Paper - Comparative studies on the growth of the cerebral cortex 8 (1918)

From Embryology
Revision as of 11:35, 17 May 2020 by Z8600021 (talk | contribs)
Embryology - 23 Sep 2021    Facebook link Pinterest link Twitter link  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)

Sugita N. Comparative studies on the growth of the cerebral cortex. VIII. General review of data for the thickness of the cerebral cortex and the size of the cortical cells in several mammals, together with some postnatal growth changes in these structures. (1918) J Comp. Neurol. 29: 241-.

Online Editor  
Mark Hill.jpg
This 1917 eighth in a series of historic papers by Sugita on the development of the cortex in the rat.



More by this author: Sugita N. Comparative studies on the growth of the cerebral cortex. I. On the changes in the size and shape of the cerebrum during the postnatal growth of the brain. Albino rat. (1917) J Comp. Neurol. 28: 495-.

Sugita N. Comparative studies on the growth of the cerebral cortex. II. On the increases in the thickness of the cerebral cortex during the postnatal growth of the brain. Albino rat. (1917) J Comp. Neurol. 28: 511-.

Sugita N. Comparative studies on the growth of the cerebral cortex. III. On the size and shape of the cerebrum in the Norway rat (Mus norvegicus) and a comparison of these with the corresponding characters in the albino rat. (1918) J Comp. Neurol. 29: 1-.

Sugita N. Comparative studies on the growth of the cerebral cortex. IV. On the thickness of the cerebral cortex of the Norway rat (Mus norvegicus) and a comparison of the same with the cortical thickness in the albino rat. (1918) J Comp. Neurol. 29: 11-.

Sugita N. Comparative studies on the growth of the cerebral cortex. V. Part I. On the area of the cortex and on the number of cells in a unit volume, measured on the frontal and sagittal sections of the albino rat brain, together with the changes in these characters according to the growth of the brain. V. Part II. On the area of the cortex and on the number of cells in a unit volume, measured on the frontal and sagittal sections of the brain of the Norway rat (Mus norvegicus), compared with the c responding data for the albino rat. (1918) J Comp. Neurol. 29: 61-117.

Sugita N. Comparative studies on the growth of the cerebral cortex. VI. Part I. On the increase in size and on the developmental changes of some nerve cells in the cerebral cortex of the albino rat during the growth of the brain. VI. Part II. On the increase in size of some nerve cells in the cerebral cortex of the Norway rat (Mus norvegicus), compared with the corresponding changes in the albino rat. (1918) J Comp. Neurol. 29: 119-.

Sugita N. Comparative studies on the growth of the cerebral cortex. VII. On the influence of starvation at an early age upon the development of the cerebral cortex. Albino rat. (1918) J Comp. Neurol. 29: 177-.

Sugita N. Comparative studies on the growth of the cerebral cortex. VIII. General review of data for the thickness of the cerebral cortex and the size of the cortical cells in several mammals, together with some postnatal growth changes in these structures. (1918) J Comp. Neurol. 29: 241-.

Modern Notes: cortex | rat

Neural Links: ectoderm | neural | neural crest | ventricular | sensory | Stage 22 | gliogenesis | neural fetal | Medicine Lecture - Neural | Lecture - Ectoderm | Lecture - Neural Crest | Lab - Early Neural | neural abnormalities | folic acid | iodine deficiency | Fetal Alcohol Syndrome | neural postnatal | neural examination | Histology | Historic Neural | Category:Neural



Historic Disclaimer - information about historic embryology pages 
Mark Hill.jpg
Pages where the terms "Historic" (textbooks, papers, people, recommendations) appear on this site, and sections within pages where this disclaimer appears, indicate that the content and scientific understanding are specific to the time of publication. This means that while some scientific descriptions are still accurate, the terminology and interpretation of the developmental mechanisms reflect the understanding at the time of original publication and those of the preceding periods, these terms, interpretations and recommendations may not reflect our current scientific understanding.     (More? Embryology History | Historic Embryology Papers)

Comparative Studies on the Growth of the Cerebral Cortex

VIII. General Review Of Data For The Thickness Of The Cerebral Cortex And The Size Of The Cortical Cells In Several Mammals, Together With Some Postnatal Growth Changes In These Structures

Prof. Naoki Sugita (1887-1949)
Prof. Naoki Sugita (1887-1949)

Naoki Sugita

From The Wistar Institute of Anatomy and Biology

Three Figures And Two Charts

I. Introduction

Years ago Schwalbe ('81) pointed out as characteristic somatic expressions, which might be taken to indicate the grade of intelhgence of a species of animals, the following four measurements on the brain: 1) total weight of the brain; 2) total number of nerve cells in the brain; 3) total area of the surface of the hemispheres of the brain, and 4) the thickness of the cerebral cortex. Since then he and many other neurologists have endeavored to gather data on the morphological evidence for the development of mental ability. Donaldson and Hatai ('The Rat,' Donaldson, '15) have made systematic observations on the growth changes in the central nervous system as well as in other organs and systems, using exclusively the albino rat. As a result of their investigations, the postnatal growth of the brain and the spinal cord, in gross measurements, and the relations of these to the other systems during growth have been determined. In line with these studies, I also made further researches on the growth in the thickness of the cerebral cortex, the size and shape of the cortical nerve cells and the relative number of the cortical cells in both the Norway and albino rats. The results of these researches have been already presented (Sugita, '17,, '17 a, '18, '18 a, '18 b, '18 c, '18 d), with references to some similar studies by other authors. These data give us a general idea of the postnatal development of the cerebral cortex in a representative mammal (albino rat), and we may fairly infer that similar changes occur in other mammals during the growth of the brain. To test how^ far my conclusions on the mode of the development of the cerebral elements during postnatal life may be extended, I shall review and summarize in the present paper the results obtained by several authors on the development of the cortex in other mammals and make a comparison of their results with the data obtained by me.


II. Thickness of the Cerebral Cortex in the Albino Rat

The results obtained by me regarding the cortical thickness in the brain of the albino rat may be summarzed as follows (Sugita, '17 a):

1. The cortex at the frontal pole of the hemisphere is the thickest and that at the occipital pole is the thinnest. Speaking in general terms, the cortex diminishes in thickness from the frontal to the occipital pole and from the dorsal to the ventral aspect.

2. After birth, the general average of the cortical thickness increases very rapidly during the first ten days, thickening from 0.74 mm. at birth to 1.73 mm. at ten days, more than twice the thickness at birth, while the brain weight increases from 0.25 gram to 0.95 gram during the same period. This is designated by me the first phase of the cortical development.

3. Between the tenth and the tw^entieth day after birth, the cortical thickness increases more slowly, attaining at twenty days to within 4 per cent of the full thickness of the cortex, namely, 1.84 mm., or about 2.5 times the thickness at birth, while the brain weight increases to 1.15 grams. This is designated the second phase of the cortical development.

4. From the twentieth to the ninetieth day, the cortical thickness increases but little on the average, attaining at ninety days the thickness of 1.93 mm., or 2.6 times the thickness at birth, w^hile the brain weight has increased to about 1.80 grams. This is designated the third phase of the cortical development.


After the ninetieth day, there is no significant change in the thickness of the cortex, but the area of the cortex increases as the brain weight rises and at 2.0 grams is greater than at 1.15 grams (20 days) by about 45 per cent.

5. In the first phase the cortex increases its thickness by receiving some newly formed cells from the matrix and many already formed from the transitional layers and at the same time by the general enlargement of the neurons, especia ly the cell bodies; in the second phase, however, it grows main'y by the enlargement of the cell bodies and the growth of the axons and dendrites; while during the third phase it thickens only slightly, but extends in area as the result of the ingrowing axons and the formation of the myelin sheaths and non-nervous structures.

6. The cortex at the frontal pole increases its thickness very rapidly and steadily, continuing to do this even after the end of the second phase, while at all the other localities the cortex thickens in the same proportion, so that at the end of the second phase all the localities reach nearly the full thickness, but maintain their initial relations. The localities heterogeneous in their cell lamination show in the course of thickening some deviation from the localities which are typical.

7. The cortex generally attains nearly its full thickness before myelination, as shown by the Weigert staining method, occurs in it. In the Albino, the cortex has nearly its mature thickness at twenty days, just before the young rat is weaned and when the brain has attained only a trifle more than half its final weight. The growth of the cortex in thickness is therefore precocious.


III. Increase in Cortical Thickness during Growth of the Brains of the Mouse and the Guinea-Pig

Mouse

Isenschmid ('11) has made a study of the cortical cell lamination in the brain of the mouse and given a map of the topographic localization in the hemisphere, which is reproduced here as figure 1. De Vries ('12) and Rose ('12) have also presented a brain map of the mouse according to their studies on the cell architecture of the cortex ; a map which resembles that



Fig. 1 Cortical area of the mouse (Mus musculus) — reproduced from the original by Isenschmid ('11); the thickness of the cortex at each area designated on the map is tabulated in table 1 of this paper. Double lines show borders of three — the dorsolateral, the frontomedial, and the suboccipital — regions of the neopallium. Shaded parts (areas t, s, and h) do not lie in the same (median) plane as the other areas. A = Dorsal view of the right hemisphere; B = Lateral view of the right hemisphere; C = Medial view of the right hemisphere. B.olf. = Bulbus olfactorius; C = Corpus callosum; c.A. = Cornu Ammonis; cl. = Claustrum; s.;?. = Septum pellucidum. s — s' = the level corresponding to that from which the sagittal sections of the Albino brain were taken by me; /— /' = the evel corresponding to that from which the frontal sections of the Albino brain were taken by me; h — h' = the level corresponding to that from which the horizontal sections of the Albino were taken by me.



TABLE 1 Giving for each localiUj of the brain of the adult mouse the characteristics of the cell lamination, the thickness of the cortex on the slide as deterinined from the photograms given by Isenschmid {'11), and the relative thickness of the outer and inner layers as presented by the same author. For the localities consult figure 1 in this paper, which was reproduced from the original of Isenschmid {'11)


AKEA

(fig. 1)


CHARACTERISTICS OF. THE AREA IN CELL-LAMINATION


e

f

i

k

1

m

r

q

s t


Largest ganglion cells contained (18 X 20 m)- Not so large cells

IV layer thick

Transitional part

Paleopallium

IV layer not so well developed

Adjoins to fovea limbica, cell lamination not clear

Transitional part (ganglion cells: 13 X 15 m)At the corner (ganglion cells: 12 X 14 /u)

(Ganglion cells : 15 X 18 m)

Similar to area q


THICKNESS OF THE CORTEX


0.73

0.86 0.50 0.53


0.62 0.44 0.81 0.78 0.71-0.61 1.201 0.56 0.26 0.34


RELATIVE

THICKNESS OF

THE OUTER AND

INNER LAYERS

OF THE

CORTEX

outer: inner'


48:52

45:55 45:55 45:55


42:58


34:66 23:77

22:78 28:72


1 Section cut obliquely.

^ The outer layer = the lamina granularis externa plus the lamina pyramidalis plus the lamina granularis interna. The inner layer = the lamina ganglionaris plus the lamina multiformis.

of Isenschmid. Isenschmid ('11) has recorded the thickness of the cerebral cortex of the mouse on the sHde at every locahty mapped in his figure (fig. 1). But the actual thickness not being given explicitly for each locality, I calculated the values from the direct measurements made on the photograms. The brain was fixed in alcohol, imbedded in paraffine and cut in 10-micra sections and stained with kresyl- violet. The thicknesses of the cortex on the slide as thus obtained are given for each locality in table 1 and also are condensed in table 2, in which the data


246


NAOKI SUGITA


TABLE 2

A com'parison of the thicknesses of the cerebral cortex at several corresponding localities in the albino rat and in the mouse. The data for the albino rat were taken from table 11 in ?ny previous paper (Sugita, '17 a, p. 578) and the data for the mouse were taken from a paper by Isenschmid {'11). The order of increasing thickness is the same in both animals


ALBINO RAT


MOUSE


Locality


Average

thickness of

cortex by

locality


Corresponding locality


Thickness of cortex at each of the localities


Average

thickness of

cortex by

locality



mm .



mm.


mm.


V and XIII


1.24


C


0.50


0.50


IV


1.42


d


0.53


0.53


XII and VIII


1.67


e and i


0.65 and 0.44


0.55


III and XI


1.91


a and e


0.73 and 0.65


0.69


VI


2 01


1 (corner)


0.78


0.78


II and X


2.03


k and b


0.81 and 0.86


0.84


VII


2.29


b


0.86


0.86


I and IX


2.99


frontal pole


1.00


1.00


Average


1.94


Average


0.72




for the Albino are so entered that the cortical thicknesses at the corresponding localities in the two forms may be compared. The order of the localities is arranged according to the increasing thickness in the Albino (taken from table 11, Sugita, '17 a, p. 578). The average value of the cortical thickness in the mouse is, on the slide, 0.72 mm., and if corrected to the fresh condition would probably be somewhat thinner than one-half the average thickness of the Albino cortex. The order of the thickness according to localities is quite the same, so that in both forms the cortical thickness decreases from the frontal to the occipital pole and from the dorsal to the ventral aspect. Moreover, the cortex at the frontal pole is the thickest and has double the thickness of that at the occipital pole.

As seen in figure 1, the cerebral hemisphere is divided by Isenschmid into three main regions — the dorsolateral, frontomedial and suboccipital regions — separated by the double line in figure 1.

The average cortical thickness in the dorsolateral region (fig. 1 a) is 0.56 mm. at its hinder-medial part and 0.90 mm.



at its fore-lateral part, and in this region the lamina zonalis is about one-twelfth, the main outer layers (the lamina granulans externa plus the lamina pyramidalis plus the lamina granulans interna) about two-fifths and the main inner layers (the lamina ganglionaris plus the lamina multiformis) about one-half the total thickness of the cortex. In the frontomedial region (fig. 1 c) the cortical thickness at the frontal pole is 1.00 mm. and that at the caudal part is 0.35 mm., while in the suboccipital region the cortical thickness ranges between 0.2 and 0.3 mm.




Fig. 2 Showing diagrammatically the thickness of the cerebral cortex at locality a in the mouse at different ages. Reproduced from the original given by Isenschmid ('11). B = at birth. M = at maturity. The other arable numbers show the age in days. I = lamina zonalis; II = lamina granularis externa; III = lamina pyramidalis; IV = lamina granularis interna; V = lamina ganglionaris; VI = lamina multiformis. The cell outlines found between the last two diagrams indicate the relative size and shape of the cells in each cortical layer.

Isenschmid has given also diagrams illustrating the growth in cortical thickness at locality 'a' (fig. 1 a, corresponding approximately to locality III in my study, fig. 2, Sugita, '17 a, p. 525), sampled from material at several different ages and magnified uniformly. These are also reproduced here as figure 2. The diagrams show that as age advances the lamina pyramidalis (II and III) thickens steadily and continuously while the lamina ganglionaris (V) and especially the lamina multiformis (VI) grow much less rapidly. Chart 1 gives a comparison of the increase in the cortical thickness at corresponding localities (locality 'a' of the mouse and locality III of the albino rat) in the albino rat and the mouse, the data being from Isenschmid ('11) and Sugita ('17 a). In the Albino the cortex attains nearly its full thickness at twenty days (weaning time), while in the mouse this stage was reached between twelve and seventeen days of age, very closely corresponding to the weaning time of


mm. 2.0r

18

16

lA

J.2

1.0

0.8

0.6

0.4

Q2













Chart 1 Giving the cortical thickness of the albino rat and of the mouse according to age. The data for the albino rat are taken from Sugita ('17 a) at locality III measured on the sagittal section and the data for the mouse are taken from Isenschmid ('11) at locality 'a.' These two localities approximately correspond.

the mouse, which is fifteen days. The remarkable phase during which the rapid increase in cortical thickness takes place in the Albino (first ten days after birth) cannot be clearly identified on the graph for the mouse cortex. It must be recalled, however, that data on the mouse cortex have not been corrected for the action of the reagents, while the data for the rat have been so corrected. The outstanding fact, however, is that the cerebral cortex in both forms attains nearly its full thickness just before the weaning time.


Guinea-pig

I have had the opportunity at The Wistar Institute to examine the sections of the guinea-pig brains prepared by Allen ('04) for her study on the myelination of the nervous system of that animal. The sections were cut in series in the frontal plane from material fixed in Miiller's fluid, imbedded in celloidin and stained by Weigert's method for the myelin sheaths. The thickness of the cerebral cortex in the adult guinea-pig (body weight, 618 grams; brain weight not recorded) is on the average 1.90 mm. (1.80 mm., 1.88 mm,, and 2.01 mm., respectively, at the localities corresponding to localities VI, VII, and VIII examined by me on the frontal section of the Albino brain at the level of the commissura anterior). The corresponding measurements at birth (body weight, 108 grams) are 1.71 mm. (and 1.51 mm., 1.75 mm., and 1.86 mm., respectively) and those at thirty-five days (body weight, 250 grams) are 1.85 mm. (and 1.77 mm., 1.86 mm., and 1.92 mm., respectively). So, from birth on to the maturity, the cortical thickness has on the average increased only 11 per cent. According to Allen, the guineapig at birth is covered with hair, has complete muscular development, and is almost independent of the mother, the central nervous system being practically completely myelinated, whereas, by contrast, the albino rat is born quite naked, extremely helpless and undeveloped, and myelination in the brain has not begun. The guinea-pig is psychically mature soon after birth (three days after birth) ; the degree of development of the central nervous system of the new-born guinea-pig corresponds to that of the albino rat at twenty-three to twenty-seven days or its period of first psychical maturity. A new-born guinea-pig is fobnd to have a cerebral cortex in which the myelination is going on.

Comparing the sections from the guinea-pig brain with those from the albino rat brain, it appears that the new-born guinea-pig corresponds to the albino rat of about ten days in cortical thickness, but seems to be older when judged by the myelination of the cortex. This coincides with observation that the guineapig is, almost from the start, relatively independent of the mother.


250 NAOKI SUGITA

IV. THE CORTICAL THICKNESS AT SEVERAL LOCALITIES IN THE BRAINS OF SOME MAMMALS OTHER THAN THE RAT

Few papers have been published regarding the differences in the thickness of the cerebral cortex at given localities of the brain in mammals other than the rat, except for man. Yet even in these cases, the techniques of hardening, imbedding, and staining used by the different authors are dissimilar and their results are accordingly not precisely comparable. Despite this, however, it has seemed worth while to make a survey of the data at hand.

Rabbit. Bevan Lewis (^81) has given as the natural thickness^ of the cerebral cortex of the adult rabbit the following figures (table 3) according to localities. For the localities, the map made by him and reproduced by me in a previous paper (Sugita, '17 a, fig. 10, p. 544) should be here consulted. He has presented the thicknesses of every layer of the cortex separately, but here only the total cortical thicknesses, as computed by me from his data, are given in round numbers.

Pig. Lewis ('79) has also determined the cortical thickness at several localities in the pig brain (the names of the localities

TABLE 3

The thickness of the cerebral cortex of the rabbit, quoted from Bevan Lewis {'81)

Depth of cortex on a plane with genu of corpus callosum :

mm.

Gyrus fornicatus 1.72

Sagittal angle 2.23

Extra-limbic 2.81

Near limbic sulcus 2.31

Depth of cortex on a jilane with posterior border of corpus eallosum:

Gyrus fornicatus 1 . 70

Sagittal angle 1.91

Extra-limbic 2 . 46

Depth of cortex of the modified lower limbic t3'pe 2.23 to 2.47

Depth of cortex in the cornu Ammonis:

Anterior regions 2 . 27

Average at six different sites 2 . 23

1 Lewis measured the cortical thickness on sections cut by the freezing microtome from fresh material and then hardened by osmic acid, stained by aniline black and mounted in Canada balsam. According to his statement we obtain, by this method, the natural depth o"" the cortex, no shrinkage occurring if the preparations have been carefully made (Lewis, '78).


Limbic lobe <


Upper parietal convolutions <


Lower parietal convolutions.


GROWTH OF THE CEREBRAL CORTEX 251

TABLE 4

The thickness of the cerebral cortex of the pig, quoted from Bevan Lewis {'79)

Depth of cortex from before backward:

mm.

'4.97 4.48 3.70 4.98 3.53 3.77

Average 4.22

fa. 28 2.65 3.08 3.91 4.23 3.44

Average 3 .50

■3.44 3.91 3.95 3.35 3.02 3.67

Average 3.64

are analogous to those given for the rabbit brain, loc. cit.)- His results are summarized in table 4. These values are distinctly high compared with those for other mammals, as shown in the various tables in this paper. These results taken together with those for the rabbit just given, which are also noticeably high, suggest that the determination by Lewis are for some reason systematically too high.

Marsupials to man. Table 5 is quoted (slightly modified) from Brodmann ('09) and gives for several species of mammals, including man, the cortical thickness at six localities (areae precentralis, frontalis, parietalis, occipitaUs, hippocampica et retrosplenialis) in the brain of each animal. The sections were made by hardening the material in 4 per cent formaldehyde, imbedding in paraffine, and staining by the modified Nissl's method, and the cortical thickness was measured by the micrometer directly on the slide. The average thickness was calculated by me for the four areas, excluding the areae hippocampica et retrosplenialis which are heterogeneous in cell lamination.


252


NAOKI SUGITA


TABLE 5

The cortical thickness at the corresponding parts of the cerebral hemisphere in different mammals, quoted from Brodmann i'09). According to his nomenclature, area precentralis = type 4, area frontalis agranularis = typed, area parietalis = type 7, area occipitalis = type 17, area hippocampica = type 28, and area retrosplenialis = type 29, as given in his 'Hirnkarte' {Brodmann, '09)


Homo sapiens (man) Cercopithecus (longtailed ape)

Lemur

Hapale (marmoset) .

Pteropus edwardsii (vampire bat). . . .

Erinaceus europaeus (hedgehog)

Cercoleptes caudivolvulus (kinkajou)

Lepus cuniculus (rabbit)

Spermophilus citillus (ground squirrel)

Macropus giganteus (kangaroo). .


grams

60,000

2,500

1,800

200


375 700

2,000

2,200

200

5,000


grams

1,400

85 23


7 3.5


10


2.2


3.0-4.5

3.0 2.3 2.15


1.9 1.87

2.17

2.7

2.1

2.8-3.1


O < 03 t^


3.0-3.8

2.5 2.3

2.17


1.6 2.1

2.0

2.33

2.18


3.08

2.0

1.67

1.73


1.7 1.78

1.7 2.2 1.73 2.2


2.3-2.6

1.7

1.55

1.26


1.76 1.5

1.9


1.37


1.9


mm.

2.5

1.6

1.35

1.14


1.52 1.6

1.9 1.2 1.13 1.7


< 2


2.3

1.1

1.19

1.07


1.4-1.76 0.8

1.67

0.8-1.5

0.75

1.2


3.0

1.95 1.73 1.59


1.66 1.61

1.89 1.79 1.54 2.15


Reviewing this table, it is readily seen that, within each order, the animal which has a greater brain weight shows also a greater cortical thickness, but a fixed relation between the brain weight and the cortical thickness has not been here revealed. In different orders, this relation is not true; the lemur and the kangaroo have a similar brain weight (23 to 25 grams),


GROWTH OF THE CEREBRAL CORTEX


253


while the cortical thickness in the latter is much greater (by about 25 per cent).

Prosimiae and primates. The following table (table 6) is summarized from a paper by Marburg ('12) and shows for some species of the prosimiae and primates the total cortical thickness measured at four representative localities (gyri centralis, frontalis, temporalis et occipitalis) . The average values were taken by me.

