement that the germ cell is structureless and composed of throughout equivalent parts so absolutely schenuitic and fictitious as to omit all the truth!

Taking verifiableness as our aim will likewise leave biology and physiology resting peacefully in union. We shall be interested in the ])lan of the organism so far as it is vcrlfiahlr; and to work out the verifiable plan we shall be forced to consider the actual forces, materials and arrangements, not fictitious ones. Doubtless physiology has in practice become narrowed ; the remedy lies in broadening it till it includes everything verifial)le in the study of the processes of organisms.

Criticism of theoretical points is not a proper close for a consideration of work of such solid value as that of v. Uexkiill. Though we may differ from him in theoretical ideal and in method of presentation, we must recognize the fundamental soundness of his methods of actual work. ITever was a truer j^rinciple set forth for

33^ ^^urnal of Comparative Neurology and Psychology.

successful biological investigation than that the first requirement is ^'^Tlie continued and accurate observation of the living animal in its environment" (29, p. 75). And v. Uexkiill has done more toward an analysis of the internal processes in the behavior of lower animals than perhaps anyone else.

The Journal of

Comparative Neurology and Psychology

Volume XIX July, 1909 Number 4

IMITATIO^T IN MONKEYS.

M. E. HAGGERTY.

From the Harvard Psychological Laboratory.

With Thirteen Figures. CONTENTS.

PAGE

I. Introductory Statements 337

II. Description and Care of Animals Studied 340

III. Method of Investigation 348

IV. Experiments and Results 355

Y. General Summary of Results and Conclusions 433

I. Introductory Statements.

1. Statement of Problem. Popular opinion has generally attributed to monkeys the ability to learn by imitation. As will appear later, experimental evidence on the matter has been of a conflicting nature, but in the main it has not supported the popular belief. The general problem of imitation presents itself in the form of two questions: Do monkeys imitate human beings? and T)o they imitate one another ? It is conceivable and, indeed, quite probable that an animal which fails to copy the acts of persons, may yet imitate individuals of its own species. In the native state, monkeys must have innumerable opportunities to imitate one another, whereas they rarely, if ever, have opportunity to imitate human beings. Further THE JorRNAL OF COMrAEATIVE NEUROLOGY AND PSYCHOLOGY. VOL. XIX, NO. 4.

33^ yournal of Comparative Neurology and Psychology.

more, a monJiey lifting a latch is a very different stimulus for an observing monkey from a person lifting the same latch. In view of these considerations it is imj)ortant in an experimental study of imitation in monkeys to deal separately with the two questions proposed above. The first question, Do monkeys imitate human beings 'f is only indirectly related to the natural activities of the animals ; the second. Do they imitate one another ? is extremely important for an understanding of the behavior and mental life of monkeys. To discover in what w^ays certain sj^ecies of monkeys are influenced by one another's acts has been the chief aim of the investigation which I have here to report.

2. History of Present Investigation. The investigation was begun in the Harvard Psychological Laboratory in October, 1907. From that time until June, 1908, the experimenter devoted himself (a) to studying the behavior of three Cebus monkeys; (b) to making experiments with these individuals for the purpose of developing methods of testing imitative ability, and (c) to devising and constructing apparatus for experimental work.

In June, 1908, the investigation was transferred to the Xew York Zoological Park in order to make use of the large collection of monkeys available there. The apparatus which had been built in Cambridge, and two of the Cebus monkeys which had been used in the preliminary experiments were taken to the Park. Here, under peculiarly favorable conditions the investigation was continued until September. Well-prepared apparatus and methods of experimental procedure, the fine collection of animals and the excellent local conditions provided by those in charge of the Park, greatly facilitated the work and within the short space of ten weeks much was accomplished in the way of results.

3. The ^Vorh of Other Investigators. — I^oteworthy observations concerning the imitative ability of monkeys have been made under experimental conditions by Thorndike\ by Kinnaman^, by Hob 'Thorndike, Edward L. The Mental Life of the Monkeys. Psycholoffical Review, Monograph Supplement, vol. 3, no. 5, 57 pp. 1901.

=KiNNAMAN, A. J. Mental Life of Two Macacus Rhesus Monkeys in Captivity. American Journal of Psyclioloyy, vol. 13, pp. 98-148; 173-218. 3902.

Haggerty, hnitation in Monkeys. 339

hoiTse,^ and liy Watson/ In the main these observations are but indirectly related to the present inv^estigation, for they are largely concerned with the animal's ability to copy the acts of hmiian beings. On this ground, the work of Hobhouse, which gave positive results, ma}^ be excluded from this discussion. The other three investigators, who studied the tendency of monkeys to imitate one another, used, in one form or another, the problem method. One monkey was taught to get food by manipulating a mechanical device; then another monkey was allowed to learn the act by watching the trained animal perform. None of the investigators has given the problem an extended study, since the observations in this particular were incidental to studies of wider scope.

Thorndike reports a series of five experiments on a Cebus monkey. This animal, "JSTo. 3," was, at the time of the experiments, "on terms of war" with J^o. 1, the animal he was to imitate. In none of the imitation tests did "ISTo. 3" learn to do the act. Thorndike concludes : "There is clearly no evidence here of any imitation of -N^o. 1 by E^o. 3. There was also apparently nothing like purposive watching on the part of ISTo. 3."^ '"This lack of any special curiosity about the doings of their own species characterized the general behavior of all three of my monkeys and in itself lessens the probability that they learn much from one another."*'

Kinnaman observed two cases where the conduct of a male rhesus caused the female to learn an act. The problem was to get food by manipulating a mechanism — in one case, the pulling of a plug, in the other, the bearing down of a lever. In each case, the female was given opportunity to get food but failed. The male was then allowed to get food while she was present and watching. In each case she went at once, after seeing the male get food, and operated the mechanism and repeated the performance numerous times later. Kinnaman says : "Here we have a copy in the form of an act. It was copied almost in detail, and that, too, so far as the place of

'Hobhouse, L. T. Mind in Evolution. Chap, X, London. 1901, 'Watson, John B, Imitation in Monlveys, Psychological Bulletin, vol, 5, pp, 169-178, 1908, =P. 40. «P, 42.

340 "Journal of Compa