TABLE 6

Thickness of the cerebral cortex at several localities in monkeys, as presented by

Marburg {'12). Averages are calculated by me


Simla satyrus

Hylobates (sp.?)

Semnopithecus nasicus. . .

Macacus rhesus

Cynocephalus hamadryas

Ateles niger

Lemur varius






AVEI


CENTRAL, GYRUS


FRONTAL GYRUS


TEMPORAL GYRUS


OCCIPITAL, GYRUS


Of the four

localities

m »i .


7n7n .


7)1 711 .


mm.


mm.


3.11


2.97


2.43




3.78


3.24


2.51


1.78


2.83


3.78


2.43


2.43


1.35


2.50


"2.84


2.70


2.15


1.49


2.30


2.97


2.70


2.03


1.35


2.26


2.97


2.84


2.43




1.30


1.76


1.76


1.67


1.62


Of the three localities


2.84 3.18 2.88 2.56 2.57 2.75 1.61


This table also suggests that, in the order of monkeys, the average thickness of the cortex varies so that those which have the greater brain weight have also t|ie greater thickness of the cerebral cortex, but the brain weights are not available for comparison.

V. THE THICKNESS OF THE CEREBRAL CORTEX IN MAN

Man. There are scores of papers giving the measurements of the thickness of the cerebral cortex in man, but they are diverse in the techniques used for preparing the material, in the localities selected for measurement, and also in the manner of measurement. The results published before 1891 were all summarized by Donaldson ('91), but the table is not reproduced here as, owing to the lack of the information necessary for the interpretation of the values found, it has mainly an historical interest.


254


NAOKI SUGITA


Donaldson ('91) measured also the thickness of the cerebral cortex at fourteen localities from each hemisphere of nine normal brains (six males and three females), as shown in figure 3 reproduced from his original paper, in order to obtain control




Fig. 3 This figure shows the localities on the hemispheres from which the samples of cortex were taken by Donaldson ('91). For the thickness of cortex at each locality see table 8 and chart 2. A = Lateral aspect. 3 is used to designate the insula, here not exposed. B = Ventral aspect; C = Mesial aspect.

values for the study of the brain of a blind deaf-mute, Laura Bridgeman. The technique employed by Donaldson was fixation in bichromate and alcohol (potassium bichromate 2| per cent plus I its volume of 95 per cent alcohol) for six to eight


GROWTH OF THE CEREBRAL CORTEX


255


TABLE 7

Giving the average cortical thickness of man, arranged according to age and sex, together with the brain weight. Quoted from Donaldson {'91)


BRAI^f WEIGHT


AVERAGE CORTICAL THICKNESS


Males


years


grams


7nm.


35


1419


2.81


35


1443


2.98


39


1393


2.82


45


1367


2.92


57


1464


2.94


?


1210


3.11


Females


40


1196


2.74


45


1173


2.90


?


1312


3.07


Average


1331


2.92


weeks, washing in water for twenty-four hours, 95 per cent alcohol for two days, final preservation in 80 per cent alcohol, and imbedding in celloidin. The sections were cut about 100 micra thick and measured unstained under a low magnifying power with a micrometer eyepiece, at the summit of the gyrus arid at the side, midway between the summit and the bottom of the bounding sulcus. To obtain the average thickness at the locality, the smaller figure was multiplied by 2, added to the larger figure, and the sum divided by 3.

Table 7 shows the average thickness of the cortex (taken from the fourteen localities) arranged according to sex and age, quoted from Donaldson ('91). If we take the nine cases in this table as the basis for computation, we find the mean thickness of the cortex to be 2.92 mm., with a probable error of the mean equal to ±0.026 mm.

I wish to cite also the average thickness of the cortex, as thus obtained by Donaldson ('91), according to locality (table 8). These localities are shown in figure 3 and the relative thickness


256


NAOKI SUGITA


of the cortex at each is graphically presented in chart 2. Generally summarized, the average thicknes of the cortex of the adult man is 2.92 mm.; females have a slightly thinner cortex than males (differences less than 1 per cent, or 0.02 mm.) and the right hemisphere usually has a cortex a few per cent less thick than the left (maximmn difference 7 per cent).

With the foregoing determinations are to be compared the measurements by three other observers.


wm


\


1^






]--















\


\


\.


-^

















A.Q














4 A





























A















t


5 I


' 6


)4


2


, £


1(


) 1


i


M\


li


l^


> S


M4


Chart 2. -The curve was plotted according to table 8 to show the cortical thickness at each locality as measured by Donaldson ('91. The numbers placed by the ordinates indicate the thickness of the cortex in millimeters. The numbers for the localities are given below, and correspond to those in figure 3, A. B and C.

In accordance with this plan, the results obtained in a careful study by Hammarberg ('95) are tabulated in table 8. The material used for this study was a brain of a male, twenty-eight years old, mentally normal, and who died of typhoid fever. The technique employed was fixation in 95 per cent alcohol, imbedding in paraffine by means of xylol, sections 10 micra in


TABLE 8 Giving for several localities on the hemisphere of the adult human brain the thickness of the cortex, as measured by different authors. The general average thickness was taken, averaging all measurements presented by each author. For reasons given in the text, these averages as they stand are by no means comparable with each other. The data were taken from Donaldson ('91), Hammarberg {'95), Campbell {'05), and Brodmann {'08)



ocaJLtu


Au/A-<jT


Tlanaldsan


Hamtnurfc&rj


C(Xtm


pbel/


Brocf


T)ann


L


Ki^iti o/ secti'uTv


Cell


Cell


Cell


Fi ber


Cell


Fiber 1



U.n,U


rvvrn.


m^


m^


n\/yn.


«^



^


Gyi-ixs centralis


o/n tenor



2.?7


2.^-0


2.62.


a gi'


4 05

^.


Gyrus fenfroJIs pos/erior


oral sic^e


3.08


2.70


2.20


^.1:1,


/ ?6


/.9i


^


ir\Ter mediate p^rf


Z.9S


3./6



ccL^otf/ai. s^ale.


2.6


/■ <Jo


2.4 3


2. ST/


Q.


IiOTver end of su


cus Rolo/nJ;.






2.53


P


Lobu-lus pa-ra.ce


■ntra/is


2.?6







O o s p

1:


Gyr«s fron-talis Superior


fwnc/e<* po^t



a/0


2.62^


Z.SZ.


3.8 2


3.84


m-i^<J/e fjcw-r



3.93



fore fa^f



2.60


3.45



G^rus -fronfa/is


«ec/ius


3.09


3.4-0


2.A-0


%.I0


3.5 7



Gyrus froTitcdis i/nferioT


?iM-s ofjerccilarij


ao8


2.50




3.sa



Ta^s tria/*^£^a.fs


2.98


3.00




3.34



Po/rs or ()i Tali's






3.60



Gyrus rectus


2. J3





3.17



Frcmtal pole




2.37


1. EZ


3.07



» o 1"^


Gyrus ^(wictalis


su.f)e^ioT




2.3 7


%2,5


3.08


3.2.0


extr&mc fore f!(irt






2.93


2.SS

Gyrus ficurieTaJis mfe/ricn

Gyrus a^gn-laris



2.4-3


2.50


2,0


3.3 5


3.n


G)frus suprama/rsmalis





3.3 1


3.25


§ ^

u


Gyrus occt ti/oVci


fore p<M-t


2.6/


(. 80


2.50


Z.5Z


2.. 6 8


2.8 3


polflu- pa^t


2-.S4

a. 33


pr

Canei« 



%.52,


2.3 8


1. g%


1. 4


a. 3 8


a.47



Gyrws lu*7gK.a/i's


2.65







t

Gyrus Temboro-lis


iu./)erlor


3. /O


2.64

Z.GO


2.4


38 1


3.8 3



in -f'lssu-ra. Sy/vli




Z.SI


1. 90


3.3 5


3.80


w


tio. exfarMO,! pa^f




3.57


3. 9^


1


Pole of fem^o


cd lobe






3 70


3.8 7


-k

Gyrus Te-rK-fKrraJis


m^JUujti


3./ 5



2.68


2.25


3.SZ


3.64


L


Gyrus TewJioraliS


mf&rior




3.47


3.4 2/


extreme lundie^ pmrt




Z.91



lr,s^Ia.


3.38


2.34

2.67


2.6 2.




G


yrcs


pe^i rhyiina.1 pcwt


3.04




2.07



ec/or/tina-l ^a-rt"





2.93



presu>bi(U<.(ar fort





Z.Z5


2.70


It^CU-S





2.53


3.10


Sic-t'i C-M-lu/m





2 33


2.60


Gv/rns


— fjosterior y/eirfra/is


27 S





2.<?7


2,.94

- posl&rior clorsaJ'\s





3 10


303


-otifericnr ve^tra^is





2.0s


3.) 7


— cuiTerror tJorso/is





3.48


3 44


Prege^uoJ piwt





I. 80



Su^bgc/nuo.! pO/rt





2.35



rerro5f>l&»via.l fiO/rt





2.3 2.9 7




G-e*icro.| cxve


rage.


2.<?2.


£.i"7


Z.^6


2-/7


3.00


3. 17


257


THE JOUBNAL OF COMPARATIVE NEUROLOGY, VOL. 29, NO. 3


258 NAOKI SUGITA

thickness and staining with methyleneblue. Hammarberg claims that after twenty-four hours in 95 per cent alcohol the brain piece shrinks about 20.5 per cent in volume and the cortical thickness diminishes by 0.1 to 0.2 mm., but that during the subsequent procedures no significant size changes occur. According to his results, the gyri frontales have the thickest cortex (about 3.0 mm.) and the lobus centralis or insula is the thinnest among localities typical in cell lamination. This latter part as measured by Donaldson shows the thickest cortex.

Campbell ('05) gave two series of determinations of the cortical lamination of the human brain, after cell staining and after fiber staining, represented by uniformly magnified illustrations of the sections at the several localities. Making use of his illustrations, I obtained a series of cortical thicknesses at different localities (table 8), reduced to the actual thickness on the slide, by dividing the direct measurement on the illustration by the magnification. His sections were taken from the material fixed in Miiller's or Orth's fluid and imbedded in celloidin, cut at 25 micra, and stained with thionine. The general average thickness thus obtained, the two series combined, is about 2.3 mm.

Brodmann ('08) also has measured the thickness of the cortex on the human brains at forty-two different localities on sections prepared by two different methods: one set was fixed in 4 per cent formaldehyde, imbedded in paraffine, and stained by Nissl's method for cell study, and the other, fixed in Miiller's fluid, imbedded in celloidin, and stained by Weigert's method for the myelin sheaths. His results, which are the averages from brains between seventeen and forty-five years in age, are also tabulated in table 8 for a comparison. The general average thickness given by Brodmann is about 3.09 mm.

Kaes ('07) also studied the growth in thickness of the human cerebral cortex, measured at twelve different localities on the hemisphere, using sections fixed in Miiller's fluid and stained by Weigert's method. His results are remarkably high, giving 4.9 mm. on the general average. His method of measuring the cortex is so arbitrary and peculiar, however, that his results are not included in this table 8.


GROWTH OF THE CEREBRAL CORTEX 259

Bevan Lewis ('79) has given as the average depth of the human cortex the figures as high as 4,84 to 5.70 mm., a higher vahie even than that of Kaes. His results for both the pig and rabbit cortex were also very high, compared with those obtained for other mammals. These results suggest that his technique, which he claims gives the natural depth of the cortex, is likely to produce very high values.

Reviewing the table (table 8) , the values for the cortical thickness given for a fixed part of the hemisphere by different authors are by no means in accord ; the results by Brodmann stand close to the results by Donaldson, while those given by Campbell are the lowest, less than one-half the values given by Lewis. These differences are probably due mainly to differences in technique and are not to be attributed to variations within the sanie species, as the series of Donaldson (table 7) and my previous study (Sugita, '17 a) both have shown that individual variations in cortical thickness, obtained by the use of the same technique, are low as compared with the variations for other body measurements.

On the average, the figures given by Donaldson and Brodmann are fairly close and the former being somewhat lower, probably because Donaldson took the average from the values at the summit and at the sides of the gyrus, while Brodmann has measured the thickness at the summit only. The figures given by Hammarberg and Campbell are low, probably owing to the shrinkage of the material during preparation, as may be inferred from the descriptions by the authors and from the studies on the effects of fixing fluids by King ('10) and by me (Sugita, '17 a, '18 b, '18 c).

Despite the apparent irregularity among the figures given for the cortical thickness at different localities by the several authors, as shown in table 8, there are some general relations which are fairly clear. If we examine table 9 in which has been entered for each region the average thickness obtained by each author, it may be safely said that this table (and also table 6 for the monkeys) shows that in man (and the primates) the cerebral cortex differs normally according to locality. The


260


NAOKI SUGITA


TABLE 9

Giving the average cortical thickness for several lobes and regions {with typical cell lamination) of the cerebral hemisphere as given by different authors, and the order of localities according to the cortical thickness, together with the difference in thickness between. the temporal and occipital regions. R = regio Rolandica, F = lobus frontalis, P = lobus parietalis, = lobus occipitalis, T = lobus temporalis. Based on table 8 in this paper


LOCALITY


DONALDSON

('91) (Cell)


HAMMARBERG

('95) (Cell)


CAMPBELL

('05) (Cell)


CAMPBELL

('05) (Fiber)


BRODMANN

('08) (Cell)


BRODMANN

("08) (Fiber)


Regio Rolandica

Lobus frontalis

Lobus parietalis

Lobus occipitalis

Lobus temporalis


mm.

2.92 2 92

2.59 3.21


mm.

2.34 2.92

2.43

2.09

2.49


mm.

2.43 2.46

2.44

2.16

2.64


77im.

2.21 2.15

2.13

1.96

2.29


mm.

2.74 3.50

3.17

2.47

3.48


mtn.

2.93 3.84

3.12

2.54

3.75




Average


2.91


2.45


2.43


2.15


2.92


3.16


Order of the above five localities as to the thickness


TFRO?


FTPRO


TFPRO


TRFPO


FTPRO


FTPRO


Difference between T and


0.62


0.40


0.48


0.33


1.01


1.21




frontal and temporal regions have in all cases the thickest cortex and the occipital region is the thinnest, while the position for the cortex of the parietal and Rolandic regions is less fixed. These thickness relations support the earlier statement made by me for the rat cortex that the thickness diminishes from the frontal to the occipital pole and from the dorsal to the ventral aspect.

Brodmann ('08) has concluded from his careful study that regional characteristics for the cortical thickness clearly exist. Diese sind in alien normalen Gehirnen gesetzmassig und kon


GROWTH OF THE CEREBRAL CORTEX


261


stant und bilden ein Hauptmerkmal der struktuellen Verschiedenheiten der Gehirnoberflache; jedes Strukturfeld besitzt demnach eine bestimmte, mittlere Durchschnittsbreite, durch welche es sich von den Nachbarfeldern auszeichnet." On the other hand, local variations within a fixed area are small, while individual differences between different brains for each locality may run sometimes as high as 0.5 mm. or more.

VI. INCREASE IN CORTICAL THICKNESS DURING THE GROWTH OF THE BRAIN OF THE MAN


From the point of view of the growth changes, there have been only few studies on the human cerebral cortex ever published. Kaes ('05, '07, '09), employing forty-one human brains (twentyeight males and thirteen females, normal and pathological combined) of different ages and of different grades of intelligence, studied the cerebral cortex for the purpose of following the growth changes in it. He took his sections from twelve localities in each hemisphere, stained the fibers by Weigert's method and measured the so-called cortical thickness from the ectal border of the Meynert's arcuate fibers (or fibrae propriae) to the ectal border of the zonal layer. His conclusions on the growth

^ His ('04) has given the following values as the cortical thicknesses measured at different localities of the hemisphere of the human embryos in early months, at different stages of intrauterine development- — measured directly on the sections imbedded in paraffine.


AGE OF EMBRYOS


AT CORPUS STRIATUM

M


AT LATERAL WALL OP THALAMUS


AT LATERAL

WALL OF HEMISPHERE (BASAL PART)

M


AT LATERAL

WALL OF

HEMISPHERE

(MID PART)

M


AT MEDIAN

WALL OF HEMISPHERE


AT BOTTOM OF

SULCUS CINGULI

M


1 2


50-55 65-75


4 5 6

7 8


150

360

800

1300

2000


130

160 300 600 900


300

400


110 120 130

170 200


90 110

130


60

50 40 30 30


262 NAOKI SUGITA

changes, briefly stated, are as follows: The average thickness of the cortex diminishes rapidly from his first entry (three months old, 5.58 mm.) to the twenty-third year (4.44 mm.) and is followed by an increase up to the forty-fifth year (5.71 mm.), where it is to be noted that the thickness attained is even greater than that at birth. Then it undergoes a second thinning up to the old age (at ninety-seventh year, his last entry, 4.62 mm.).

These conclusions have been disputed by Donaldson ('08) and by Brodmann ('09), and I am in agreement with these critics that Kaes' results cannot be taken seriously.

Brodmann ('08), in his paper on the cortical measurement, has noted only in a general way the average cortical thickness at the lateral surface of the hemisphere at several ages, as shown in table 10 (columns A and C). Nevertheless, these data can be used for a comparison.

Donaldson ('08) has compared the albino rat with man in respect to the growth of the brain and reached the conclusion that man and the rat show growth curves for the brain which are similar in form when the data are compared at equivalent ages, and the condition of the brain of the rat at five days of age is taken as like that of the human brain at birth. The relative growth rates of the rat and man are as 30 to 1 and the brain of the child at one year corresponds to that of the albino rat at seventeen days of age in its stage of development (Donaldson, MS.). These statements are also confirmed by me for the cortical thickness, as shown in table 10 (see below), and I have already noted that the transitional cortical cell layers, which are no longer to be seen in a new-born child, do not disappear in the albino rat until after four days of age (Sugita, '17 a, p. 539).

From these relations, we conclude that the course of growth in the thickness of the cerebral cortex in man and the albino rat would probably be similar, if the brains were compared at the equivalent ages. Such a comparison is attempted in table 10. Here the increase in cortical thickness in man and in the albino rat is compared, employing data given by Brodmann ('08) and by me (Sugita, '17 a). From the age (column A) given by Brodmann, the approximate brain weight (column B) was de


GROWTH OF THE CEREBRAL CORTEX


263


TABLE 10

Giving a comparison in the course of increase in cortical thickness in ynan and in the albino rat, according to data given by Brodmann {'08) and by Sugita {'17 a). Approximate brain weight in man and in the Albino for the equivalent ages were assumed in round numbers according, respectively, to Vierordt {'90) and Donaldson {'08)


A


B


C


D


E


F


G


MAN


ALBINO RAT








Correspond




Approximate




ing cortical




Observed


brain



Thickness


thickness in



Approximate


thickness of


weight, at the


Equivalent


of the cortex


human brain,


Age


brain


the cortex.


equivalent


(observed)


at ages


when the



weight


Brodmann


(observed)


age


given in


adult values




('08)


ages (Donaldson)



Column E


in the both are taken as the standards



grains


m m .


grams


days


mm.


7nm,


Fetus








8-9 months



1.0-1.5



Birth


0.80


1.25


Birth


380


1.5-2.0


0.50


5


1.10


1.75


1 year


950


2.0-3.0


1.10


17


1.75


2.76


Adult


1400


2.0-4.0


1.90


Adult


1.90


3.00


termined according to Vierordt ('90) and then the final weight (1400 grams) was entered corresponding to the adult brain weight of the albino rat (1.9 grains). The other corresponding brain weights of the Albino of the equivalent ages were entered also according to Donaldson ('08) (column D). The cortical thickness (column F) for the given brain weights of the Albino were then entered according to my former determination (Sugita, '17 a). If the cortical thickness of the adult man be assumed as 3.00 mm. (the mean value of 2.0 to 4.0 mm.) and the corresponding thickness at each age be calculated on the basis of the course of increase in cortical thickness in the Albino (given in column F), the results given in column G — a mere inference, to be sure — are fairly in accord with the figures presented by Brodmann (column C).

In this connection, I had the opportunity, through the courtesy of Dr. W. H. F. Addison, to prepare sections and examine the cortical thickness at the dorsal part of the gyrus centralis anterior (regio Rolandica) from a child thirteen months old (material hardened in 4 per cent formaldehyde, imbedded in paraf


264 NAOKI SUGITA

fine, and stained by Nissl's method). The mean value of the cortical thickness at the summit of the gyrus was 3.55 mm., or within 10 per cent the value obtained by Brodmann at the same locality in the adult brain and on a section similarly prepared and measured (table 8) . So far, then, as this observation goes, it helps to support my conclusion presented earlier that the human cortex has attained nearly its full thickness at the age of fifteen months (Sugita, '17 a).^

VII. THE BRAIN WEIGHT, THE CORTICAL VOLUME, AND THE BODY

WEIGHT

Dhere and Lapicque ('98) and DuBois ('98 a, '98 b), working independently, found several important relations existing between the body and the brain weights in man and a number of other vertebrates. Recently DuBois ('13) has obtained results which he has formulated in following terms:

1) In species of vertebrates that are alike in organization of their nervous system and their shape, but differ in size, and also in the two sexes of one and the same species, the quantity of the brain increases; A) as the quotient of the superficial dimension divided by the cube root of the longitudinal dimension. B) as the product of the longitudinal dunension by the square of its cube root.

2) In individuals of one and the same species and of the same sex, but differing in size, the quantity of brain increases as the square of the cube root of the longitudinal dimension of the body.

So, briefly stated, 1) reads: in any species of vertebrates that are equal in organization, in form of activity and in shape, the weights of the respective brains are proportional to the 0.55 power

' According to a study by Fuchs ('83), the child is born without any myelinated fibers in the cerebral cortex. In the lamina zonalis the first myelination appears at five months, in the lamina pyramidalis at the end of the first year, while in the innermost layers we see some faintly stained fibers at two months. The fibrae arcuate (association fibers) appear clearly at seven months. Later the myelinated fibers increase in caliber and number as the age advances, and at eight years they attain nearly the appearance which they have in the adult cortex.


GROWTH OF THE CEREBRAL CORTEX 265

of the weights of bodies, and 2) the exponent of correlation within the same species is for all vertebrates the 0.22 power.

These relations were based on a series of observations, and this illuminating idea is now generally accepted as true.

The brain in general consists of the white and the gray matter, and in higher animals the gray matter as represented by the cerebral cortex occupies a relatively large part of the entire cerebrum. This cortex is the seat of a complex series of physiological nerve centers, and the possibility that it has definite quantitative relations with the body as a whole is suggested by the following statement made by Du Bois ('13) :

If the quantity of brain does not increase proportionally to the volume of the body, exprassed by the weight, it might be that this is really the case with regard to the superficial dimension, as being proportional with the receptive sensitive surfaces and with the sections of the muscles, thus measuring the passive and active relations of the animal to the outer world, for which in this waj" the quantity of brain can be a measure.

This statement, to be sure, is applied by DuBois to the weight or volume of the entire brain, but if the volume of the cortex stands in some definite relation to the volume of the entire brain, then the cortical volume should be also in a definite relation to the size or weight of the body.

The cortica' volume is determined by the area of surface of the cerebral hemisphere and the thickness of the cortex. The former factor is not easy to determine exactly, even in lissencephala, while in higher animals the hemispheres have many convolutions which increase still further the difficulty of this determination. In lissencephala, the surface area of the hemispheres in two brains, which are nearly similar in the form of cerebrum, are approximately comparable with squares of the corresponding diameters of the cerebra.

The cortical thickness, on the other hand, is not so hard to determine exactly. The average thickness of the cortex in different mammals is given in table 11, quoted from various sources, and, as seen from this table, it is not directly related to the size or weight of the brain, since, as Marburg's ('12) table shows, the


266 NAOKI SUGITA

cortical thickness in several primates ranges within rather narrow limits (2.3 mm. to 2.8 mm.), while the brain weight shows a distinctly wider range (82 grams to 400 grams) (table 11). In some cases indeed the smaller brain has a thicker cortex, even in the same family (e.g., the smaller hapale has a thicker cortex than the larger lemur) . But in general we may conclude with Brodmann ('09) that, within one and the same order or family of mammals, the large brain tends to have a larger average value for the cortical thickness.

The relative cortical volume has been formerly computed by me, employing the formula especially devised for this purpose, in the albino and the Norway rat brains, so that the two forms may be compared directly (Sugita, '18 b). The ratio of the cortical volumes in the adult Albino (brain weight, 2.0 grams) and the Norway (brain weight, 2.3 grams) is 1.31, as the relative cortical volumes are, respectively, 393 and 517 (Sugita, '18 b, table 15), and the ratio of the body surfaces in the two animals amounts also to 1.30, when the body weights of the adult albino and the Norway rats are taken as 300 grams and 450 grams, respectively. Moreover, the ratio of cortical volumes in the two forms at any given age will prove to be almost equal to the ratio of body surfaces of the two at the same age.'*

As above tested, the body weight and the cortical volume of the animals in the same family stand in a definite relation, at least in this instance. But, as we cannot compute the volume of the cortex in other mammals from the data given in table 11, the relation can not be tested further.

■' For example, according to my former presentation (Sugita, '18 b), the computed cortical volume in the Albino Group XV (brain weight, 1.54 grams) is about 346 and that in the Norway Group N XVIII (brain weight, 1.83 grams) is about 423, and according to another determination (Sugita, '18 a) these two groups may be regarded nearly equal in age, as the Albino brain weight would be about 18 per cent less than the Norway brain weight of the like age. The ratio in cortical volume of the above two is 1.22. The body weight corresponding to the brain weight of 1.54 grams in the albino rat is 64 grams and that corresponding to the brain weight of 1.83 grams in the Norway rat is 90 grams ('The Rat,' Donaldson, '15). The ratio of the body surface in the above two, therefore, is about 1.25, quite near to the ratio in cortical volume.


TABLE 11 Giving for several species of mammals the adult body weight and brain weight, the average cortical thickness and the name of author from whom the data for the cortical thickness or for the brain and body weights were cited, arranged in the order of decreasing body weight within each family of inammals. The abbreviations of the names of authors are as follows: B = Brodmann {'09), I — Isenschmid {'11), L = Lewis {'79), M= Marburg, {'12), S = Sugita {'17 a, '18 a, MS.)


OF MAMMALIA


Rodentia


Chiroptera


Marsvipialia


Primates


Prosimiae


Artiodactyla f et Carnivo- \ ra I


Insectivora


NAME OF SRECIES


Simia satyrus (orang-outang).

Hylobates

Cynocephalus hamadryas

Macacus rhesus (macaques). . .

Cercopithecus (long-tailed ape)

Lemur varius

Lemur

Hapale (marmoset)

Microcebus

Ovis musimon (sheep)

Felis domestica (cat)

Erinaceus europaeus (hedgehog)

Talpa europaea (mole)

Lepus cuniculus (rabbit)

Cavia cobaya (guinea-pig)

Mus norvegicus (Norway rat) . Mus norv. albinus (albino rat) Spermophilus citillus (groundsquirrel)

Mus musculus (mouse)

Pteropus edwardsii (vampire

bat)

Vespertilio murinus (bat)

Macropus giganteus (kangaroo)

Didelphys


BODY WEIGHT'


7,350 950 920 356

2,500

2,170

1,800

200

62

23,000 3,000


700 75

2,200 600 450 300

200 20


375 23


5,000 1,100


BR.\IN WEIGHT'


grains

400.0

130.0

142.0

82.0

85.0

28.7

23.0

8.0

1.9


100.0 30.0


3.5 1.3

10.0 4.5 2.5 2.0

2.2 0.4


AVERAGE CORTICAL THICKNESS


7.0 0.3


25.0 5.5


mm. 2.8 2.8 2.3 2.3

2.3 1.6 1.7 2.0 1.5

1.6(2.6)2

1.5(2.6)2


1.8 1.0

2.2 1.9 2.1 1.9

1.8 0.8


1.7 0.4


2.3 1.2


K tS

fa p S <


M

M M M

B M B B B

L L


1 The body and brain weights of some animals were not given by the author who has given the cortical thickness. In such cases the body and brain weights were taken from the list given by Weber ('96).

2 According to Lewis (79), the values given here without brackets were taken from Meynert and show the value measured on the slide and the values given within brackets were obtained by his own observation and represent the natural depth of the cortex.

267


268 NAOKI SUGITA

VIII. SIZE AND GROWTH CHANGES IN SOME NERVE CELLS IN THE

^VIAMMALIAN BRAIN

Albino rat. The results obtained by me regarding the size and the growth changes of the pyramidal cells and of the ganglion cells in the cerebral cortex of the albino rat were summarized in a previous study (Sugita, '18 c). Four of the conclusions are here quoted:

1. The full size of the pyramids in the lamina pyramidalis is cell body 21 x 27 m and nucleus 18 x 20 /x in the fresh condition (on the slide, respectively, 16 x 21 ju and 14xl5yu). The full size of the ganglion cells in the lamina ganglionaris is cell body 27 X 37 M and nucleus 23 x 25 ^ in the fresh condition (on the slide, respectively, 21 x 29 ^ and 18 x 19 m) 2. The cell body and the nucleus of the pyramids attain their maximum size at twenty to thirty days in age. Up to ten days they still retain their fetal morphology. After having passed the maximum size at about twenty-five daj^s, they diminish somewhat in size, but the internal structure differentiates as the age advances.

3. The cell body and the nucleus of the ganglion cells attain nearly their maximum size at ten days, when they remain still in fetal form. After this stage, the size of the cell body still increases slowly but steadily as the age advances, while the nucleus remains nearly unchanged in size throughout life.

4. Taking a general view of the data already presented in this series of studies, it is very interesting to observe that the thickness of the cortex, the total number of the cortical nerve cells, and the size of the cortical cells, all attain nearly their full values at the same age of twenty days; that is, at the weaning time of the albino rat.

For comparison with these results on the cells of the cerebral cortex, there are some observations by Addison ('11) on the postnatal growth of the Purkinje cells in the cerebellar cortex of the albino rat. His material was also obtained from the rat colony at The Wistar Institute and the cerebellum was fixed in Ohlmacher's solution, imbedded in paraffine, and stained with


GROWTH OF THE CEREBRAL CORTEX 269

carbol-thionine and acid fuchsin. A part of his results on the Purkinje cells is here quoted:

The Purkinje cells are easily distinguishable at birth along the inner boundary of the molecular layer by their relatively large size and lightly staining nucleus. These cells measure 12 x 7 m and nuclei 8 X 6.3 At. During the first week, there is great increase in size of both nucleus and cytoplasm. The main bulk of the latter is at the ectal pole and from it several fine processes radiate into the molecular layer. At eight days the cells measure 18 x 12 /x and nuclei 10 x 8 ^ to 12 x 9 IX. At eight to ten days there is definite change in form by the elongation of the cytoplasm of the ectal pole to form the main dendrite, the previously existing fine processes becoming its branches. At the same time all the dendries become arranged in one plane, and this plane is parallel to sections directed across the folia. Nissl granules appear in the cj^oplasm at eight to ten days. The arrangement of Purkinje cells changes with the increase in the surface area of the cortex. At birth they are arranged in two to three irregular rows; at three days in one to two irregular rows, and at five days in one continuous row. As growth of the cortex continues, the space intervening between the Purkinje cells becomes greater. Some nuclei reach their maximum size of 12 x 9 /x at eight days, while the cell bodies usually continue to grow, reaching a maximum size of 24 x 19 ^ at twenty days. The dendrites reach the outer limiting membrane when all the outer granule eel's have migrated (twenty-one to twentyfive days), and continue to develop new branches until a much later period as is .diown by a comparison of cells from a 31 day with cells from a 110-day cerebelhun.

From this it is plain that the Purkinje cells (cell bodies) of the albino rat cerebellum have also reached full size at about the weaning time (twenty days of age) .

From the foregoing, we see that the functional cortical cells both in the cerebrum and in the cerebellum reach their full size at an early age — before the weaning time — and though they continue to mature after that they change only slightly in size, sometimes even diminishing. Thus the cortical nerve elements are all precocious in their growth, which is nearly complete when the young become independent of the mother and their education begins. Addison ('11) has stated also that the development of motor control in the young rat is closely correlated with the completion of the cerebellum and the rat attains its full motor control when the cerebellum has attained structural


270 NAOKI SUGITA

maturity at twenty-one to twenty-five days of age. At that age the cells are nearly full size. We may conclude, therefore, at least regarding some of the nerve cells, that the beginning of functional education of the cells at twenty days is preceded by the attainment of nearly full size, and after this period there is very little change in size, though the internal structures mature as the age advances.

Mouse. A study in this field was made by Stefanowska ('98) on the cortical cells of the mouse. She stained the cells by the method of silver impregnation and studied mainly the development of the cell attachments. Her conclusions may be condensed as follows:

1. In the new-born mouse most of the cortical nerve cells have a simple morphology. 2. The cells are usually arranged in chains, disposed perpendicularly to the surface of the cortex. 3. Besides these, there are some groups of cells more advanced in developmen and having many dendrites, and cells which have the adult form having many, long, ramified dendrites. 4. The different parts of the cortex do not attain the same degree of development at the same time. Some cell groups are more precocious. 5. In the lamina multiformis and in the lamina ganglionaris, we find always the most advanced cells in large numbers. 6. In the lamina pyramidalis the development of the cells is very slow. On the ectal surface, near the pia mater, many cells not at all differentiated are often found. 7. At one day after birth, the dendrites of cortical cells are covered with varicosities. The axis-cylinders have also many nodal swellings. 8, As the neurons develop, the varicosities become more and more rare. At fifteen days, varicosities are no longer seen on the dendrites and the neurons at this age have completed their development. 9. The appearance of the piriform appendices on the dendrites is somewhat delayed. At ten days all pyramidal cells show these appendices. These latter are the constant feature of the neuron, while the varicosities are only a temporary formation. The piriform appendices may be the terminal apparatus of the dendrites. 10. The piriform appendices are the last element which appears on the cortical cells during growth. This fact seems to suggest the high importance of these appendices for this nerve function.

As seen from the foregoing, the morphological completeness in respect of the dendrites and the axis-cylinder of the cortical cells is attained at fifteen days or at the weaning time of the mouse also.


GROWTH OF THE CEREBRAL CORTEX


271


TABLE 12 Giving for man and other mammals the size of the largest ganglioncells in the lamina ganglionaris of the cerebral cortex as presented by different authors. Data are arranged according to the order of the average diameters


NAME OF SPECIES


Homo sapiens (man)

Homo sapiens (man)

Homo sapiens (man)

Homo sapiens (man)

Felis leo (lion)

Felis tigris (tiger)

Cercoleptiis caudivolvulus (kinkajou).

Ursus syriacus (bear)

Indris (babakoto)

Felis domestica (cat)

Cercopithecus mona (African monkey)

Elephas (elephant)

Lemur

Mus norvegicus (Norway rat)

Ovis musimon (sheep)

Sus (pig)

Mus norvegicus albinus (albino rat) . .

Lepus cuniculus (rabbit)

Lepus cuniculus (rabbit)

Pteropus edwardsii (vampire bat)

Mus musculus (mouse)


MAXIMUM SIZE


REPORTED IN MICBA


Linear diameters


Average

diameter or

square root

of the



product


60X120


85


55X126


83


53X106


75


40 X 80


57


60X133


90


60X100


78


50X110


74


53X100


73


44 X 80


59


32X106


58


40 X 72


54


35 X 60


46


SOX 70


46


33 X 48


40


23 X 65


39


27 X 48


36


30 X 42


36


18 X 60


33


18X 40


27


16X 36


24


18X 20


19


Author


Betz Lewis Brodmann Hammarberg

Brodmann

Brodmann

Brodmann

Brodmann

Brodmann

Lewis

Brodmann

Brodmann

Brodmann

Sugita

Lewis

Lewis

Sugita

Lewis

Brodmann

Brodmann

Isenschmid


There are no other systematic investigations on the postnatal development of the cortical nerve cells in mammals, although there are some studies on the growth of nerve cells in the fetus, among which the researches by His ('04) (see footnote 2), Koelliker ('96), and Vignal ('89) are the most important.

Table 12 was compiled by me in order to compare the size of the largest ganglion cells in the lamina ganglionaris (the fifth layer of Brodmann) of the cerebral cortex of man and some other mammals. The tabulated data were taken from Brodmann ('09), Lewis ('79, '82), Hammarberg ('95), and others.


272 NAOKI SUGITA

The results obtained by me (Sugita, '18 c) in the albino and the Norway rats have been also entered.

IX. THE SIZE OF THE LARGEST CORTICAL CELLS IN MAN AND SOME OTHER MAMMALS

From table 12 we can draw only very general conclusions as to the significance of the size of the largest cortical cells. The giant Betz cells even in man vary rather widely in size according to the different authors, probably owing largely to the different technical methods used, as has been pointed out repeatedly in the course of this paper.

From time to time attempts have been made to formulate a general interpretation of the size of the Betz cells and of the nerve cells in general. From the examination of table 12, it is seen that the values for the mean diameters do not, except in the very most general way, follow the size of the animal, but that the Felidae, even the cat, stand high in the series.

We are not able to contribute any general explanation for the size of these cells, although it may not be out of place to repeat that in the Norway rat with the heavier brain these cells are larger than in the albino rat with the lighter brain (Sugita, '18 c), and so will merely call attention to the various authors who have had something to say in the matter: Lewis ("79), Hughlings Jackson ('90), Schwalbe ('81), Barratt ('01), Dunn ('00, '02), Herrick ('02), Donaldson ('03), Campbell ('05), Boughton ('06), Johnston ('08), and Kidd ('15).

X. SUMMARY

1. In the present paper I have attempted to compare my conclusions regarding the development of the cortical elements in the brains of the albino and the Norway rats with the corresponding changes in other mammals. The data used for these comparisons were taken from various sources, but the comparisons are in many instances hampered by differences in technique or the lack of essential information.

2. The relations of the cortical thickness at different locali


GROWTH OF THE CEREBRAL CORTEX 273'

ties in the cerebrum are quite the same in the mouse and rabbit as in the rat. The development of the cortical thickness has proved to be similar in the mouse and guinea-pig: it attains nearly its full value at the weaning time of the animal.

3. The statement that the cortical thickness diminishes from the frontal to the occipital pole and from the dorsal to the ventral aspect probably holds true throughout mammals, including man.

4. The results given by different authors for the cortical thickness of human brain (averages or for each locality) are by no means in accord. Even for the same locality there are wide deviations. The best data indicate that the average cortical thickness of the adult human brain is about 3 mm.

5. The mode of increase in cortical thickness in man according to age appears to be similar to that in the albino rat, if the brains are compared at equivalent ages. The developmental stage of the brain of a new-born child corresponds to that of an albino rat of five days of age, and throughout the postnatal life the relative growth rate of the rat and man are as 30 to 1. The span of life 30 for man corresponds to 1 for the rat and the equivalent ages are represented by like fractions of the span of life. The human cortex probably attains nearly its full thickness at fifteen months, equivalent to twenty days of rat age.

6. The relative cortical volumes of the albino and the Norway rat brains, computed formerly by me (Sugita, '18 b), appear to be proportional to the surface areas of the entire bodies at the like age. This relation may be generally applicable within a given order of mammalia. The cortical thickness or the brain weight is in general only loosely correlated with the body weight or size of the animal.

7. The cortical nerve cells in the cerebruni and in the cerebellum of the albino rat are precocious in their growth, attaining almost the full size at twenty days, the weaning time. The maturation of the intracellular structures probably continues after the size is apparently completed. This process is shown also in the mouse.

8. The size of the Betz giant cells in the adult human cortex

THE JOURNAL OF COMPABATIVE NEUBOLOGT, VOL. 29, NO. 3


274 NAOKI SUGITA

(found ill the gyrus centralis anterior) is reported differently by different authors. The mean value is about 75 micra in average diameter.

9. The size of the cortical cells, especially the Betz motor ganglion cells, of adult animals has no clear relationship to brain size or body size. These cells are notably large in the Felidae.

10. As a general conclusion to this series of studies the following statement may be made:

The morphological organization of the cerebral cortex is generally precocious. The size of individual cortical nerve cells, the total number of cortical cells, and the thickness of the cortex, all attain nearly their full values at the same time and very early in life (corresponding to the weaning time in some rodents) , after which the maturation of internal structures of the cell body and the nucleus continues. The brain weight and the cortical volume continue to increase even after this stage throughout the postnatal life, though not so rapidly as during the early period. This later growth is due principally to the development of the cell attachments, intercellular tissues (neuroglia tissue and bloodvessels), the ingrowth of axons into the cortex and their myelination, which together separate the cells from each other, and cause an increase in cortical volume. The cortical volume is primarily dependent on the size of individual cortical cells and their total number and it appears in animals belonging to a given zoological order to have a definite relationship to the size (or area of surface) of the body of the animal.


GROWTH OF THE CEREBRAL CORTEX 275

LITERATURE CITED

Addison, VV. H. F. 1911 The development of the Purkinje cells and of the cortical layers in the cerebellum of the albino rat. Jour. Comp. Neur., vol. 21, no. 5.

Allen, Ezra 1912 The cessation of mitosis in the central nervous system of the albino rat. Jour. Comp. Neur., vol. 22, no. 6.

Allen, Jessie Blount 1904 The associative processes of the guinea-pig. A study of the psychical development of an animal with a nervous system well medullated at birth. Jour. Comp. Neur. and Psychol., vol. 14, no. 4.

Barratt, J. O. Wakelin 1901 Observations on the structure of the third, fourth, and sixth cranial nerves. Jour. Anat. and Physiol., vol. 35, p. 214.

BouGHTON, T. H. 1906 The increase in the number and size of the medullated fibers in the oculomotor nerve of the white rat and of the cat at different ages. Jour. Comp. Neur. and Psychol., vol. 16, pp. 153-165.

Brodmann, K. 1908 Uber Rindenmessungen. Centralbl. f. Nervenheilkunde u. Psychiatrie, Bd. 19.

1909 Vergleichende Lokalisationslehre der Grosshirnrinde. Leipzig. 1909 Antwort an Herrn Dr. Th. Kaes. tJber Rindenmessungen. Neurolog. Centralbl., Jahrgang 28, p. 635.

Campbell, A. W. 1905 Histological studies on the localisation of cortical function. Cambridge.

Dhere and Lapicque, Louis 1898 8ur le rapport entre la grandeur du corps et le developpement de I'encephale. Archives de Physiologie normale et pathologique, no. 4.

Donaldson, H. H. 1891 Cerebral localization. Am. Jour, of Psychol., vol. 4, no. 1.

1891 Anatomical observations on the brain and several sense-organs of the blind deaf-mute, Laura Dewey Bridgeman. II. On the thickness and structure of the cerebral cortex. Am. Jour, of Psychol., vol. 4, no. 2.

1897 The growth of the brain. New York.

1903 On a law determining the number of medullated nerve fibers innervating the thigh, shank, and foot of the frog— Rana virescens. Jour. Comp. Neur., vol. 13, no. 3.

1908 Review "Die Grosshirnrinde des Menschen" von Dr. Th. Kaes. Am. Jour. Anat., vol. 7, no. 4. Anat. Rec, no. 8. 1908 A comparison of the albino rat with man in respect to the growth of the brain and of the spinal cord. Jour. Comp. Neur., vol. 18, no. 4. 1915 The Rat. Memoirs of The Wistar Institute of Anatomy and Biology, no. 6. Dubois, Eugene 1898 Uber die Abhangigkeit des Hirngewichtes von der Korpergrosse bei den Saugetieren. Archiv f. Anthropologic, Bd. 25.

1898 tJber die Abhangigkeit des Hirngewichtes von der Korpergrosse beim Menschen. Archiv f. Anthropologie, Bd. 25.


276 NAOKI SUGITA

Dubois, EuGE^fE 1913 On the relation between the quantity of brain and the size of the body in vertebrates. Proceedings of the meeting of December 27, 1913. Koninklijke Akademie van Wetenschappen te Amsterdam, vol. 16.

Dunn, Elizabeth Hopkins 1900 The number and size of the nerve fibers innervating the skin and muscles of the thigh in the frog (Rana virescens brachycephala, Cope). Jour. Comp. Neur., vol. 10, no. 2. 1902 On the number and on the relation between diameter and distribution of the nerve fibers innervating the leg of the frog, Rana virescens brachycephala. Cope. Jour. Comp. Neur., vol. 12, no. 4.

FucHS, SiGMUND 1883 Zur Histogenese der menschlichen Grosshirnrinde. Sitzungsber. der K. Akad. der Wissenschaft, Wien., Bd. 88. III. Abtheil.

His, Wilhelm 1904 Die Entwickelung des menschlichen Gehirns wahrend der ersten Monate. Leipzig.

Hammarberg, Carl 1895 Studien liber Klinik und Pathologic der Idiotie nebst Untersuchungen tiber die normale Anatomie der Hirnrinde. Upsala.

Herrick, C. Judson 1902 A note on the significance of the size of nerve fibers in fishes. Jour. Comp. Neur., vol. 12.

Isenschmid, Robert 1911 Zur Kenntnis der Grosshirnrinde der Maus. Abh. Akad. Wiss. Berlin, physik-math. CI. Jahrg. 1911 Anh. no. 3.

Jackson, J. Huglings 1890 On convulsive seizures. British Medical Journal, vol. 1.

Johnston, J. B. 1908 On the significance of the caliber of the parts of the neurone in vertebrates. Jour. Comp. Neur. and Psychol., vol. 18, no. 6.

Kaes, Theodor 1905 Die Rindenbreite als wesentlicher Faktor zur Beurtheilung der Entwickelung des Gehirns und namentlich der Intelligenz. Neurolog. Centralbl., Jahrgang 24, Nr. 22. 1907 Die Grosshirnrinde des Menschen in ihren Massen und in ihren Fasergehalt. 2 volumes. Jena.

1909 tjber Rindenmessungen. Eine Erwiederung an Dr. K. Brodmann. Neurolog. Centralbl., Jahrgang 28, p. 178. 1909 Replik. Zu "Dr. Brodmanns Antwort an Rindenmessungen." Neurolog. Centralbl., Jahrgang 28, p. 639.

KiDD, Leonard J. 1915 Factors which determine the calibre of nerve cells and fibres. Review of Neurology and Psychiatry, vol. 13, pp. 1-27.

King, Helen Dean 1910 The effects of various fixatives on the brain of the albino rat, with an account of a method of preparing this material or a study of the cells in the cortex. Anat. Rec, vol. 4, pp. 214-244.

Lapicque, Louis 1907 Tableau gen6ra' du poids encephalique en ionction du poids du corps. Paris.

Lewis, W. Bevan 1878 Application of freezing methods to the microscopic examination of the brain. 'Brain,' Part 3, pp. 348-359. 1879 Re^arches on the comparative structure of the cortex cerebri. III. Phil. Trans., pp. 36-64.

1882 On the comparative structure of the brain in rodents. Phil. Trans., pp. 699-749.


GROWTH OF THE CEREBRAL CORTEX 277

Marburg, Otto 1907 Beitrage zui Kenntniss der Grosshirnrinde der Affen.

Arbeiten aus dem Neurologischen Institute an der Wiener UniversitJit

(Obersteiner). Bd. 16. Mayer, Otto 1912 Mikrometrische Untersuchungen iiberdie Zelldichtigkeit

der Grosshirnrinde bei den Affen. Jour. f. Psychol, u. Neurol., Bd.

19, Heft 6. Rose, M. 1912 Histologische Lokalisation der Grosshirnrinde bei kleinen

Saugetieren (Rodentia, Insectivora, Cheiroptera). Jour. f. Psychol.

u. Neurol., Bd. 19, Ergonzungshefte 2. ScHWALBE 1881 Lehrbuch der Neurologie. Erlangen. Stefanowska, Micheline 1898 Evolution des cellules nerveuses corticales

chez la souris apres la naissance. Annales de la Societe Royale des

Sciences med. et naturelles de Bruxelles, vo . 7. Sugita, Naoki 1917 Comparative studies on the growth of the cerebral cortex. I. On the changes in the size and shape of the cerebrum during

the postnatal growth of the brain. Albino rat. Jour. Comp. Neur.,

vol. 28, no. 3.

1917 a Comparative studies on the growth of the cerebral cortex.

II. On the increase in the thickness of the cerebral cortex during the postnatal growth of the brain. Albino rat. Jour. Comp. Neur., vol. 28, no. 3.

1918 Comparative studies on the growth of the cerebral cortex.

III. On the size and shape of the cerebrum in the Norway rat (Mus norvegicus) and a comparison of these with the corresponding characters in the albino rat. Jour. Comp. Neur., vol. 29, no. 1.

1918 a Comparative studies on the growth of the cerebral cortex.

IV. On the thickness of the cerebral cortex of the Norway rat (Mus norvegicus) and a comparison of the same with the cortical thickness in the Albino. Jour. Comp. Neur., vol. 29, no. 1.

1918 b Comparative studies on the growth of the cerebral cortex.

V. Part I. On the area of the cortex and on the number of cells in a unit volume, measured on the frontal and sagittal sections of the albino rat brain, together with the changes in these characters according to the growth of the brain. Part II. On the area of the cortex and on the number of cells in a unit volume, measured on the frontal and sagittal sections of the brain of the Norway rat (Mus norvegicus), compared with the corresponding data for the albino rat. Jour. Comp. Neur., vol. 29, no. 2.

1918 c Comparative studies on the growth of the cerebral cortex.

VI. Parti. On the increase in size and on the developmental changes of some nerve cells in the cerebral cortex of the albino rat during the growth of the brain. Part II. On the increase in size of some nerve cells in the cerebral cortex of the Norway rat (Mus norvegicus), compared with the corresponding changes in the albino rat. Jour. Comp.

. Neur., vol. 29, no. 2.

1918 d Comparative studies on the growth of the cerebral cortex.

VII. On the influence of starvation at an early age upon the development of the cerebral cortex. Albino rat. Jour. Comp. Neur., vol. 29, no. 3.


278 NAOKI SUGITA

ViERORDT, H. 1890 Das Massenwachstum der Korperorgane des Menschen.

Archiv f. Anatomie u. Physiologic, Anat. Abtheil., pp. 62-94. ViGNAL, William 1889 Developpement des elements du systeme nerveux

cerebro-spinal. Paris. De Vries, I. 1912 tjber die Zytoarchitektonik der Grosshirnrinde der Maus

und iiber die Beziehungen der einzelnen Zellschichten zum Corpus

Callosum auf Grund von experimentellen Ltisionen. Folia Neuro Biolog ca, Bd. 6, Nr. 4. Weber, Max 1896 Vorstudien iiber das Hirngew'cht der Saugetiere. Fest schr It iir Carl Gegenbaur. Pp. 105-12].


AUIHORS'S ABSTRACT OF THIS PAPER ISSUED BY THE BIBLIOGR.tPHIC SERVICE, .\PR1L 20


THE PERIPHERAL TERMINATIONS OF THE NERVUS LATERALIS IN SQUALUS SUCKLII

SYDNEY E. JOHNSON

From the Anatomical Laboratory of Northivestern University Medical School^

TEN FIGURES

The observations set forth below supplement the writer's previous paper on the structure and development of the lateral canal sense organs of Squalus acanthias and Mustelus canis.* In the investigation referred to the peripheral terminations of the lateral nerve were demonstrated in Mustelus canis, but not in Squalus acanthias, as fresh specimens of the latter species were unobtainable at that time. Last summer (July, '17), while at the Puget Sound Biological Station, I secured a number of living specimens of the Pacific coast dogfish, Squalus sucklii, which appears to be practically identical to the Atlantic form, Squalus acanthias. The histological structure of the lateral sense organs of these specimens was examined and the peripheral terminations of the lateral nerve were demonstrated by the pyridine silver method and also with methylene blue. These observations supply the omission which was necessitated in the paper referred to above.

The papers which deal specifically with the peripheral terminations of the nervus lateralis and which are of more than historic value are those of Retzius '92, v. Lenhossek '92, Bunker '97, Heilig '12, and Pfliller '14. They are discussed briefly in the writer's previous paper and need no further comment except to say that most attempts to stain the peripheral terminations of the lateral nerve have heretofore yielded rather meagre results.

1 Contribution No. 60.

-Jour. Comp. Neur., Vol. 28, No. 1.

279


280 SYDNEY E. JOHNSON

In comparing the lateral sensory canals of Mustelus canis and Squalls sucklii there are a number of differences to be noted. Perhaps the most striking is the difference in calibre of the sensory tubes. The sensory tubes (or canals) of Squalus are much smaller than would be found in a Mustelus specimen of the same size. The column of sensory epithelium is proportionately narrower in Squalus. A slight but apparently constant difference in the course of the lateral canals of the two species is seen in the slight elevation of the canal above the anal fin in Mustelus. There are other differences in the distribution of the canals, but they are less striking and have not been carefully examined. The lateral canals of both species lie chiefly in the dermis and their tubules pass directly ventrad for a short distance before making a sharp bend laterally to open on the surface of the integument. The surface tubules correspond in number with the ramuli of the lateral nerve and there are approximately five tubules for every four segments of the vertebral column.

The lateral nerve lies at a considerable depth from the sensory canal, especially in the anterior region, and its ramuli pass obliquely to the basilar membrane of the sensory column, where their fibers diverge caudad and cephalad to form a continuous longitudinal fiber zone just outside of the basilar membrane. This fiber zone differs from that described for Mustelus only in the fact that it contains a considerably smaller number of nerv^e fibers.

The sensory epithelium of Squalus sucklii differs considerably from that of Mustelus canis. It is much less extensive and the sensory cells are aggregated in smaller groups. This can be seen readily in transverse and longitudinal sections. In the former one to three sensory cells can ordinarily be seen in the cell clusters (fig. 1), and in the latter, usually three to six (figs. 2 and 10). The groups of sensory cells are somewhat more widely separated from each other than they are in Mustelus, and the sensory column appears to show a stronger tendency towards segmentation. This apparent segmentation of the column of sensory epithelium, however, bears no relationship to the normal body segments for there are usually more than ten


LATERAL SENSE ORGANS OF SQUALUS SUCKLII


281


clusters of these cells between adjacent surface tubules, and the tubules, in turn, are more numerous than the segments of the vertebral column. Nor is there any marked regularity in the number and size of the individual clusters of sensory cells. While the sensory column is thus essentially continuous throughout the entire length of the sensory canal it shows considerable



S71.G0L


FkZn


Fig. 1 Transverse section of the entire sensory canal of a Squalus sucklii garter. Camera lucida sketch. Iron haem. tech. X 432, | off. Can., canal wall; F6.Zn., longitudinal fiberzone; Sn.CL, secondary sensory cell; Sn. Col., sensory column; Spn., spindle cells.

variation in thickness. It becomes gradually thinner posteriorly and, as in Mustelus, it is usually thinner between adjacent ramuli of the lateral nerve. The base of the column of sensory cells is limited by a continuous basilar membrane.

The same types of cells can be distinguished in the lateral sensory epithelium of Squalus sucklii as were found in the sensory


282


SYDNEY E. JOHNSON


column of Mustelus and of Squalus acanthias. The hair cells or secondary sense cells are large, pear-shaped, and have centrally placed nuclei. In many specimens hair-like processes could be seen at their distal ends, but whether one or more for each cell has not been determined. The relative length of the cells is usually one-half to two-thirds the thickness of the sensory




T

fbZn



-Rml


Fig. 2 Longitudinal section o the lateral sensory column (Sii. Col.) of Squalus sucklii (adult). The sensory epithelium was drawn with the aid of a camera lucida from an iron haematoxylin preparation, and the nerve fibers were put in free hand from pyridine silver sections. The outlines of the canal wall {Can.) and the surface tubule {Tub.) are not drawn to scale but are greatly reduced in order to conserve space. For correct proportions, see figure 1. Sensory column, X 650, J off. Fhr., terminal fibrillae; Fb.Zn., longitudinal fiber zone; Gr-p., one group of secondary sensory cells (hair cells).

column. Spindle-shaped cells, basilar cells, and columnar cells constitute the supporting elements (see figs. 1, 2, 3 and 10). The rest of the canal wall is formed by a double layer of epithelial cells, both layers of which are continuous with the walls of the surface tubules and also with the columnar and stratified layers of the epidermis.


LATERAL SENSE ORGANS OF SQUALUS SUCKLII 283

The peripheral terminations of the lateral nerve. On reaching the base of the sensory column the fibers of the lateral ramuli diverge caudad and cephalad in the subbasilar fiber zone. This fiber zone is shown in longitudinal section in figures 2 and 4, and in transverse section in figures 1 and 3. The majority of the fibers are medullated but a few non-medullated fibers can be found. These can be traced back through the ramulus to the lateral nerve, which indicates that they are not simply nonmedullated branches of the large medullated fibers.

Two zones of distribution or branching of the nerve fibers appear well marked. Primary distribution takes place from the longitudinal fiber zone and the branching is almost entirely subbasilar (figs. 4 and 10), while a secondary zone of distribution or branching is located roughly between the limits marked by the .nuclei of the basilar cells and the proximal ends of the hair cells. It is from this zone that the fine fibrillae arise which pass out freely between the hair cells.

The primary branches are large and coarse as a rule (fig. 7), although many fine branches arise from this zone also (fig. 4). Branching of the fibers appears frequently to be dichotomous but not uncommonly three or more branches arise at the same level. This statement holds for both zones of distribution. Enlargements of considerable size are commonly seen at the level of branching of the nerve fibers (fig. 9), but it seems likely that the majority of these extra large varicosities" are caused by an over-deposit of silver at the points of branching. One or more fibers may rise from the subbasilar fiber zone to supply a single cluster of hair cells, and occasionally the fibrillae of a given fiber ramify in adjacent groups of hair cells (fig. 10). The medullary sheath is usually lost just outside of the basilar membrane.

The primary branches rise to a considerable height in the sensory epithelium — usually beyond the nuclei of the basal cells — where they form a rather rich plexiform network (figs. 4, 7, and 10) . This network forms the secondary zone of distribution and it is from it that the ultimate distribution of fibrillae to the hair cells takes place. While this secondary zone of distribution is present in the lateral sensory epithelium of Mustelus canis, it is


Grp.


. \ -' 1^ v"^- *^H T


hT^ r


%




K


Grf>,




Grp. ^%




/I6^.^i^.



/=Z>.Z;7.



'i-C^ ZJ^S^^^"


284


LATERAL SENSE ORGANS OF SQUALUS SUCKLII 285

not as uniformly developed and is much less conspicuous than it is in Squalus sucklii.

The fine fibrillae which arise from the secondary zone of distribution rise to various levels in the sensory epithelium. In many instances they can be traced to within a short distance of the outside limiting membrane (figs. 8 and 9). Varicosities of various sizes and shapes appear on the fibrillae at practically all levels and not infrequently at their distal extremities. In many cases the fibrillae appear to surround the bases of the hair cells (figs. 8 and 9), and in others, to pass out freely and separately between the hair cells.

The observations set forth above corroborate the results obtained on Mustelus canis. Only minor differences exist in the structure and innervation of the sensory epithelium of the two species. In Squalus sucklii the sensory epithelium is less extensive, there is a stronger suggestion of segmentation, and in nerve supply there is a more definite and conspicuously secondary zone of distribution.

A number of features which stand out in the embryonic and adult structure of the lateral canal system of Squalus and Mustelus appear to me to reflect doubt on the theory that this sytem of sense organs has a phylogenetic relationship with the segmental sense organs of certain invertebrates and that the system itself is segmental in the sense suggested by John Beard^ and W. H. Gaskell.^ The evidence, in part, against such a view may be

Fig. 3 Transverse section of the sensory column, showing the peculiar condition of two groups of hair cells (Grp.) existing side by side. Camera sketch, X 650. Nf., nerve fibers of the subbasilar fiber zone.

Fig. Longitudinal section of the lateral sensory column and the subbasilar fiber zone (Fb.Zn.). The secondary zone of distribution (Snd.Zn.) is also shown. Camera sketch. Pyridine silver tech. X 650, f off. Grp., group or cluster of hair cells; N.M.Fb., non-meduUated nerve fibers.

Fig. 5 Transverse section of the sensory column showing large fibers, and fibrillae diverging at a large varicosity. Pyridine silver tech. X 1525, | off.

Fig. 6 Transverse section of sensory epithelium showing long, fine fibers, and varicosities. Pyridine silver. X 650, j off.

3 See Zool. Anz., Bd. 7, 1884, p. 125 et seq., and also Bd. 8.

4 The Origin of Vertebrates, 1908.


286


SYDNEY E. JOHNSON


A-'nj-r



STfJZn.



VAn


l/ar.



Grp.




B.(ll.


MMSh.


LATERAL SENSE ORGANS OF SQUALUS SUCKLII 287

summarized briefly. The lateral sense organs do not develop in situ from successive or segmental patches of ectoderm along the side of the body, but each lateral sensory column arises from a thickened area of ectoderm located on the side of the head; this invades the posterior segments of the body not as a segmental structure, but in the form of a continuous column of epithelial cells. The grouping of the sensory cells in small clusters occurs comparatively late in the development of the embryo. It has been pointed out that these groups, when they do appear, are not segmental in the sense of the term as here employed. It is only in a degenerating or breaking down condition of the sensory ridges that isolated groups (pit-organs) of hair cells are found (e.g., dorsal series of sense organs in Squalus acanthias). These so called pit-organs show no relationship to the body segments either in their number or in their innervation. Further, their early development is identical with that of the lateral sense organs, the separated organs simply representing parts of what was earlier a continuous ridge of epithelium. So much for the developmental aspect.

The opinion has already been expressed that the slight tendency towards segmentation as seen in the lateral sensory column of the adult is probably of no significance as an argument for the segmentation theory. This is one of the anatomical features, however, which might be considered as pointing in that direction. Another one is seen in the innervation of the sensory epithelium by separate and successive ramuli (of the lateral nerve) which correspond in number and level with the surface tubules. The first condition named loses segmental significance when one remembers

Fig. 7 Longitudinal section of lateral sensory epithelium showing the extensive branching of a single large nerve fiber. Pyridine silver. X 1525, f off.

Fig. 8 Longitudinal section of a group of hair cells, showing various relations of the terminal fibrillae. Pyridine silver. X 650, J off. Var., varicosity.

Fig. 9 Section showing several slender fibrillae diverging from a large varicosity (Var.). Pyridine silver. X 650, I off.

Fig. 10 Longitudinal section of the lateral sensory column, showing two groups of hair cells (Grp.), and a network of fibers arising from the subbasilar fiber zone (Fb.Zn.). Pyridine silver. X 650, i off. B.CL, basal cell; N.M.Fb., non-meduUated nerve fibers.


288 SYDNEY E. JOHNSON

that there are from fifteen to twenty clusters of hair cells for every vertebral segment. Evidence based on the arrangement of the lateral ramuli and the surface tubules is unsatisfactory partly for the same reason and partly for other reasons. As shown above, the lateral ramuli and the surface tubules are considerably nore numerous than the vertebral segments and a constant ratio between the number of vertebrae and ramuli of the lateral nerve is wanting. Furthermore, these ramuli are merely the branches of distribution of a cranial nerve which differs from other cranial nerves only because of the fact that it supplies this remarkable type of sense organ and extends from the head to the caudal fin. In this connection it must be remembered that the fibers of the ramuli diverge at the ba§e of the sensory epithelium to form a continuous fiber zone from which the ultimate distribution takes place.

Further difficulty is met in attempting to relate the numerous organs of the head canals and of the cross-commissures to a corresponding number of ancestral segments.

In view of these considerations it seems improbable to me that the organs of the sensory canals have a phylogenetic history which would relate them either to the segmental sense organs of certain invertebrates, as claimed by Beard, Gaskell, and others, or to the posterior (body) segments of primitive vertebrates. To assume that the lateral sense organs have had such a past history involves the necessity of explaining why the innervation of the body organs should change from a segmental spinal nerve supply to a cranial nerve supply, and also, why the organs do not arise in situ on each segment of the body rather than from cephalic ectoderm which invades the posterior segments and carries with it its own nerve supply, probably from a corresponding primitive cephalic segment. It appears to me more likely that if the lateral sensory apparatus is segmental it is so only in relation to a limited number of cephalic segments. The several lines of organs, then, would represent simply an invasion or extension of a primitive cephalic sensory apparatus into other segments of the body.

Clearly the evidence at hand is not sufficient to warrant dogmatic statements or conclusions. The need is emphasized


LATERAL SENSE ORGANS OF SQUALUS SUCKLII 289

for further histological and embryological work, to be conducted on a comparative basis. The amphibia, especially, need further investigation along this line.

LITERATURE CITED

Bunker, F. S. 1897 On the structure of the sensory organs of the hxteral line

of Arneiurus nel)ulosus. Anat. Anz., Bd. 1.3. IIeilig, Karl 1912 ZurKenntnisderSeitenorgane von Fischen und Ampliihien.

Arch, fiir Anat. und Physiol. Lenhossek, M. v. 1892 Der feinere Bau und die Nervenendigungen der

Geschmacksknospen. Anat. Anz., Bd. 8. Pfuller, Albert 1914 Beitriige zur Kenntnis der Seitensinnesorgane und

Kopfanatomie der Macruriden. Jen. Zeitschr., Bd. 52. Retzius, G. 1892 Ueber die peripherische Endigungsweise des Gehornerven.

Biol. Unters., Bd. 1.


THE JOURNAL OF COMPARATIVE NECROLOOY, VOL. 29, NO. 3


author's abstract of this paper issued

BT THB bibliographic SERVICE, JUNE 1


ON THE DEVELOPMENT OF THE NERVE ENDORGANS IN THE EAR OF TRIGONOCEPHALUS JAPONICUS

TOKUYASU KUDO Anatomical Institute, Medical High School, Niigata, Echigo, Japan

ONE PLATE

The endorgans of the auditory nerve in reptiles have been investigated morphologically with considerable thoroughness. Many authors have interested themselves particularly in the macula neglecta (described for the Amphibia by Deiters in 1862 and given the name now in common use by Retzius) and this endorgan has been studied in various vertebrates, especially in the fishes, the Sauropsida, the mammals and even in man.

Relatively few embryological investigations, however, have been published on this subject. Concerning the genesis of the macula neglecta, Retzius and Alexander concluded that this organ originates from the crista acustica posterior, the former basing his opinion on its comparative anatomy and the latter on observations of its innervation. In Hertwig's Handbuch Krause briefly states that a small region of common neuroepithelium differentiates upon the separation of the saccular from the utricular portions. Fleissig, who, working on reptiles (Gecko), was the first to investigate extensively the development of the macula neglecta, disagrees with both of these statements and is of the opinion that the organ arises from the macula sacculi. The same conclusion is reached by Okagima in the case of Hynobius; but this author remarks that because in the Amphibia the macula neglecta lies within the sacculus, its origin in these forms is easier to determine than in the reptiles, where the macula is found in the utriculus. Corroboration of this view, according to which the macula neglecta arises from the neuroepithelium of the pars inferior, is found in Okagima's study of the salmon embryo and Wenig's recent work on Pelobates fuscus.

291


292 TOKUYASU KUDO

This simple interpretation of the genesis of the macula neglecta has been considerably complicated by the studies of P. and F. Sarasin, who claim to have found a second endorgan in the Caecillidae, for they distinguish two different maculae, one of which lies in a small evagination of the sacculus (macula neglecta of Retzius), the other in the floor of the utriculus (macula neglecta fundi utriculi). The existence of the latter was, however, denied by Retzius, in which opinion he is joined by Ayers. Retzius states: "Es geht nicht hervor, dass die am Boden des Utriculus der Caeciliiden gefundene Nervendstelle einer neu entdeckten Nervendstelle entspricht. Denn gerade am Boden des Utriculus liegt die von mir bei vielen Fischen, Reptilien und Vogeln entdeckte Nervendstelle, Welche von mir schon langst 'Macula neglecta ' genannt wurde. Es ist deshalb ganz unrichtig, wenn die Herren Sarasin die von ihnen bei Ichthyophis am Boden des Utriculis beschriebene Nervendstelle als von ihnen neu entdeckt bet achten und sie als eine 'Macula fundi utriculi' auffiihren. Die echte 'Macula neglecta' hegt am Boden des Utriculus oder Offnung des Canalis^utriculo-saccularis, oder auch-nach meiner Ansicht — bei den niederen Amphibien in der eigentiimlichen Ausstiilpung dieses Canalis, welches ich 'Pars neglecta' gennannt habe, bei den hoheren aber in einer von ihm abgetrennten Ausstiilpung der Sacculuswand." He adds that it would be interesting to know whether both of the endorgans as described by P. and F. Sarasin really do occur, in view of the fact that in all Amphibia that have been thoroughly studied a single macula neglecta occurs. Ayers contends that the new endorgan of the Sarasins is probably none other than the macula neglecta of Retzius. But Fleissig, from his study on the development of the labyrinth in Gecko, was able to demonstrate a transitional condition between the two described above. According to this author the macula neglecta of Retzius is to be regarded as a persisting organ in the sinus inferior, while only traces of the macula neglecta of the Sarasins occurs in adult individuals; and these traces may well be regarded as vestiges of Sarasin's macula, which is present as a developing organ only at a certain stage.


NERVE ENDORGANS IN THE EAR 293

To this much mooted and interesting question, then, I wish to contribute the modest results which I have been able to obtain from my study of Trigonocephalus japonicus.

The viperid embryos which were placed at my disposal comprise more than 27 stages^ (Suzuki-Okajimas series), of which I have employed four for the present study. The embryos were fixed in formol-alcohol, potassium bichromate-acetic and corrosive sublimate-acetic and were stored in alcohol until stained and imbedded in paraffin. Mainly frontal sections 10-15/* thick were made through the heads of the embryos. These were stained in toto with alcoholic borax carmine and Weigert's iron haematoxylin, and in the latter case orange G was employed as a counter stain. The two adult specimens were fixed in potassium bichromate, imbedded in celoidin and cut vertically through the head. These sections, 30/^ thick, were stained in haematoxylineosin and orange G.

Stage 1 (fig. 1). The embryo is coiled up in 4}-^ turns. Olfactory pit very deep. Wall of optic cup thickened anteriorly; lens solid. Fixation: corrosive-acetic. Stain: Weigert's iron haematoxylin ; sections 15^. Frontal sections of head and body.

The auditory vesicle, which is distended into a sac-like structure, is already oval in shape and, since it runs through 47 sections, is about 0.705 mm. in antero-posterior diameter. It lies some distance removed from the brain. Differentiation in the epithelial lining of the wall of the auditory vesicle is already apparent. Laterally the epithelium is flattened, while the medial and lower walls are stratified several cells deep and show here and there a mitotic figure. This thickened portion represents the common neuroepithelium which will later separate into the pars superior and the pars inferior. The ductus endolymphaticus is already tubular in form, with the dilated saccus endolymphaticus at the end.

Stage 2 (fig. 2) The embryo consists of 33^ coils. The parietal elevation is prominent. The lens is approximately as in the preceding stage; the retina moderately pigmented. The

^ The number includes 7 sectioned by the writer.


294 TOKUYASU KUDO

pocket-shaped olfactory pit is deep and the oral sinus deeply cleft. Fixation: formol-alcohol. Stain: alcoholic borax carmine. Sections 15^ in thickness cut frontally through head and entire body. The antero-posterior diameter of the auditory vesicle is calculated to be 0.48 mm., since it runs through 32 sections.

The auditory vesicle has at this stage undergone considerable development. The pars superior and the pars inferior are distinctly separated. The anterior and the posterior semicircular canals are now completely constricted off; but this is not the case with the lateral canal; i.e., this canal is not yet independent of, but still broadly in communication with the main lumen of the vesicle. The pars inferior is well differentiated and possesses an elongated oval swelling on the ventro-medial wall of the vesicle. The ductus endolymphaticus appears as a long slender tube.

In correspondence with the external change in form the epithelial lining is also well differentiated. The anterior canal, which is flattened in a medio-lateral (partly dorso-ventral) direction, widens out at its anterior end into an ampulla, and the crista acustica anterior is here represented by high epithelium which is continuous, without any decrease in thickness, with the macula utriculi. The same holds true for the crista laterahs, the epithelium of which is somewhat lower than that of the anterior crista. The medial and ventral walls of the utriculus are made up of especially high stratified epithelium, which, bending upon itself at the entrance of the pars inferior, passes over into this without any sharp boundary line. The tallest epithelium of the medial wall decreases somewhat in thickness as it passes over into the medial wall of the endolymphatic duct. The flattened lateral wall of the utriculus presents no points of especial interest. The crista posterior has moved back some distance and appears as a thickened zone of cells in several layers at the ventro-medial portion of the semicircular canal.

Stage 3 (fig. 3) The embryo consists of about 2| coils. On the surface of the body striations are observed which are transverse on the ventral and crossed on the dorsal surface. Fixation: Corrosive-acetic. Stain: alcoholic borax carmine.


NERVE ENDORGANS IN THE EAR 295

The 15 sections cut frontally through head and body. The membranous labyrinth runs through 102 sections and hence has an antero-posterior diameter of 1.53 mm.

The utriculus and the sacculus communicate by a narrow foramen, the canahs utriculo-saccularis ; the lateral semicircular canal is now an independent structure. * Each nerve endorgan is well developed. The crista anterior is mound-shaped; the crista lateralis is a thick cell mass which appears as a crescent in the sections. Both structures still maintain their connection with the crista utriculi.

The tall epithelium of the utricular froor, which diminishes in thickness as it passes upward, doubtless represents the first anlage of the macula neglecta Retzii. It is continuous with the macula partis inferioris through the still cylindrical epithelium of the canalis utriculo-saccularis. The macula partis inferior consists in this stage of an extended zone of neuroepithelium on the medial wall of the pars inferior and already there is to be seen on its margin several minimal though unmistakable points devoid of nuclei. The fine nerve-fiber bundles that arise from the ganglion acusticum show excellent mitotic figures where the fibers enter the macula. The low cylindrical epithelium of the ductus endolymphaticus is continuous with the tall neuroepithelium of the medial wall of the sacculus.

Stage 4 (fig. 4). The embryo, which is made up of 2^ coils, has the appearance of a fuUy developed individual. Its peculiar dermal spots are prominently displayed over the entire body. Fixation: formol. Stain: Alcoholic borax carmine. Sections: 15 M in thickness, cut frontally through the head.

The nerve endorgans are nearly all differentiated and on each the marginal zone free of nuclei may be recognized. The cristae anterior and posterior are separated from the macula utriculi by a low epithelium.

It is worthy of notice that the thick epithehum of the utricular wall shows clearly a border without nuclei and that it is differentiated from the epithelium of the canal by its greater thickness. It soon becomes thinner as it passes gradually over into the undifferentiated epithelium lining the vesicle. This thickening just


296 TOKUYASU KUDO

referred to may well be considered as the first anlage of the macula neglecta Retzii. In the wall of the canal there is no zone marked out by a cell-free border, although the epithelium is still rather thick, and this in turn is continuous with the mound-shaped swelling, the macula sacculi.

Corresponding to the external changes in form, the macula partis inferioris is now separated into the papillae basilaris and lagenae, which are still united by cubical epithelium. The crista posterior is quite separated from the macula sacculi by an unspecialized epithelium.

Stage 5. The embryo consists of 2| coils. The external characters are quite comparable to those of the preceding stage. Fixation: potassium bichromate. Stain: alcoholic borax carmine. The 15 M sections are cut frontally through the head.

The macula neglecta Retzii, which lies closely adjoining the canalis utriculo-saccularis, is mound-shaped and consists of two or three layers of cells. The maculae neglecta and sacculi are united by means of cubical epithelium except in the wall of the canal, where the epithelial cells are still tall.

The Adult Animal (fig. 5). Fixation in potassium bichromateacetic. Stain: haematoxylin-eosin and haematoxylin-orange G. The section are cut frontally through the head.

Among the endorgans the cristae anterior and posterior are composed of two- to three-layered epithelium and project as rounded protuberances into the lumen. The macula utriculi lies on the anterior-medial wall of the utriculus and is composed of auditory and supporting cells. The macula neglecta appears as a swelling in the proximity of the canalis utriculo-saccularis on the floor of the utriculus; its vesicular auditory cells rest upon one or two layers of supporting cells. The macula diminishes in thickness as it passes over into the simple cylindrical epithelium which makes up the wall of the canal and which is continued beyond in the wall of the sacculus. The tall epithelium found on the medial wall of the canal is also to be seen on and near the lateral wall. In several places within and near the canal the lining is thrown up into wave-like folds.


NERVE ENDORGANS IN THE EAR 297

DISCUSSION

The results of my studies, as presented above, agree on the origin of the macula neglecta with the view of Fleissig, for it has been shown that this macula is derived directly from the macula partis inferioris. Even after the neuroepithelium has been completely separated by the undifferentiated epithelium from the pars inferioris, the macula neglecta remains for a long time in connection with the macula sacculi.

The common neuroepithelium on the ventro-medial wall of the auditory vesicle of stage 1 begins to divide into the utricular and the saccular portions (stage 2), the histological changes in the epithelium keeping pace with the external changes in form. The more strictly utricular portion swells to form the crista anterior, crista lateralis and macula utriculi, which are united by means of a tall epithelium. The more strictly saccular portion, separated from the utricular portion by flattened epithe lium (stage 3) still extends from the medial wall of the canalis utriculo-saccularis upwards further into the floor of the utriculus.

After the macula saccularis has been differentiated (stage 3) the macula neglecta gradually protrudes more and more into the lumen and in stage 4 discloses a border free of nuclei, but is still connected by means of a cubical epithelial layer with the macula sacculi. Furthermore, the crista ampuUaris posterior becomes entirely free from the saccular portion, while the papillae basilaris and lagenae still maintain their connection with the macula saccularis by means of a bridge of cubical epithelium. In stage 5 the well developed macula neglecta may be seen as a moundshaped structure as in adult specimens.

The existence of two maculae neglectae I have failed to demonstrate in my Trigonocephalus material, although I have minutely examined the rather comprehensive series of the different stages. Fleissig says: "1) die macula sacculi, welche nicht mehr die ganze mediale Sacculuswand, sondem nur mehr deren unteresten Abschnitt einnimmt. Ein Epithel, das etwas hoher ist als das indifferente Wandepithel und ganz typisch in der Umgebung der Nervendstellen vorkommt, erstreckte sich von der Macula


298 TOKUYASU KUDO

sacculi nach aufwarts zum Foramen Utr.-Sacc, wo es zu einer zweiten Neuroepithelstelle — 2) Macula neglecta Sarasinianschwillt, die im Foramen Utr.- sacc. (an dessen hinterem Rand) gelegen, zum kleineren Teil in den Sacculus, zum grosseren in den Utriculus hineinragt. Von dieser erstreckt sich wieder ein niedriges Epithel in den Sinus inferior hinein zu persistierenden 3) Macula neglecta (Retzii). Beide Maculae neglectae stehen auf derselben Entwicklungsstufe."

Now even if the bulging endorgan found in the floor of the utriculus of stages 4 and 5 were not to be regarded as the macula neglecta Sarasini but rather as the macula neglecta Retzii, I would not feel justified in interpreting the thickened epithelium which extends through the canalis utriculo-saccularis to the macula saccularis as the macula Sarasini. The further the development progresses the thinner does the epithelium of the inner w^all of the alveus become as compared with the early stage of the auditoryvesicle. One may readily see that the medial wall of the alveus communis is lined with relatively taller epithelial cells in stage 2 than in stage 3. From this it is apparent that the neuroepithelium, except where it progressively develops into nerve endorgans, is destined to be reduced to indifferent epithelium, even though the time when it retrogresses be very variable.

According to my opinion, therefore, the tall epithelium of medial wall of the canal and its proximity represents a developmental stage in the neuroepithelium which later retrogresses. If this epithelium were to be interpreted as a nerve endorgan, the tall epithelium of other regions, as e.g., of the lateral wall of the canal and the medial wall of the utriculus and the ductus endolymphaticus, would have to be regarded as neuroepithelium, since these latter regions are quite similar in structure and arrangement of their epithelial cells to those in the medial wall of the canal. At any rate, the macula neglecta does not occur in my material as it has been pictured by Fleissig in his work. But it should be noted that in the adult snake the epithelium of the canalis utriculo-saccularis and its immediate environs is relatively much thicker as compared with the medial and lateral walls.


NERVE ENDORGANS IN THE EAR 299

From the above it appears, then, that the macula neglecta Retzii, which comes to He in the floor of the utriculus, arises from the neuroepithehum of the pars inferior, as was first estabhshed by Fleissig in the case of Gecko; but, as stated above, I am unable to demonstrate in my material any progressively developing endorgan which could represent the macula neglecta Sarasini.

Alexander has suggested that in the embryo of Echidna the tall epithelium at the mouth of the ductus endolymphaticus may represent the vestige of the Amphibian macula neglecta Sarasini. This tall epithelium, which is continuous with the neuroepithehum of the medial utricular wall, Fleissig has also observed in the embryo of Gecko, but his interpretation is a totally different one, for he does not consider it remarkable that the mouth of the ductus endolymphaticus, which is still in active growth, should possess tall epithelium where it passes suddenly into the neuroepithelial anlage of the medial utricular wall.

In conclusion I desire to record the observation that the three semicircular canals of Trigonocephalus japonicus do not develop synchronously, the medial and posterior canals anticipating the lateral canal in their development.

SUMMARY

1. The macula neglecta arises directly from the macula partis inferioris.

2. The occurrence of two maculae neglectae is not to be observed in my material : while the macula neglecta Retzii is well developed, there does not form a persistent macula Sarasini nor does this endorgan even develop temporarily as in Gecko (Fleissig).

3. The anterior and the posterior semicircular canals are separated off much earlier than the lateral canal,

Kyoto, Sept. 15, 1914.


300 TOKUYASU KUDO

LITERATURE CITED

The references marked with an asterisk (*) were available to the author.

ALEXA^^DER, G. 1900 tJber Entwickelung und Bau der Pars inferior labyrinthi der hoheren Wirbeltiere. Denkschr. d. k. Akad. d. Wiss. Math.Naturw. Kl. 70. Alexander, G. 1904 Entwickelung und Bau des inneren Gehororgans von Echidna aculeata. Jenaische Denkschr., Bd. 6.

1904 Zur Entwickelungsgeschichte und Anatomic des inneren Gehororgans der Monotremen. Centralbl. f. Phys. Bd. 17.

1905 Zur Frage der phylogenetischen, vicariierenden Ausbildung der Sinnesorgane (Talpa europaea und Spalax typhlus). Zeitschr. f. Psych, u. Phys. d. Sinnesorg. Bd. 38.

Ayers, H. 1892 Vertebrale Cephalogenesis. 2. A Contribution to the morphology of the Vertebrate Ear, etc. Journ. of Morph. vol. 6. 1893 The macula neglecta again. Anat. Anz. Bd. 8.

Deiters, D. 1862 Ueber das innere Gehororgan der Amphibien. Reichert u. Du Bois Re.ymonds Arch.

Fleissig, J. 1908 Die Entwickelung des Geckolabyrinthes. Ein Beitrag zur Entwickelung des Reptilienlabyrinthes. Anat. Hefte, Bd. 37.

Krause, R. 1906 Entwickelungsgeschichte des Gehororgans. Hertwigs Handbuch d. Vergl. u. Experim. Entw.-Lehre.

Krause, R. 1906 Das Gehororgan der Petromyzonten. Anat. Anz. Erg. -Heft 7. Bd. 29.

Okajima, K. 1911 Die Entwickelung des Gehororgans von Hynobius. Anat. Hefte. Bd. 45.

Okajima, K. 1911 Die Entwickelung der Macula neglecta beim Salmoembryo. Anat. Anz. Bd. 40.

Retzius, G. 1878 Zur Kenntniss von dem membranosen Gehorlabyrinth bei den Knorpelfischen. Arch. f. Anat. u. Phys. Anat. Abt. Jahrg.

Retzius, G. 1880 Zur Kenntniss des inneren Gehororgans der Wirbeltiere. Arch. f. Anat. u. Phys. Anat. Abt. Jahrg.

Sar.\sin, p. u. F. 1890 Ergebnisse naturwissenschaftlicher Forschungen auf Ceylon. Bd. 2.

Sarasin, p. u. F. 1892 Ueber das Gehororgan der Caeciliiden. Anat. Anz. Bd. 7.

StIjtz, L. 1912 Ueber sogenannte atypische Epithelformation im hautigen Labyrinth. -Eine rudimentiire Mac. negl. Morph. Jahrb. Bd. 44.

Wenig, J. 1913 Untersuchungen liber die Entwickelung der Gehororgane der Anamnia. Morph. Jahrb. Bd. 45.

WiTTMAACK 1911 Ueber sogenannte atypische Epithelformation im membranosen Labyrinth. Verh. d. Deutsch. Otol. Gesell.


PLATE


301


PLATE 1

Weigert's Iron haematoxylin, Leitz Achromat 6; Ocular I. Boraxcarmine. 3X1. Boraxcarmine. 3X1.


1 Stage 1. Stain:

2 Stage 2. Stain:

3 Stages. Stain:

4 Stage 4. Stain: Boraxcarmine. 3X1.

5 Adult. Stain: Haematoxylin-eosin, IXI,


ABBREVIATIONS


C.u.s. Canalis utriculo-saccularis

B., Brain

A. v., Auditory vesicle

Lag., Lagena

L.c, Lateral semicircular canal

A.c, Anterior semicircular canal

U., Utriculus


M.n., Macula neglecta M.S., Macula sacculi 0., Otolith P.i., Pars inferior P.S., Pars superior S., Sacculus


302


NERVE ENDORGANS IN THE EAR

TOKUYASr KUDO


PLATE 1



"* '0.


303


author's abstract of this paper issued b? the bibliographic service, may 11


AN INTRODUCTION TO A SERIES OF STUDIES ON THE SYMPATHETIC NERVOUS SYSTEM

S. W. RANSON From the Northwestern University Medical School^

ONE FIGURE

Anatomists have devoted little thought to the functional pathways within the sympathetic nervous .system. Yet it is obvious that no account of the structure of any part of the nervous system is complete which does not include an analysis of the more important conduction paths. Such an analysis cannot, as a rule, be made by purely morphological methods, but requires the aid of physiological procedures including degeneration experiments. Above all, the investigator must approach his subject from the right point of view; he must regard the structures to be analyzed as parts of a functional mechanism and strive to understand how it works.

While histologists have given ais many details concerning the structure of the ganglia, they have ignored the composition of the various nerves and plexuses in the sympathetic system and have made little effort to analyze what seemed to them a hopeless confusion of interconnected elements. In the anatomical and histological texts we find no hint that the sympathetic nervous system is made up of definite functional groups and chains of neurones as distinct and sharply limited as are any of the conduction systems of the brain and spinal cord. Nevertheless, such is the case; it is even probable that the functional groups and chains of neurones are more sharply limited in the sympathetic than in the central nervous system. The latter is provided with a mechanism for the widest possible diffusion of incoming impulses, while such diffusion does not occur in the former. Strong stimulation of a single small cutaneous nerve will give

» Contribution No. 53, February 15, 1918.

305

THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL. 29, NO. 4 AUGUST, 1918


306


S. W. RANSON


rise to nerve impulses which are distributed throughout the brain and spinal cord and may call into action any part of the smooth or striated musculature of the body. Nothing in any way comparable to this occurs in the sympathetic system.

Excluding the terminal ganglionated plexuses which require further study, we may say that there is probably no more opportunity for diffusion of nerve impulses in the sympathetic nervous system than there is in an ordinary spinal nerve. This can



Fig. 1 Diagram of two conduction paths from which all purely topographic details, such as spinal nerves, rami communicantes, and sympathetic trunk, have been omitted: a, somatic path with branching efferent fiber; b, autonomic path with branching preganglionic efferent fiber, the branches ending in relation to two postganglionic neurones.

be made clear by a diagram (fig. 1). So far as the possibility for diffusion of nerve impulses is concerned, it is immaterial whether the efferent fiber branches in the course of a nerve or within a ganglion and whether its branches come in contact with the innervated structure directly or through the mediation of a second neurone, provided there is in the ganglion no other type of synapse than that indicated in the diagram.

Thanks to the work of Langley, we have reason to believe that the sympathetic system, with the probable exception of the terminal ganglionated plexuses, is built up on the simple lines


STUDIES ON THE SYMPATHETIC NERVOUS SYSTEM 307

indicated in the diagram; and, if so, the working out of conduction pathways should not be as difficult as we had supposed. In fact, a great deal along this line has already been accomplished by the physiologists; but there yet remains a large amount of work to be done before the course of nerve impulses through the sympathetic nervous system can be mapped with accuracy

Since there is considerable confusion in the use of terms referring to this division of the nervous system, we wish at the outset to define those which we shall have occasion to use.

The sympathetic nervous system is an aggregation of ganglia, plexuses, and nerves through which the glands, heart, and all smooth muscle receive their innervation. It is a term belonging primarily to descriptive anatomy and includes the ganglionated plexuses associated with the fifth nerve and the vagal plexuses of the thorax, as well as the sympathetic trunk and the parts more directly associated with the latter. Since it is connected at many points with the cerebrospinal nerves, it is necessary to decide what shall be included in it. The logical point of separation is that at which the cerebrospinal nerves give off branches which run exclusively to the sympathetic system. These branches of the cerebrospinal nerves form an integral part of this system. This is well recognized in the case of .the rami communicantes; but the principle has never been carried through systematically. On this basis it would include the radix brevis of the ciliary ganglion, the cardiac and pulmonary rami of the vagus, and the visceral rami of the second, third, and fourth sacral nerves. We pass now to a consideration of the terms selected from the vocabulary of the physiologists.

The autonomic nervous system is that functional division of the nervous system which supplies the glands, heart, and all smooth muscle with their efferent innervation. It is the sum total of all general visceral efferent neurones both pre- and postganglionic.

The preganglionic visceral efferent neurones have their cells located in the cerebrospinal axis, and their fibers make their exit from this axis in three streams: 1) cranial — via the III, VII, IX, X, XI cranial nerves; 2) thoracicolumbar — via the white


308 S. W. RANSON

rami communicantes from the thoracic and upper lumbar spinal nerves; 3) sacral — via the visceral rami of the II, III, and IV sacral nerves. The fibers of the thoracicolumbar stream run to the sjTnpathetic trunk and are distributed through it to ganglia at higher and lower levels. The fibers of the cranial and sacral streams make no connection with the sympathetic trunk, but run directly to the various plexuses. While the fibers of the thoracicolumbar stream end in the ganglia of the trunk or in collateral ganglia, those of the cranial and sacral streams end in terminal ganglia. In these two respects the cranial and sacral streams agree with each other and differ from the thoracicolumbar stream. Also physiologically and pharmacologically the two former agree with each other and differ from the latter. It is therefore desirable to divide the autonomic nervous system into two divisions:

1. The thoracicolumbar autonomic system (called by many physiologists the sympathetic nervous system).

2. The craniosacral autonomic system (called by many physiologists the parasympathetic system).

The importance of this division is further emphasized by the fact that most of the structures innervated by the autonomic system receive a double nerve supply, being furnished with fibers from both divisions of that system. The thoracicolumbar fibers are accompanied in most peripheral plexuses by craniosacral fibers of opposite function, so that an analysis of these plexuses is greatly facilitated by subdividing the autonomic system in this way. These statements may be summarized in the form of three definitions:

The autonomic nervous system is that functional division of the nervous system which supplies the glands, the heart, and all smooth muscle, with their efferent innervation and includes all general visceral efferent neurones both pre- and postganglionic.

The thoracicolumbar autonomic system is that division of the autonomic system, the preganglionic fibers of which make their exit from the spinal cord through the thoracic and upper lumbar spinal nerves.


STUDIES ON THE SYMPATHETIC NERVOUS SYSTEM 309

The craniosacral autonomic system is that division of the autonomic system, the preganglionic fibers of which make their exit from the cerebrospinal axis through the III, VII, IX, X, and XI cranial nerves and the II, III, and IV sacral nerves.

The preganglionic neurones are those, the cell bodies of which lie in the brain or spinal cord and whose axons run through the cerebrospinal nerves to enter the sympathetic system and end in its ganglia. The autonomic nervous system therefore includes certain cells in the brain and spinal cord and certain fibers in the cerebrospinal nerves and is not contained exclusively in the sympathetic system. The postganglionic neurones are those whose cellbodies lie in the sympathetic ganglia and whose axons run to end on cardiac or smooth muscle or in glandular tissue.

In order to show how these terms will aid in the presentation cf the facts of visceral innervation, we may give a few examples. While some points are still obscure, the outlines given below are as nearly correct as our present knowledge enables us to make them. They are given not as an ultimate statement of fact, but as an illustration of the sort of information which we should strive to perfect.

IMPORTANT FUNCTIONAL PATHS IN THE AUTONOMIC SYSTEM

1. Paths for the efferent innervation of the eye.

a. Ocular craniosacral pathway.

Preganglionic neurones. Cells in the oculomotor nucleus, fibers by way of the III cranial nerve to end in the ciliary ganglion.

Postganglionic neurones. Cells in the ciliary ganglion, fibers by way of the short ciliary nerves to the ciliary muscle and the circular fibers of the iris.

Function — accommodation and contraction of the pupil.

b. Ocular thoracicolumbar pathway.

Preganglionic neurones. Cells in the intermediolateral column of the spinal cord, fibers by way of the upper white rami and sympathetic trunk to end in the superior cervical ganglion.


310 S. W. RANSON

Postganglionic neurones. Cells in the superior cervical ganglion, fibers by way of the internal carotid plexus to the ophthalmic division of the Vth nerve, the nasociliary and long ciliary nerves to the eyeball: other fibers pass from the internal carotid plexus through the ciliary ganglion, without interruption, into the short ciliar}'- nerves and to the eyeball.

Function — dilation of the pupil by the radial muscle fibers of the iris.

2. Paths for the efferent innervation of the submaxillary gland,

a. Submaxillary craniosacral pathway.

Preganglionic neurones. Cells in the nucleus salivatorius superior, fibers by way of the seventh cranial nerve, chorda tympani and lingual nerve to end in the submaxillary ganglion on the submaxillary duct.

Postganglionic neurones. Cells in a number of groups along the chorda tympani fibers as they follow the submaxillary duct, fibers distributed in branches to the submaxillary gland.

Function — increases secretion. h. Submaxillary thoracicolumbar pathway.

Preganglionic neurones. Cells in the intermediolateral column of the spinal cord, fibers by way of the upper white rami, and the sympathetic trunk to end in the superior cervical ganglion.

Postganglionic neurones Cells in the superior cervical ganglion, fibers by way of the plexuses on the external carotid and external maxillary arteries to the submaxillary gland.

Function — increases secretion.

3. Paths for the efferent innervation of the heart,

a. Cardiac craniosacral pathway.

Preganglionic neurones. Cells in the dorsal motor nucleus of the vagus, fibers through the vagus nerve to the intrinsic ganglia of the heart in which they end.

Postganglionic neurones. Cells in the intrinsic cardiac ganglia, fibers to the cardiac muscle.

Function — cardiac inhibition.


STUDIES ON THE SYMPATHETIC NERVOUS SYSTEM 311

b. Cardiac thoracicolumbar pathway.

Preganglionic neurones. Cells in the intermediolateral column of the spinal cord, fibers by way of the upper white rami and the sympathetic trunk to the superior, middle, and inferior cervical ganglia.

Postganglionic neurones. Cells in the cervical ganglia of the sympathetic trunk, fibers byway of the corresponding cardiac nerves to the musculature of the heart. Function — cardiac acceleration. 4. Paths for the efferent innervation of the musculature of the stomach exclusive of the sphincters. a. Gastric craniosacral pathway.

Preganglionic neurones. Cells in the dorsal motor nucleus of the vagus, fibers by way of the vagus nerve to end in the intrinsic ganglia of the stomach.

Postganglionic neurones. Cells in the intrinsic gastric ganglia, fibers to end on the gastric musculature. Function — excites peristalsis. b. Gastric thoracicolumbar pathway.

Preganglionic neurones. Cells in the intermediolateral column of the spinal cord, fibers by way of the white rami from the 5th or 6th to the 12th thoracic nerves, through the sympathetic trunk without interruption, and along the splanchnic nerves to the coeliac ganglion where they end.

Postganglionic neurones. Cells in the coeliac ganglion, fibers by way of the coeliac plexus and its offshoots to the stomach to end on the musculature of the stomach. Function — inhibits peristalsis. It will be noted that the organs receive a double autonomic innervation and that the impulses transmitted along the craniosacral pathways are usually antagonistic to those transmitted along the thoracicolumbar paths.

The afferent innervation of the viscera. General visceral afferent fibers are found in the IX and X cranial nerves and in the spinal nerves. Their cells of origin are located in the cerebrospinal ganglia. The fibers run through the sympathetic


312 S. W. RANSON

nervous system, passing through the ganglia and plexuses without interruption, to end in the viscera. There is no satisfactory -evidence that any afferent neurones have their cell bodies located in the sympathetic ganglia. The function of these afferent fibers is to convey to the central nervous system impulses giving rise to vague sensations, and other impulses, which never rising into consciousness, give rise to visceral reflexes.

Visceral reflex arcs. In the gastrointestinal tract there may be a mechanism for purely local reflexes, i.e., there are probably reflex arcs complete within the gut wall. With this exception the evidence strongly indicates that all visceral reflex arcs pass through the cerebrospinal axis and involve a series of three neurones: 1) visceral afferent; 2) preganglionic autonomic, and 3) postganglionic autonomic. The purely local reflexes which seem to occur within the gut wall after section of all the nerves leading to the intestine are known as the myenteric reflexes and must depend upon a mechanism different from that of other visceral reflexes. We do not know what this mechanism is, but it must be located in the enteric plexuses. The term enteric nervous system should be restricted to the elements responsible for the myenteric reflex.

In the papers which follow there will be presented some of the evidence that has led me to take the general position in regard to the sympathetic nervous system outlined in the preceding pages. For much of the evidence, however, it will be necessary for the reader to refer to the papers of Langley. To this evidence Dr. Johnson has made an important contribution in showing that there are no commissural neurones in the ganglia of the sympathetic trunk of the frog. The papers of Dr. Billingsley and myself are primarily concerned with details of structure, a knowledge of which will be necessary for any future attempt to map the functional pathways of the sympathetic nervous system.


authors' abstract of this p.vper issued by the bibliographic service may 11


THE SUPERIOR CERVICAL GANGLION AND THE CERVICAL PORTION OF THE SYMPATHETIC

TRUNK

S. W. RANSON AND P. R. BILLINGSLEY

From the Anatomical Laboratory of Northwestern University Medical School^

FIFTEEN FIGURES

In this paper we shall report observations on the superior cervical ganglion and the nerves immediately associated with it. But in dealing with the literature it has been necesssary to treat the subject in a somewhat broader way and to set forth what is known concerning the sympathetic ganglia in general.

The general plan of the cephalic end of the sympathetic trunk, according to the evidence obtained by the nicotine and degeneration methods, is as follows : The trunk below the superior cervical ganglion consists of fibers ascending to end in that ganglion (fig. 1). These are preganglionic fibers, the axons of cells located in the intermediolateral cell column of the spinal cord, which have entered the trunk through the upper thoracic white rami and are ascending to the ganglion. Having reached the superior cervical ganglion, these fibers end in synapses with the postganglionic neurones, whose cell bodies are located there, and to which belong the postganglionic fibers that leave this ganglion through its various branches of distribution. Those branches which run to the internal carotid artery, known collectively as the internal carotid nerve and forming the internal carotid plexus, carry postganglionic fibers which are distributed to the eyeball, lacrimal gland, mucous membrane of the nose, mouth, and pharynx and many of the blood-vessels of the head. The fibers to the salivary glands run by way of the branch to the external carotid artery

1 Contribution No. 54, February 15, 1918.

313


314


S. W. RANSON AND P. R. BILLINGSLEY


(jrland.lac»

A.JcLC.

i ri.cil.opev. T^.cillonq.

'■ a oil ^


A..opnthalm.

Al.opntka/m. / Plex.cavern.

Jy.trj^eTn.

Jx.man. Npetr. sup. m aj.

' Gland.

pdrot



G.cerv.sup. .Jx.carot.ext.

.A..carot.int.


fflZX


Fig. 1 Diagram representing the arrangement of the more important thoracicolumbar autonomic pathways to the head in man. The preganglionic fibers are indicated by solid lines. The cells of the postganglionic neurones are located in the superior cervical ganglion and their fibers are indicated by dotted lines. 1, Postganglionic fibers to sweat glands of the face; 2 and 3, to the mucous membrane of the nose; 4, N. cardiacus superior; 5, Rr. laryngopharyngei; 6, branch to the N. hypoglossus; 7, branch to the N. vagus; 8, n. caroticus internus; 9, branch to the N. glossopharyngeus; 10, 11, 12, 13, Rami communicantes (gray) to Nn. cervicales I, II, III, and IV.


THE CERVICAL SYMPATHETIC TRUNK 315

and follow along its branches to the glands. Through the superior cardiac nerve postganglionic fibers run to the heart in man. Other postganglionic fibers join the upper four spinal nerves and the ninth, tenth, and twelfth cranial nerves to be distributed to the blood-vessels and glands in the regions supplied by these nerves, and still others run by the laryngopharyngeal branches of the superior cervical ganglion to the larynx and pharynx.

This will serve as a general survey of the field to be studied. In the pages which follow we will take up in detail the structure of the superior cervical ganglion, sympathetic trunk ^nd internal carotid nerve and pay particular attention to the synapses which occur in the ganglion.

MATERIAL AND METHODS

The superior cervical ganghon of man, the dog, and cat were prepared by the pyridine silver method and cut into sections 12 to 20 micra thick. Osmic acid preparations were also made from the dog and cat. Many of the preparations were cut into serial sections at right angles to the long axis of the ganglion, beginning at the internal carotid nerve and extending through the ganglion and some distance along the sympathetic trunk. Other ganglia were also examined, such as the stellate ganglion of the cat and the superior cervical ganglion in the rabbit.

In addition to the study of these parts in normal animals, experiments were carried out to determine the effect of partial and of complete degeneration of the preganglionic fibers. It had been noticed in a stud}' of degenerating and regenerating nerves, made several years previously that certain fibers in the early stages of degeneration showed an increased affinity for silver. It was hoped that this might furnish a clue which would lead to the development of a differential stain for degenerating axons. A number of ganglia were prepared by the pyridine silver method sixteen or seventeen hours after section of the sympathetic trunk in the neck to see if by this method the preganglionic fibers might be made to stain more intensely through an increased affinity for the silver. So far we are not convinced


316 S. W. RANSON AND P. R, BILLINGSLEY

that any advantage was obtained by this procedure. It is true that the majority of our best preparations of the preganghonic fibers were obtained in this way, but since we occasionally obtained just as good stains in normal animals we are in doubt as to the value of the preliminary division of the fibers. We shall consider these as preparations of the normal ganglia since if there is any change it is only in the direction of an increased affinity of these fibers for the silver.

In order to obtain complete degeneration of the preganglionic fibers the sympathetic trunk was divided in the neck. The operation was perfoi'med aseptically on cats and dogs, the nerve being cut about 2 inches below the ganglion. After periods of from eight to fifty days some of the animals were killed. It was found that after the longer periods some regeneration had occurred and the shorter periods were scarcely adequate for full degeneration. In order to avoid these difficulties, a second operation was performed on some of the animals twenty to fifty days after the first, the nerve being cut cephalad to the neuroma. Eight days after the second operation the animals were killed.

In dealing with small nerves and ganglia we have found that the pyridine silver stain often fails to give good results apparently because the volume of the tissue is too small. In order to overcom.e this difficulty we find it desirable to imbed the small nerve or ganglion in the spinal cord. For this purpose we have tied a fine silk thread to the sympathetic trunk and with a long fine needle have drawn the trunk with the attached superior cervical ganglion and internal carotid nerve into a lateral half of the spinal cord along the line of the ventral gray column. After fixation for two hours in ammoniated alcohol the block of spinal cord can be pared down with a razor until it forms a bar the crosssection of which is not morq than 4 mm. square. Within this block of cord the nerve is held extended and straight and is protected from the two direct action of the reagents. The cord is dissected away from the nerve just before it is dehydrated and cleared in preparation for imbedding.


o


THE CERVICAL SYMPATHETIC TRUNK 317

STRUCTURE OF THE CEPHALIC END OF THE SYMPATHETIC TRUNK

As has been said, the cervical portion of the sympathetic trunk serves to convey preganglionic fibers from the upper white rami to the superior cervical ganglion. Whether it also contains other than preganglionic fibers is a question which we will consider in this paper. In the cat this nerve, a short distance below the superior cervical ganglion, has the structure shown in figure 2. In cross-section it is uniform throughout except for one or two small well-defined bundles at the periphery. These bundles are not constant and, as we shall see, represent fine branches of distribution from the ganglion which have been incorporated for a short distance in the trunk.

Fig. 2 From a section of the truncus sympathicus a short distance below the ganglion cervicale superius in the cat. a, area occupied by a bundle of unmyelinated fibers. Osmic acid. X 425.

Exclusive of these peripheral bundles which really do not belong to it, the sympathetic trunk below the superior cervical ganglion in the cat consists almost exclusively of myelinated fibers as shown in figure 2. These are uniformly distributed and closely packed. It is as well myelinated a nerve as there is anywhere in the body. The fibers are all rather fine. The majority vary in diameter from 1.5)U to 3.5^- Between these two extremes there are fibers of all sizes and about as many of one size as another. Fibers larger than 4.5;u are few in number but there may be two or three as large as 6.5 or 7^. Pyridine silver preparations show rather small axons, each surrounded by an unstained halo representing a myelin sheath; these are uniformly distributed, each well separated from its neighbor. There are


318 S. W. RANSON AND P. R. BILLINGSLEY

no bundles of closel}' packed unmyelinated axons and no individual ones can be made out with certainty. From a study of the normal truncus sympathicus we may conclude that it is composed almost exclusively of small myelinated fibers.

The fine peripheral bundles, which represent branches of distribution from the superior cervical ganglion, can usually be followed in serial sections to the point where they are given off as fine branches from the trunk. They do not degenerate after section of the nerve more caudally. The structure of these peripheral bundles is entirely different from that of the rest of the nerve and corresponds to that of the other branches of distribution given off from the superior cervical ganglion. They contain a few small myelinated fibers, 1.5^ to 6ju in diameter, scattered among the unmyelinated fibers. Such a bundle is seen at a in figure 2 where the area occupied by the umyelinated fibers is indicated by stippling. In osmic acid preparations bundles of unmyelinated fibers are recognized by their being somewhat more darkly stained than the rest of the background. A fascicle of axons, even though lightly stained, is easily differentiated from connective tissue. Additional information may be obtained by the study of the degenerated nerve. In an osmic acid preparation taken from a cat eight days after neurotomy of the sympathetic trunk in the neck most of the medullated fibers are degenerated, although a few cannot be distinguished from normal fibers. But eighteen days after the operation all the medullated fibers were degenerated except for a small number in a single peripheral fascicle, such as has been described and which is not to be regarded as belonging to the nerve. There were 16 myelinated fibers in this bundle varying in size from l.S^i to 3.6;Lt. All the other myelinated fibers in the nerve were degenerated. From this we may conclude that all the myelinated fibers in the cephalic end of the sympathetic trunk (exclusive of branches of the superior cervical ganglion which may be incorporated with it for a short distance) are ascending fibers. There are no medullated fibers arising in the superior cervical ganglion and running to the ganglia placed more caudally in the truncus sympathicus.


THE CERVICAL SYMPATHETIC TRUNK 319

Waller and Budge showed long ago that the sympathetic trunk after section in the neck degenerated toward the superior cervical ganglion. Their results have been confirmed by Langley ('96, '00). This author used the rather unsatisfactory method of examining the degenerated nerve in teased preparations stained with osmic acid. He found, however, just as we, in sections stained with osmic acid, that some fine branches of the superior cervical ganglion ma}^ accompany the nerve for a certain distance. He also found that occasionally a branch from the vagus might run to the superior cervical ganglion and accompany the nerve for a way. This may have been the depressor nerve (p. 374).

After the sympathetic trunk below the stellate ganglion and the rami communicantes to the first and second thoracic nerves were cut and time allowed for degeneration, he found no sound myelinated fibers in the cervical portion of the nerve, aside from the bundles just mentioned which may happen to be included in the same sheath with it. He concluded that no myelinated fibers run from ganglion to ganglion through this nerve and none join it from the cervical rami communicantes.

We have pyridine silver preparations of the degenerated nerve in both cat and dog. In each case the structure is the same. Take, for example, Cat XII which was killed fifteen days after the division of the sympathetic trunk in the neck. In that part of the nerve just below the superior cervical ganglion the sections stained with silver showed two fascicles of fine undegenerated axons mostly unmyelinated at the periphery of the trunk. Following the sections caudally through the series, one of these fascicles can be seen to leave the trunk, but the other remains with it as far as our series goes, although it would no doubt separate off a little farther down.

Aside from these two peripheral fescicles, which, properly speaking, do not belong to the nerve, almost all of the axons have degenerated. Here and there throughout the section there seems to be an isolated unmyelinated axon of normal appearance. These normal unmyelinated fibers are not numerous. In fact, since we have never seen such isolated unmyehnated axons


320 S. W. RANSON AND P. R. BILLINGSLEY

elsewhere except in regenerating nerves, we are somewhat skeptical of this observation. The presence of these few axons descending from the superior cervical ganglion, however, raised the questions, are there commissural fibers joining the superior cervical with the stellate or other ganglia?

Here we can take up only the question of the existence of fibers connecting cells in different ganglia, and will leave out of account for the moment that of the interconnection of the cells within a single ganglion. According to Langley, there is no evidence which would justify us in assuming the existence of commissural fibers between the cells of different ganglia, and in certain parts of the sympathetic nervous system he has given strong evidence that no such connections exist. The mass of evidence which he has presented is very convincing, but is too extensive to be summarized here. The reader is referred to the account in Schaffer's Physiology, vol. 2, p. 683, and other articles by Langley in the Journal of Physiology, vol. 25, p. 468, and vol. 31, p. 244. We can refer here only to that part of the evidence which concerns the cervical portion of the sympathetic trunk. After this nerve was cut below the ganglion stellatum, and the rami communicantes to the first and second thoracic nerves divided and time allowed for degeneration, stimulation of the trunk in the neck produced no effect en the pupil, nictitating membrane, eyelids, hairs, or blood-vessels. Hence the cells of the ganglion stellatum or the middle cervical ganglion do not send nerve fibers to the superior cervical ganglion or to the head by way of this nerve. Even in the normal cat stimulation of this nerve produces no vasomotor, pilomotor, or secretory effect in the territory supplied with such fibers by the ganglion stellatum. It is clear, then, that the superior cervical ganglion does not send commissural fibers to the vasomoter, pilomoter, or secretory nerve cells of the ganglion stellatum which include the great majority of the cells in the ganglion. It is easy to show that stimulation of the sympathetic trunk in the neck is without appreciable effect on the heart of the cat. Hence no fibers descend from the superior cervical ganglion to the cardio-accelerator neurones of the middle cervical and stellate gangha.


THE CERVICAL SYMPATHETIC TRUNK 321

Langley has shown that stimulation of the sympathetic trunk in the neck causes no general body reflexes of any kind. It must, therefore, be devoid of sensory fibers, at least of those carrying painful afferent impulses. We have been able to confirm this physiological observation and our histological results are also in agreement with it. On page 432 we wdll shoV that the characteristic sensory fibers of the sympathetic trunk are the large myelinated and the unmyelinated. Except for two or three large myelinated fibers, there are no fibers which would be interpreted as sensory ascending in the cervical portion of the sympathetic trunk.

STRUCTURE OF THE NERVUS CAROTICUS INTERNUS

The chief set of branches given ofif by the superior cervical ganglion ascends from its upper pole to the internal carotid artery. Of these one or two are of large size in the cat. These large ones are easily and positively recognized in serial sections of the ganglion and its branches. The entire group of from three to five branches forms the nervus caroticus internus. It consists of both myelinated and unmyelinated fibers the latter of course predominating. Figure 3 shows the relative size, number, and arrangement of the myelinated fibers in this nerve in the cat. These fibers are rather widely separated by great numbers of unmyelinated axons and are of about the same size as those of the sympathetic trunk. They vary in diameter from 1.5/x to 4.5^1 with an occasional larger fiber up to 7/i. Their distribution is quite uniform throughout the nerve. The thickness of their myelin sheath seems to be somew^hat less than that of those in the sympathetic trunk.

These mj^elinated fibers are so numerous that interest is at once aroused as to their source, and the possibility suggests itself that they are preganglionic or perhaps afferent fibers from the trunk which have run through the ganglion without interruption. This possibility is easily excluded, however, by section of the trunk below the ganglion. After all the myelinated fibers in that trunk have degenerated the structure of the internal

THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL. 29, NO. 4


322


S. W. RANSON AND P. R. BILLINGSLEY


carotid nerve remains unchanged and contains as many myelinated fibers as the nerve of the opposite side. Measurements show that fibers of all sizes from 1.5 to 7/x are present, showing that there has not occurred a dropping out of the fibers of a particular size. In fact, figure 3 represents an internal carotid nerve after the Complete degeneration of the sympathetic trunk below the superior cervical ganglion of the same side, but illustrates perfectly well the normal structure of the nerve.



Fig. 3 From a section of the nervus caroticus internus in the cat. Osmic acid. X 425.


One must also consider the possibility of these myelinated fibers being contributed through the rami connecting the superior cervical ganglion with the upper cervical and certain of the cranial nerves. Against this assumption are the observations that can be made on serial sections through the superior cervical ganglion and , the internal carotid nerve after degeneration of the trunk.


THE CERVICAL SYMPATHETIC TRUNK 323

The myelinated fibers in such a ganglion are extremely few at the caudal pole, but increase gradually toward the cephalic end of the ganghon. They are scattered uniformly through the cross-section of the ganglion, until they begin to assemble at the upper pole to enter the internal carotid nerve. The few myelinated fibers that can be seen in the various side branches of the ganglion (to the cervical and cranial nerves) are at once lost in the ganglion. There are no bundles of medullated fibers running through the ganglion from one branch to the other. We believe that all or at least most of the myelinated fibers in the branches of the superior cervical ganglion arise from cells located in that ganglion. This will receive additional support from more detailed study of the structure of the superior cervical ganglion to follow.

Langley ('96) has shown that after section of the branches peripherally of the superior cervical ganglion nearly all of the myelinated fibers which remain connected with the ganglion are normal, while nearly all of those separated from the ganglion have degenerated, showing that the cells of origin of the great majority of these fibers are located in that ganglion. These observations were made on the cat. In the dog he has traced two small bundles of fibers from the tympanic plexus by way of the internal carotid artery to the superior cervical ganghon.

It is therefore evident that a considerable number of the axons arising in the superior cervical ganglion acquire a myelin sheath. This is in keeping with the results of v. KoUiker ('94), Dogiel ('95), Langley ('96), Michailow ('11), and others. It is interesting to note, however, that Cajal ('11) is of the opinion that the axons of the cells of the sympathetic ganglia never acquire myelin sheaths. It is easy to understand how he may never have been able to trace such an axon into a myelinated fiber, but as we have seen this is not the only fine of evidence that can be brought to bear on the problem. All things taken into consideration, the evidence is conclusive that postganghonic axons not uncommonly acquire myelin sheaths.


324 S. W. RANSON AND P. R. BILLINGSLEY

STRUCTURE OF THE SUPERIOR CERVICAL GANGLION

While we have examined a number of ganglia, including the stellate and coeliac, the observations which we have to report are restricted to the superior cervical ganglion. In the account which follows we will consider the results obtained by others, topic by topic, as we present our own. Unless otherwise stated, citations from the literature are applicable to the collateral ganglia and to all the ganglia of the sympathetic trunk. They should not be carried over without qualification to the terminal ganglia. These present special problems and require separate consideration.

Ganglion cells. It is well known that almost all of the neurones in the sympathetic ganglia are multipolar, although there are also a restricted number of unipolar and bipolar cells located near the poles of a ganglion or within its longitudinal fiber bundles, Huber ('99), Like other nerve cells these neurones have but a single nucleus, except in rodents. In the rabbit we have seen many cells with two nuclei. These have been figured and described with a summary of the related literature b}^ Huber ('99), The neurofibrils of the cells of the sympathetic ganglia have been described by a number of authors, including Michailow ('08) and Cajal ('11). The Nissl granules have been described and figured by Carpenter and Conel ('14),

Dendrites. The dendrites of the cells of the sympathetic ganglion may be divided into two chief categories — intracapsular and extracapsular. The former, although presenting great variety in length and form, are all situated beneath the cell capsule. Although these intracapsular dendrites are common in the sympathetic ganglia of man, they are rarely met with in the other mammals. Michailow ('11), in his careful study of the collateral and trunk ganglia in horses, dogs, cats, rabbits and guinea-pigs, described and figures only one form of subcapsular dendrite. These are present on his cells of types II and V. They are short and club-shaped (fig. 8, a). There are usually from five to seven of them and they begin as relatively thick fibers soon going over into bulbous endings. A fiber may divide and end in two such clubs. The expanded ends of these


THE CERVinAL SYMPATHETIC TRUNK 325

dendrites usually contain pigment in large quantities and are sometimes vacuolated. Cajal ('11) does not describe any subcapsular dendrites in the sympathetic ganglia of animals, although they are very prominent in his descriptions and figures of these ganglia in man. But these dendrites were demonstrated in the human ganglia by means of his silver stain which was not used in his earlier studies on animals.

It might be supposed that the use of the newer silver stains would demonstrate their general occurrence in the mammalian sympathetic ganglia, but in pyridine silver preparations of the superior cervical ganglia of cats and dogs we have seen no cells with subcapsular dendrites. This shows that they must be relatively rare here and establishes a very striking contrast between the superior cervical ganglion of man and that of the carnivora.

The intracapsular dendrites reach their highest development in man. Here they give rise to complicated subscapsular formations which were first described by Cajal ('11), whose observations have been confirmed by Marinesco ('06). Both investigators worked with the superior cervical ganglion stained by the Cajal method. Their observations are confirmed by our own observations on the human superior cervical ganglion stained by the pyridine silver method. The account which follows is based on our own preparations, but is in accord with the results of the two investigators who preceded us. The subcapsular dendrites are arranged in a great variety of ways underneath the capsule of the cell from which they take origin. In general they may be said to give rise to two types of complicated intracapsular networks which Cajal has called dendritic crowns and glomeruli.

Figure 4 furnishes a good example of a dendritic crown. Numerous dendrites of varying caliber come off from the cell and run toward the inner surface of the capsule where, with or without branching, they turn to run in the subcapsular space. Here they cross and recross, but do not anastamose, and form an open network more or less uniformly distributed around the cell. In some cases these dendrites can be seen to end in small bulbs


326


S. W. RANSOJSr AND P. R. BILLINGSLEY



THE CERVICAL SYMPATHETIC TRUNK 327

or rings. The long thick process which is seen piercing the capsule in the illustration is probably the axon, although it might be an extracapsular dendrite. In that case one would have to assume that the axon came off from that surface of the cell which has been cut away at the plane of section.

According to Cajal, these dendrites frequently apply themselves against the capsule to terminate on its internal surface or among the satellite cells by pear-shaped thickenings. Sometimes they are more delicate and bend in beneath the capsule to terminate by fine pale extremities. Sometimes they run beneath the capsule in great circles about the cell. The dendritic nest which envelops the cell is easy to distinguish from the ramifications of axons by the greater caliber of its fibers and the rarity of its divisions. Cajal's figures show that the spaces among the subcapsular dendrites contain small cells which he calls satellite cells.

The dendrites which enter into the formation of the glomeruli are also subcapsular, but are usually coarser than those just described. Instead of coming off from all parts of the surface of the cell, they usually arise from a restricted region. Branching repeatedly and interlacing they form a mass of considerable size over which the capsule of the cell is continued. Cajal has shown that the spaces between the dendritic branches are occupied by satellite cells. Following his classification, we may enumerate simple, bicellular, tricellular, and multicellular glomeruli according to the number of neurones the dendrites of which enter into their formation.

The simple glomeruli are formed from the dendrites of one cell. They are short and thick, come off from one side of the cell, and raise the capsule to form a pocket within which these dendrites

Fig. 4 Nerve cell surrounded by dendritic crown from the ganglion cervicale superius of man. Pyridine silver. X 800.

Fig. 5 From the ganglion cervicale superius of man. a, unicellular dendritic glomerulus; b, cell provided only with extracapsular dendrites. Pyridine silver. X 800.

Fig. 6 Tricellular glomerulus from the ganglion cervicale superius of man. Pyridine silver. X 700.


328 S. W. RANSON AND P. R. BILLINGSLEY

branch and intertwine (fig. 5, a). All transition stages are found between the simple glomeruli and the dendritic crowns. When the glomerulus is located on the side from which the axon arises it may be prolonged out for a short distance along the axon, giving rise to a comet-shaped formation.

The glomeruli formed from the dendrites of more than one cell may be called composite glomeruli and are somewhat more complicated than the simple glomeruli just described. The large subcapsular dendrites of two or more cells converge toward each other to form a circumscribed mass of branching and interlacing dendrites. Figure 6 gives a good idea of a tricellular glomerulus, which, along with the three cells, seems to be enclosed in a single capsule. The capsules and subcapsular satellite cells are not well differentiated in pyridine silver preparations, but, according to Cajal, the glomeruli are surrounded by a capsule that separates them from the fiber bundles. The capsule is better defined in the bi- and tricellular than in the multicellular forms.

The fine black fibers seen interlacing with the dendrites in figures 5 and 6 are the branches of axons and will be discussed in another place.

The extracapsular dendrites pierce the capsule and run for longer or shorter distances among the cells, helping to form an intercellular plexus of dendritic and axonic ramifications. The cells of the superior cervical ganglion of the dog and cat are provided almost exclusively with this type of dendrite. Such dendrites are also numerous in this ganglion in man. Here they may come from cells devoid of subcapsular processes (fig. 5, b) or from cells provided with dendritic crowns or glomeruli (fig. 5, a). They are usually coarse fibers and may l^ranch near the cell or may remain unbranched until they leave the section. Often it is possible to trace them much longer distances than is indicated in the figure, but in no case could they be followed to what seemed to be their true termination (fig. 7). Cajal differentiates three types of cells in the human superior cervical ganglion: 1) cells provided exclusively or almost exclusively with subcapsular dendrites; 2) cells provided only with long dendrites, and 3)


THE CERVICAL SYMPATHETIC TRUNK 329

cells provided with both kinds of dendrites. While such a classification facilitates description it must not be supposed that these types are separated by sharp lines of cleavage or that there is any reason to assign them different functions.

In pyridine silver preparations of the superior cervical ganglion of dogs and cats the dendrites have not been \'ery well stained. We could find only exti'acapsular dendrites, but could trace none of them to their termination. Figure 12 gives an idea of how they look when freed from intercellular axonic ramifications. In order to make an intelligent analysis of the functional con


Fig. 7 Cell with long extracapsular dendrites from the human superior cervical ganglion. Pyridine silver. X 400.

nections in the ganglia it is necessary to have a clear idea of the course and termination of these extracapsular dendrites.

Concerning the true endings of these dendrites our preparations, which could not be made very thick, give us no information because all the long dendrites seem to be cut off at the surface of the section. According to Cajal Cll), there are three ways in which these long intracapsular dendrites in the human superior cervical ganglion end: 1) They may run into a fascicle of dendritic fibers where they run parallel to the other fibers of the fascicle and within which they m.ay end with long interstitial appendages. At other times they end in oli\'e-shaped extremities, or in fusiform swellings which give rise to fine varicose branches.


330 S. W. HANSON AND P. E. BILLINGSLEY

2) They may end in glomeruli ^Yhe^e they encounter the branches of other dendrites of the same kind. As indicated in his figures, such glomeruli are located at a distance from the cells of origin of the dendrites concerned. 3) They may end in pericellular baskets. These dendritic baskets have been found in animals by Cajal ('11), Van Gehuchten ('90), Sala ('92), and Michailow ('11), and will be discussed more in detail in connection with the account given by the latter author.

Michailow ('11) has enumerated nine types of cells in the sympathetic ganglia of mammals. This grouping like that of other authors is chiefly of value as an aid to description, since there is no evidence that any one type is responsible for a particular function. From among the various forms, which, according to him, the dendrites of the cells in the sympathetic ganglia may assume, we have selected five as the most typical and significant. Such dendrites may be found in the ganglia of the sympathetic trunk as well as in the collateral and terminal ganglia. They are represented in figure 8.

1. Dendrites ending in a brush formation (fig. 8, a). These are given off in small numbers (1 to 4) from Michailow 's Type II cells. They run between the cells of the ganglion where some of them end; others enter bundles of fibers that leave the ganglion. He has followed such a dendrite from a ganglion of the solar plexus of the horse and seen it run as a typical unmyelinated fiber into another ganglion of the same plexus. These dendrites end in special formations in the shape of little brooms, consisting of numerous end branches beset with enlargements. These thickenings are of various shapes and sizes. Usually they are flattened and have the appearance of end plates or of large varicosities.

2. Dendrites terminating in end plates (fig. 8, 6). These are given off from Michailow's Type III cells. They begin as rather thick processes which in unipolar and bipolar cells may be so thick that it is hard to tell where the cell body ends and the dendrite begins. Sometimes these dendrites end in the same ganglion, sometimes they join bundles of nerve fibers and either end in them or run with them to end in other ganglia. Some remain thick and coarse to their end, others branch and become


THE CERVICAL SYMPATHETIC TRUNK


331



Fig. 8 Symi)athetic ganglion cells showing various types of dendrites. Redrawn from Michailow ('11). All were stained with methylene blue; a, cell of Michailow's Type II from the ganglion mesentericus superius of the horse; h, cell of Type III from the ganglion coeliacum of the horse; c, cell of Type IV from the ganglion stellatum of the horse; d, cell of Type \l from the ganglion cervicale superius of the dog; e, cell of Type IX from the ganglion coeliacum of the horse; /, cell of Type VIII from the ganglion cervicale superius of the dog.


332 S. W. RANSON AND P. R. BILLINGSLEY

thin, take on the character of unmyelinated fibers and run out of the gangUon, The endings are in the form of plates of various sizes and shapes. These may lie free in the connective tissue or may be pressed against the outside of the capsules of other cells so close as to produce an impression on the cells. Other end plates of this type are found in the fiber bundles outside the ganglia. He found great numbers of such end plates in the fiber bundles of the solar plexus.

3. Dendrites ending in a number of fine branches with end bulbs closely grouped together as illustrated in figure 8, c. Such dendrites arise from Michailow's cells of Type IV. They branch freely and occupy much space, greatly increasing the territory of these neurones. They may end in the same or in other ganglia. Near their termination they begin to divide di- and tricotomously. The branches are provided with terminal enlargements which may be rounded or pear-shaped. All the branches of a dendrite form together an end-apparatus, which may vary in size and appearance, but is always applied to the outer surface of the capsule of a cell of Type IV. That is to say, these fibers arise from cells of Type IV and end upon the surface of the capsules of other cells of Type IV.

4. Dendrites forming pericellular nests (fig. 8, d). These arise from the cells of Michailow's Type VI, are usually short, and divide repeatedly. The branches approach another cell, and anastamosing with each other form a network that encloses the cell. Sometimes such a basket-like network surrounds the cell from which the dendrite arose. Similar formations have been described by Dogiel, according to whom they are always extracapsular. As already mentioned, Cajal, Van Gehuchten, and Sala have seen such dendritic nests. The significance of these structures can best be discussed in a later paragraph.

5. Dendrites the branches of which anastomose to form a true net out of which a fine fiber, probably the axon, arises (fig. 8, /). One or more dendrites break up into a great number of fine branches which anastomose with each other, giving rise to a network. Out of the net fine filaments arise, which join together to form a smooth fiber that remains unaltered as far as it


THE CERVICAL SYMPATHETIC TRUNK 333

can be followed. Michailow thinks it probable that these smooth fibers are axons. It will be seen that these neurones resemble some described by Dogiel in the spinal ganglion.

Are there special sensory dendrites in the sympathetic ganglia? This problem has been in the foreground ever since 1896 when Dogiel published his paper on "Zwei Arten sympathischer Nervenzellen." The one type of dendrite which he thought belonged to motor cells branched repeatedly in the neighborhood of the cell and ended within the ganglion ; the other, which he thought belonged to sensory cells, resembled unmyelinated nerve fibers and could be traced long distances. Many of them could be followed out of the ganglion and were thought to end as sensory fibers in the viscera. Cajal ('11) finds no evidence in favor of the sensory function of these long dendrites and was not able to find any of them leaving the ganglia and associated nerve trunks to end in the viscera.

Carpenter and Conel ('14), working with Cajal's method on the superior cervical ganglion of the cat, could find cells answering to the description of Dogiel's two types, but were not convinced that such cells represent two distinct categories, since all gradations between the two extremes were found. In Nissl preparations all the cells of the sympathetic ganglia appeared to Carpenter and Conel to be of one type. In the cerebrospinal system it is easy to recognize sensory and motor cells by the arrangement of their chromatophile substance, but all the sympathetic ganglion cells seemed to have a structure intermediate in character between that of the cerebrospinal sensory and motor types. Since these results would indicate that there is but one functional type of cell in these ganglia and since we know that the majority of the cells are motor, the probability against the presence of sensory cells is increased.

So far as we have been able to find no one has confirmed Dogiel's account of the sensory type of cell except Kuntz ('13), who found certain structures which could be interpreted in this way. Nor has the correlated observation of Dogiel, that fibers, arising from sensory cells in the sympathetic ganglia, run to end in pericellular baskets about spinal ganglion cells, been much better


334 S. W. HANSON AND P. R. BILLINGSLEY

supported. In regard to this point Huber ('13) has recently said

the evidence presented by Cajal, Dogiel, Retzius, Huber, and others cannot be regarded as entirely conclusive, since it has not been determined that the fine medullated fibers or the unmedullated fibers which appear to enter the spinal ganglia from without and end in pericellular plexuses, are, in fact, the neuraxes of sympathetic neurones.

Very strong evidence has been presented by Langley ('03) to show that no medullated sensory fibers run from the s;yaripathetic to the spinal ganglia.

As regards the white rami, which contain most of the afferent visceral fibers, there is conclusive evidence that the very great majority of them have their trophic center in the posterior root ganglia. It consists in the fact that after intraspinal section of a nerve just peripherally of the posterior root ganglia, either all, or all but a few, of the medullated fibers in the white rami degenerate; and that after section of the sympathetic or of the splanchnic or of the inferior splanchnics no degenerated fibers are present in the white rami.

Similarly in the sacral autonomic system, the pelvic nerves contain about 1,000 afferent nerve fibers, and about twice this number of efferent nerve fibers; on cutting the roots of the sacral nerve, as shown by Anderson and myself, about half a dozen fibers only remain undegenerated in the pelvic nerve, and these are probably post-ganglionic medullated fibers.

Axons of the cells of the sympathetic ganglia. In pyridine silver preparations of the superior cervical ganglia of the cat, dog, and man, it is very difficult to follow an axon for any considerable distance. In fact, it is usually no easy matter to tell which of the several processes of a cell is to be regarded as an axon. In a preceding section of this paper it has been show-n that some of these axons acquire a myelin sheath. According to Kolliker ('96) and Langley ('00), these axons always end at the periphery, and never terminate in the sympathetic ganglia.

According to Cajal ('11), who worked with the Golgi and methylene blue stains on the sympathetic ganglia of animals and with his silver stain on the superior cervical ganglion of man, the axons of the cells of the sympathetic ganglia are rather thick, smooth, and devoid of branches. He says that his anatomical studies are in accord with the physiological experiments of Lang


THE CERVICAL SYMPATHETIC TRUNK 335

ley and indicate that the axons of these cells dispose themselves in one of the three following ways: 1) Usually they run transversely to the long axis of the ganglion to enter a gray ramus. In the initial part of their course these fibers do not give rise to branches. 2) The axons may run through a connecting nerve trunk into another ganglion. He is not able to say whether these axons only run through the second ganglion or whether they make connections with its cells. In the chick embryo he at one time described collaterals coming from those longitudinal fibers of the ganglia which take origin in neighboring ganglia. He is now inclined to doubt this observation and thinks it likely that these collaterals all come from fibers that have entered the sympathetic trunk through white rami at other levels. 3) In some cases they leave the ganglion and run toward the neighboring arteries in the visceral nerves.

Sala ('93) described two kinds of fibers in the sympathetic ganglia. Those of one variety are unbranched, varicose, and unite to form smaller or larger fascicles which run through the ganglion in every direction. These are the axons of the cells of the sympathetic ganglia. The fibers of the other kind are a little larger, non- varicose, and give off collaterals which are finer and in their turn ramify abundantly. These are less numerous than the first and are found almost exclusively in the branches from the cerebrospinal system. It is not improbable, he says, that these are of cerebrospinal origin.

In his elaborate description of nine types of cells in sympathetic ganglian jVIichailow has given very few details regarding the axons. However, it is to be noted that in none of these nine types does he describe the axon as terminating in a sympathetic ganglion and in only one does he describe it as giving off collaterals (fig. 8, e).

v. Lenhossek ('94), using Golgi preparations of the chick, traced axons of sympathetic ganglion cells into the neighboring ganglia, but did not say what became of them there. In one case he saw fibers entering a ganglion from a visceral nerve break up into branches. He considered these the axons of cells lying somewhere in the visceral ganglia. From what we know now they might just as well be interpreted as the endings of long dendrites.


336 S. W. RANSON AND P. R. BILLINGSLEY

The axons of Dogiel's Type II cell are figured by that author as passing through several ganglia giving off collaterals and finally ending by branching in another ganglion. In the text, however, he does not claim to have followed such an axon to its termination. But, as we have said before, no one has been able to confirm Dogiel's findings in regard to these cells.

Both Dogiel ('95) and Huber ('99) are of the opinion that the fine fibers which enter the ganglion through its various branches and take part in the formation of the intercellular plexuses are the axons of cells in other sympathetic ganglia. Satisfactory evidence of this is not presented, however, and in the next section of this paper we will present what seems to be conclusive evidence that these fine fibers are of cerebrospinal origin.

While it has not been shown that the axons of sympathetic ganglion cells ever end in connection with the cells of the same or adjacent ganglia, it seems to be well established that these axons may give off collaterals within these ganglia. The axons have been seen to give off collaterals either in the same or adjacent ganglia by v. Lenhossek ('94), Dogiel ('95), and Michailow ('11). These do not seem to be present on the majority of the axons. Michailow is the only one who has seen the mode of termination of these collaterals. According to him (fig. 8), they end in little plates, either in the connective tissue of the ganglion between the nerve cells or pressed against the capsule of a cell. From their mode of termination it is not evident how these collaterals could serve to transmit impulses from one neurone to another. They rather resemble certain collaterals on the axons of spinal ganglion cells, seen by Huber, Dogiel, and Ranson, which since many of them end on the cell from which the axon arose cannot serve for the spreading out of nerve impulses.

Huber ('13), in summing up the evidence concerning the interconnections of the cells of the sympathetic ganglia, concludes that "there is at hand morphologic evidence that the neuraxes of sympathetic neurones, the cell bodies of which are in one ganghon, terminate either on the cells of the same ganglion or of other ganglia." To us the evidence seems far from convincing. Such fragmentary and unsatisfactory histological evidence as


THE CERVICAL SYMPATHETIC TRUNK


337


may exist is more than offset by the strong physiological evidence against such connections. Some of this physiological evidence will be briefly presented in a succeeding paragraph.

The intercellular plexus. Throughout the ganglion there is a rich plexus of dendritic branches and fine axons. This has been described and figured by Dogiel ('95), Huber ('99), and Michailow



Fig. 9 Intercellular plexus formed by dendrites and myelinated and unmyelinated fibers from the semilunar ganglion of the cat. Redrawn from Huber ('99).


('11). The part which the dendrites take in this formation has been discussed in a preceding section. We are interested here chiefly in the axonic ramifications which help to constitute it. According to Huber, one of whose drawings is reproduced in figure 9, there are in addition to the medullated fibers entering the ganglion from the white rami, "small medullated fibers, which may be traced from this or that nerve root of a ganglion" into


THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL. 29, NO. 4


338 S. W. RANSON AND P. R. BILLINGSLEY

the ganglion where they are found branching and rebranching, and forming, with the dendritic processes of the ganglion cells, what Dogiel has described as the intercellular plexus." Huber quotes with approval the conclusion of Dogiel ('95) : "Die feinen Fasern, welche in den Ganglien mit intercellularem Geflechte endigen, zu den sympathischen, augenscheinlich ^'orzugsweise niarkhaltigen Fasern gehoren." It is interesting to note that Huber was able to trace some of the fine unmyelinated fibers of this plexus to definite endings on neighboring dendrites.

According to Dogiel ('95), whose observations were made on the terminal ganglia, the finer myelinated and unmyelinated fibers enter the ganglion, branch and intertwine, and break up into fine branches which cross in ^'arious directions and finally break up into finer fibers of uncountable number. These form a thick plexus among the cells and at the periphery of the ganglion. The fibers of the plexus are in contact with the dendrites, but separated from the cell bodies by their capsules. All the fibers of the plexus are beset with varicosities.

Michailow's ('11) conception of the intercellular plexus differs from that of the two preceding authors in that, according to him, the constituent fibers of the plexus anastomose with each other forming a closed network. By means of this network all or at least many of these fibers are united together, one neurone being in this way united with many others. As will be seen later, there are good reasons for discarding this part of Michailow's description of the intercellular plexus.

In preparations of the superior cervical ganglion of the cat or dog by the pyridine sih'er method one can readily see a plexus of fine unmyelinated fibers running among the cells in every direction through the ganglion (fig. 10). The dendrites are not well stained in these preparations and only their coarser branches are visible. The finer dendritic ramifications, which, according to those who have worked with the methylene blue stain, help to form the intercellular plexuses, are not to be seen. In these preparations the network of fibers under discussion corresponds only to the axonic constituents of the intercellular plexuses of Dogiel and Huber.


THE CERVICAL SYMPATHETIC TRUNK


339


The constituent fibers of this plexus stain rather heavily with silver and range in color from light brown to black. They also vary greatly in size, the smallest being perhaps not more than one-eighth the thickness of the largest. The larger axons can often be seen to branch, but the smaller ones seem to run for


^^..-^0




Fig. 10 Intercellular plexus in the ganglion cervioale superius of the do^ Section 20ju. Pyridine silver. X 800.


considerable distances without branching. The majority of the fibers are very fine. They run in and out among the cells, twisting and turning, crossing and recrossing and forming together a dense interlacement. That practically all of these fibers are unmyelinated can be seen at once by comparing such a prepara


340


S. W. RANSON AND P. R. BILLINGSLEY


tion with one stained by osmic acid. In the latter, in place of the dense interlacement of fine fibers just described, one sees only here and there an isolated myelinated fiber. In many parts of the ganglion these are less numerous than the nerve cells. So far as we can determine the intercellular plexus is entirely extracapsular. Although some of the fibers wrap themselves about the cells and form what might seem to be pericellular plexuses (fig. 11), these are found not to be in any way separated from the general plexus which fills in the intervening spaces. We believe that these apparently pericellular baskets are really pericapsular and represent merely portions of the general plexus which are in contact with the cell capsules. It is not clear



Fig. 11 Three cells from the ganglion cervicale superius of the dog showing fibers of the intercellu ar plexus wrapped about them. These fibers seem to be extracapsular. Pyridine silver. X 800.

whether these formations correspond to the pericapsular nets of Michailow or not. It is evident, however, that they do not correspond to Ruber's pericellular plexus which is endocapsular and forms a closed network. That all of the fibers of the intercellular plexus are extracapsular is shown by the examination of sections in which the ganglion cells have shrunken, leaving a cleft between them and their capsule. In such cases the fibers in question always remain in or upon the capsule and never lie on the shrunken cell. An additional point of distinction is found in the fact that the pericellular plexus is a closed network while as we shall see anastamoses do not seem to occur among the fibers under discussion. Furthermore, we have twice seen a fragment


THE CERVICAL SYMPATHETIC TRUNK 341

of a closed network on the surface of a cell which we were inclined to regard as a true pericellular plexus. From all these facts we conclude that the pericellular network is ordinarily not stained in the pyridine silver preparations, but that the axonic constituents of the intercellular plexuses come out with great clearness.

It will be remembered that Michailow regarded the plexus under discussion as forming a true net by means of which all or at least many of the fibers are united together, one neurone being thus united to another. This would mean diffuse conduction in the ganglion, which must then act as a whole. This is directly at variance with what is known of the physiology of the sympathetic ganglia. There is no evidence that diffuse conduction occurs in any of them, and in at least two, the superior cervical and the coccygeal ganglia, Langley ('00 a, '04) has been able to show that diffusion does not occur. We will take this up in connection with a discussion of the synapses in the sympathetic ganglia.

Neither Dogiel nor Huber gives the impression that the intercellular plexus is a closed net and we have carefully examined pyridine silver preparations for evidence in this regard. While branching fibers are common, it can usually be seen that a larger fiber is dividing into two smaller ones. The junction of three fibers of the same size as at the nodal point of a net does not seem to occur. Often two fibers could be seen crossing, one immediately over the other, but each retained its individuality and sharp contour. If the plexus were a true network, one should be able to find closed meshes surrounded on all sides by anastamosing fibers — an arrangement which does not seem to occur.

In the pyridine silver preparations of the human superior cervical ganglion the fibers of the intercellular plexus stand out prominently, as is seen in figure 5. The fibers of this plexus mingle with the branches of the long or extracapsular dendrites. Except that the plexus is not as uniformly distributed throughout the ganglion and is perhaps not quite so dense, it resembles that in the dog. There is, however, one important feature in which the human ganglion differs. Fine axons, apparently continuous


342 S. W. RANSON AND P. R. BILLINGSLEY

with those of the intercelhilar plexus, penetrate into the denckitic glomeruli and the dendritic crowns forming subcapsular plexuses in close relation to the subcapsular dendrites. This is well illustrated in figure 5, a. Cajal has considered these fine darkly staining axons as preganglionic fibers of spinal origin. From what has been said and from the accompanying illustrations it will readily be seen that these fibers are the same as those which form the intercellular plexuses in the superior cervical ganglion of the dog and cat. In a paragraph which follows evidence will be presented to show that these are fibers of spinal origin.

Here and there in this plexus in the superior cervical ganglion of the dog or cat one can see faintly stained yellow axons about the size of the larger dark fibers forming the plexus. In many places these lightly stained axons are united into bundles of parallel fibers which run as straight a course as is possible through the ganglion. These light yellow axons and the bundles into which they unite do not seem to belong to the plexus, although necessarily they run through it. The color contrast between the two kinds of fibers is quite sharp in good preparations, but since all gradations are found the color alone is not sufficient to distinguish them. The light axons are among the largest in the ganglion, are of uniform contour and apparently unbranched. They show a marked tendency to group themselves into bundles of parallel fibers in contrast with the more irregular course of the dark fibers.

Distribution of nerve fibers in the ganglion. In regard to the termination of axons in the sympathetic ganglia, Cajal (Tl) states that in his first work in this field he described two kinds of terminal arborizations, one set representing the branches of the longitudinal sympathetic fibers arising in neighboring ganglia, the other representing branches given off by fibers from the white rami. That distinction does not seem probable to him any longer because of the results of Langley's experiments and because of the presence of many medullated fibers in the commissural cords which are known to come from the spinal cord (Langley, '03, and Miiller, '09). Cajal now believes that the two kinds of terminations belong to spinal motor fibers, distinguished


THE CERVICAL SYMPATHETIC TRUNK 343

onl}^ by their course in the sympathetic trunk. One innervates the gangHon to which the ramus brings it. The other runs through two or more gangha before it terminates.

A study of serial sections through the superior cervical ganglion of the cat stained with osmic acid is instructive. At the lower pole a large bundle of myelinated fibers can be traced into the ganglion from the sympathetic trunk. This comes to lie near the center of the ganglion and breaks up into smaller bundles. Many of the fibers seem to lose their myelin sheaths while still within the smaller bundles. At least this seems to be the best explanation of the fact that the number of myelinated fibers scattered among the ganglion cells is so small.

When the sympathetic trunk is cut and time allowed for degeneration all these bundles of fibers have degenerated. There are, however, still present even in the caudal pole of the ganglion a very few scattered myelinated fibers which have their cells of origin in the ganglion. The number of such fibers increases toward the cephalic pole. Here the myelinated and the more numerous unmyelinated postganglionic fibers accumulate in bundles located especially near the periphery of the ganglion. From the pole large branches representing the internal carotid nerve are given off. Other smaller branches are given off in various places from the ganglion.

The small number of myelinated fibers which are scattered among the ganglion cells in comparison to the number entering and leaving the ganglion would indicate that they run considerable distances in the ganglia as unmyelinated fibers.

In following through a series of sections stained by the pyridine silver technique, one sees that the fine axons entering the ganglion from the sympathetic trunk are all stained a dark brown. Each fiber is surrounded by a thin unstained ring of myehn. This central bundle of the ganglion can be seen to break up into smaller and smaller bundles of dark fibers and the constituent fibers of these smaller bundles can be seen to run into and become a part of the intercellular plexus described in the preceding section.

Following the series toward the cephalic end of the ganglion, one sees bundles of axons collecting especially near the periph


344 S. W. RANSON AND P. R. BILLINGSLEY

ery of the ganglion and these can be followed into its various branches of distribution. These fibers are stained yellow or light brown in contrast to the darker fibers entering by way of the trunk. The staining reaction of these axons is exactly like that of the bundles of 'sympathetic fibers' described in the vagus and its branches by Chase and Ranson ('14) "where they are differentiated from the vagus fibers by their lighter stain." We have repeatedly noticed this characteristic light staining of postganglionic autonomic fibers in the various spinal and cerebral nerves. Here the contrast with the darker unmyelinated fibers of cerebrospinal origin could not easily be overlooked.

It is true that these lightly stained axons run among the cells and therefore through the intercellular plexus, but the great bulk of that plexus is composed of fibers whose staining reaction resembles that of the fibers entering by w^ay of the sympathetic trunk. And the impression is gained by a study of such serial sections that this intercellular plexus is formed by the fibers derived from the trunk, and that the other fibers run through the plexus as directly as possible to their point of exit from the ganglion. Were it not for the difference in the color of the two kinds of axons, however, the impression would undoubtedly be given that the plexus is formed by fibers that stream into the ganglion through all its branches. This is the impression that Dogiel, Huber, and Michailow have gained from the study of methylene blue preparations.

The proof that the intercellular plexus is formed by the ramifications of the preganglionic fibers is furnished by the experiment of cutting the sympathetic trunk in the neck and allowing time for degeneration to take place. Pyridine silver preparations of the superior cervical ganglion in which the preganglionic fibers have degenerated show no trace of an intercellular plexus (fig. 12). Our technique does not stain the finer branches of the dendrites and these do not appear in either the normal or altered ganglia, but the fine axonic ramifications that form the normal network are gone. One can readily recognize small bundles of the lightly staining postganglionic fibers and many such fibers running an isolated course. But these fibers do not coil and


THE CEEVICAL SYMPATHETIC TRUNK


345


turn about the cells, and wherever several are grouped together they run parallel to each other in small compact bundles. They do not give in any way the appearance of the intercellular network.

By way of summary we may say that the fine myelinated fibers entering the ganglion through the sympathetic trunk are preganglionic elements and form by their ramifications a complicated intercellular plexus of fine unmyelinated fibers. The other branches of the ganglion consists of many unmyelinated and a few myelinated fibers. These all represent the axons of



Fig. 12 Three cells from the ganglion cervicale superius of a dog in which the sympathetic trunk had been cut 58 days before the dog was killed. The fine fibers of the intercellular plexus are absent. Pyridine silver. X 800.


the cells in the ganglion and take no part in the formation of the intercellular plexus. They are the postganglionic fibers of Langley.

SYNAPSES IN THE SUPERIOR CERVICAL GANGLION

Where are the synapses on the paths through the superior cervica ganglion located? Langley ('00), using his method of paralyzing the endings of preganglionic fibers by nicotine, has shown that the fibers of the sympathetic trunk, destined for the*" superior cervical ganglion, come from the upper thoracic white rami and run without interruption through the upper thoracic ganglia.


346 S. W. RANSON AND P. R. BILLINGSLEY

By the same method he has shown that all these fibers end in the superior cervical ganglion. After painting that ganglion with a solution of nicotine no response can be obtained on stimulation of the upper thoracic nerves, showing that all the pathways through the ganglion are blocked. It is generally admitted that this blocking occurs at the synapse. The same effect can be obtained by the intravenous injection of nicotine. Since, however, large doses of nicotine given intravenously will not eliminate the effects of stimulating the internal carotid nerve or other branches of distribution from the ganglion, it is argued that there are no other synapses interposed between this ganglion and the tissues innervated. This conclusion is shown to be correct by the results of the method of degeneration.

That the degeneration, after section of the internal carotid branches, spreads to the periphery, is shown by stimulating the sclerotic before and after degenerative section. In the former case, there is a double effect — local contraction of the radial muscle leading to local enlargement of the pupil, and local contraction of the circular muscle of the iris; in the latter case, the i-adial contraction is lacking, the circular takes place as before.

The results obtained from section of the sympathetic trunk in the neck and of the internal carotid nerve are all in accord with the conclusions to be deduced from the nicotine experiments.

Our own observations are in full agreement with the conceptions just presented. The trunk consists almost exclusively of medullated fibers, which would not be the case if it contained postganglionic fibers ascending from the thoracic ganglia. All, with the exception of a small bundle of unmyelinated fibers, degenerate in an ascending direction and the degeneration stops in the superior cervical ganglion. The internal carotid nerves are not affected either as to their myelinated or unmyelinated constituents. The conclusion that the only synapses on the functional pathways through the superior cervical ganglion are located in that ganglion is well established. We may now ask what is the nature of the synapses which are to be found there.

Is there a mechanism within the ganglion for the general diffusion of impulses such as occurs in the central nervous system? As a result of the diffusion of impulses in the brain and spinal cord the


THE CERVICAL SYMPATHETIC TRUNK 347

stimulation of a small sensory nerve may bring about reflex activity of the skeletal and involuntary musculature over the entire body. Are impulses disseminated in a similar way in the sympathetic ganglia? Langley ('00) maintains that a preganglionic fiber branches and becomes associated with several postganglionic neurones and that these taken together form a functionally isolated unit. That is to say, there is no general diffusion of impulses through the ganglion. This is beautifully illustrated by his experiments on the pupilodilator pathway.

As pointed out by Hoffmann ('04), the stimulation of a long ciliary nerve causes local dilation of the pupil, while sthniilation of the white ramus of either the first or second thoracic nerve causes a general and symmetrical dilation. This might appear to be due to a spreading of the impulses within the superior cervical ganglion to all postganglionic pupilodilator neurones. This is not the case, however, as Langley ('04) has shown: 1) Because stimulation of a small number of postganglionic fibers as they leave the ganglion in any one of the four bundles that form the internal carotid nerve will also cause a symmetrical general dilation. Fibers from such a bundle undergoing rearrangement in the internal carotid plexus are distributed to all parts of the iris. It is therefore unnecessary to assume any spreading out of nerve impulses through diffusion in the ganglion. 2) Local dilation of the pupil can, on the other hand, be obtained by stimulating a few preganglionic fibers in one of the rootlets of the upper thoracic nerves. It is difficult to see how, on any theory of the cells being connected together to form a coordinating center, stimulation of a small number of preganglionic fibers could cause rather marked local dilation of the pupil. The spreading out of the impulses which does occur is due to the intermingling of the postganglionic fibers in the preterminal plexuses.

An even more striking case has been made out against the general diffusion of nerve impulses within sympathetic ganglia in the case of the coccygeal ganglion.

In all compound ganglia it is obvious that stimulation of certain of the preganglionic fibers running to the ganglia excites some only of the nerve cells, and no increase in the strength of the stimulus can cause


348 S. W. RANSON AND P. R. BILLINGSLEY

irradiation of nervous impulses to other cells of the ganglion. And the nerve cells which cannot then be brought into action may be nerve cells of the same class as the cells which are in a state of excitation. Of this we may give an example. In the cat, at times, when the arrangement of nerves is posterior, the fourth lumbar nerve causes erection of hairs on the tip of the tail; the nervous impulses pass through nerve cells in the coccygeal ganglion; other nerve cells in the coccygeal ganglia will, on stimulation cause erection of hairs in the greater part of the rest of the tail; but no stimulation of the fourth lumbar nerve will affect this region. Hence, pilomotor nerve cells, set in action by the fourth lumbar nerve, send no commissural fibers to the other pilomotor nerve cells of the coccygeal ganglion. (Langley, '00.)

It thus appears that there is no physiological evidence indicating that diffusion of nerve impulses occurs in the sympathetic ganglia and in certain cases, like those cited, there is positive evidence that diffusion does not occur. We shall now see that there is no histological evidence of any mechanism which could serve to bring about such diffusion.

We may picture such a diffusion mechanism in three ways. The first that suggests itself is a diffuse network formed by anastomosing branches of the preganglionic fibers. Such a network has been assumed by ]Michailow ('11), but without adequate evidence. In this respect his description of the intercellular plexus does not coincide with that given by Dogiel and Huber. Very clear pictures of the intercellular plexus are obtained in pyridine silver preparations, and these give no indication of anastomosing fibers or of a closed netw^ork. The histological evidence is therefore distinctly against the existence of this sort of mechanism for diffusion of nerve impulses.

In the second place, the purpose of diffusion might be served by purely intraganglionic neurones w^hose axons would branch repeatedly and end within the ganglion. So far as we have been able to find, no one has ever described an axon of a sympathetic ganglion cell as ending within the ganglion where it began. Wherever axons have been traced they have always been seen to leave the ganglion through one or other of its branches. The intercellular plexus of fine fibers, which Dogiel and Huber thought represented the ramifications of such axons, and which, if interrupted in this way, might serve as a mechanism for diffusing


THE CERVICAL SYMPATHETIC TRUNK 349

nerve impulses through the ganghon, we have shown to be formed by the branching of the preganghonic fibers. In a paper which follows, Johnson presents conclusive evidence that commissural neurones do not exist in the ganglia of the sympathetic trunk of the frog.

Finally, diffusion of nerve impulses might occur through collaterals given off by the postganglionic axons before they left the ganglion.

That such collaterals exist has been shown b}^ Dogiel, but we must conclude from his descriptions and figures that they do not occur on the majority of the axons. Michailow does not find them except on the axons of his cells of Type IX. He shows that they end in plates located in the connective tissue of the ganglia between the nerve cells or against the outside of the capsule of a nerve cell. This mode of termination does not speak for them as serving the function of transferring impulses from one neurone to another. In fact, they rather resemble certain collaterals from the axons of spinal ganglion cells which in all probability serve no such function.

The complete absence of fine branching axons in the superior cervical ganglion after degeneration of the preganglionic fibers is strong evidence against the existence of connections between the various cells of the ganglia. In such a ganglion the postganglionic axons can be seen to accumulate in bundles of parallel fibers and run as directly as possible toward the emerging nerves.

From all that has been said we may conclude that there is no physiological or histological evidence for the existence in the superior cervical ganglion of a mechanism for the general diffusion of nerve impulses. And the same conclusion would probably be equally valid for all the ganglia of the sympathetic trunk. We have already discussed the question of commissural fibers joining cells located in adjacent ganglia.

Are there any synapses between sensory and motor neurones within the superior cervical ganglion such as would be required by the conception of the ganglion as a center for visceral reflexes? So far as we have been able to learn, no one has ever described any reflex through this ganglion. According to Langley ('00 a),


350 S. W. RANSON AND P. R. BILLIXGSLEY

there are no sensory fibers in the cervical sympathetic trunk, since stimulation of this trunk produces no reflex effect through the spinal cord. Since no one has ever claimed that this ganglion contained sensory elements, it is not necessary to discuss this question in detail here. The question of the presence of sensory neurones in the sympathetic ganglia was discussed at some length in the section of this paper dealing with the dendrites. The negative evidence (the absence of fine branching axons in the ganglion after degeneration of the preganglionic fibers) which indicated the absence of connections between the sympathetic ganglion cells would also speak against the existence of sensorymotor synapses.

Synapses between pre-, and postganglionic neurones are the only ones of which physiological experiments have given evidence. These are also the only ones that have been demonstrated histologically. The clearest demonstration has been given by Huber ('99) on the frog (fig. 13 and 14). In preparations stained with methylene blue he was able to trace the fibers of the white rami into the trunk ganglia and see them divide repeatedly. Some of these branches he was able to follow to their termination as subcapsular pericellular baskets. In a well stained ganglion it could be seen that the cell body of each neurone was enclosed in such a pericellular plexus. As a rule, the fibrillae of the plexus form a closed network, but now and then fibrillae were found ending free. Similar pericellular plexuses were observed by him in the trunk ganglia of mammals and here again the evidence pointed to their being the endings of fibers from the white rami. These pericellular plexuses have been seen by others, including Ehrlich ('96), Retzius ('89), Arnstein ('87), Aronson ('86), Sala ('93), Van Gehuchten ('92), v. Lenhossek ('94), Dogiel ('95), and KoUiker ('96).

Dogiel ('95) and Huber ('13) could not determine whether all or only a part of the cells of a sympathetic ganglion were surrounded by pericellular plexuses. I take these statements to refer to the mammalian ganglia since Huber ('99) has himself shown that all these cells are so surrounded in the frog. For a full account of this form of synapse the reader is referred to


THE CERVICAL SYMPATHETIC TRUNK


351


Huber's three papers. It seems that the pyridine silver method usually does not stain these pericellular networks; only occasionally have we seen fragmentary impregnations of them. This is in keeping with the fact that the method does not readily yield pictures of nerve endings.

In addition to the pericellular endings thus described there are, we believe, synapses between preganglionic fibers and the dendrites of the cells in the superior cervical ganglion. This is true



Figs. 13 and 14 Preganglionic fibers and pericellular plexuses of the frog. Redrawn from Huber ('99). The preparations were stained with methylene blue. 13, preganglionic fibers, the branches of which form pericellular plexuses; 14, a sympathetic ganglion cell, uni])olar, in connection with which a preganglionic fiber is terminating.


of the subcapsular dendrites in man as well as of the long extracapsular dendrites of man and the dog and cat. As was first shown by Cajal in the superior cervical ganghon of man, the subcapsular dendrites forming glomeruli and dendritic crowns are in close relation to fine, darkly staining fibers, which run among them in every direction. This is illustrated in figure 5 and 6. These fibers have the same appearance, caliber, and


352 S. W. RANSON AND P. R. BILLINGSLEY

staining reaction as the fine fibers of the intercellular plexus in the cat and dog, and they bear the same relation to these subcapsular dendrites that that plexus bears to the extracapsular dendrites. There is every reason to believe that these fibers, Uke those of the intercellular plexus, are the branches of preganglionic axons. There seems to be no essential difference between the intercellular plexus in man and that which surrounds the subcapsular dendrites except that of location. So far as we are able to judge from our preparations, the intercellular plexus is not so well developed nor so uniformly distributed in man as in the dog. In the cat and dog there are almost no subcapsular dendrites, and so far as we have been able to see the intercellular plexus does not extend beneath the capsule.

We have already given a somewhat extended account of this intercellular plexus and shown that it consists of the ramifications of preganglionic axons. Just what is the relation of the ramifications to the dendritic branches? In pyridine silver preparations the fibers do not seem to end on the dendrites, but rather to form an interlacing feltwork with them. It is probable, however, that here the actual terminations of the axonic ramifications are not stained. In methylene blue preparations Huber ('99) was able to trace some of the fine fibers of the pericellular plexus to their termination on neighboring dendrites.

It seems to be well established that one preganglionic fiber may activate several postganglionic neurones (Langley, '00 b). Histological evidence points to three ways in which this can be brought about: 1. The branching of preganghonic fibers, each branch ending in a pericellular basket about a different neurone. The best evidence of this has been given by Huber ('99). Figure 13 is a reproduction of one of his drawings of fibers from a white ramus entering a sympathetic ganglion of the frog. One of these fibers is associated with three pericellular plexuses. This mechanism for bringing several postganglionic neurones under the control of one preganglionic fiber is illustrated diagrammatically in figure 15, b.

2. The ending of dendrites of one cell in the neighborhood of another cell so as to come under the influence of the axonic ramifications in connection with that cell. This relationship is


THE CERVICAL SYMPATHETIC TRUNK


353


illustrated diagrammatically in figure 15, c. The ending of dendrites of one cell in the immediate neighborhood of another cell has been observed by a considerable number of investigators. Such endings occur in a variety of different forms which can scarcely be accidental. A dendrite may end by forming a pericellular basket about another cell as seen inj^figure 8, d. Such formations have been seen by Cajal ('11), Dogiel (^95),



Fig. 15 Diagram illustrating three ways by which one preganglionic fiber may come into relation with two or more postganglionic neurones, a, preganglionic fibers ending in a tricellular glomerulus in connection with the dendrites of three neurons; h, a preganglionic fiber branching to form two pericellular plexuses; c, a preganglionic fiber ending in connection with the cell body of one neurone and the dendrite of another which is applied to the outer surface^of the capsule of the first neurone.

Michailow ('11), and others. According to'^Dogiel, such dendritic baskets are always extracapsular. It is obvious that such formations cannot serve to transmit impulses from one sympathetic ganglion cell to another unless we are prepared to admit exceptions to the law of the dynamic polarity of neurones. But even then the capsule would be interposed between the nerve cell and the surrounding dendritic nest. So characteristic an


THE JOURNAX OF COMPARATIVE NBUROLOGT, VOL. 29, NO. 4


354 S. W. RANSON AND P. R. BILLINGSLEY

arrangement cannot be entirely accidental; and the most obvious functional significance of the dendritic nest would be that the two neurones are thereby in position to be activated by the same pregangUonic fiber. This is now Cajal's interpretation of the pericellular dendritic baskets. Such baskets must not be confused with the basket-like appearances produced by dendrites winding their way between the cells without encircling them as has been done by Van Gehuchten and Sala. Functionally similar structures are the plate-like endings of dendrites outside the capsule of another cell as in Michailow's cells of Type III (fig. 8, b) and the smaller egg-shaped endings of the termilial branches of the dendrites of Michailow's Type IV cells which are also applied to the outer surface of the capsule of another cell (fig. 8, c). We believe that all of these formations are designed to place two neurones under the influence of the same preganglionic fiber as illustrated in figure 15, c.

3. Another arrangement of dendrites which seems designed to favor the simultaneous activation of two or more neurones by one preganglionic fiber is found in the bi-, tri-, and multicellular glomeruli in the human superior cervical ganglion. This is illustrated diagrammatically in figure 15, a. Such glomerulae, formed by the dendrites of two or more cells, are numerous in the human ganglion, and one is illustrated in figure 6. A single axon ramifying within such a glomerulus would be in position to activate each neurone contributing dendrites to the glomerulus.

SUMMARY AND CONCLUSIONS

Although attention is directed in this paper particularly to the cephalic end of the sympathetic trunk and the superior cervical ganglion, the comments drawn from the literature are for the most part applicable to the entire trunk.

A study of the literature based on the evidence obtained by the nicotine and degeneration methods shows that the cephalic end cf the sympathetic trunk consists of preganglionic fibers arising in the upper segments of the spinal cord and terminating in the superior cervical ganglion, and that the cells located in this ganglion give rise to fibers which run to terminate in the glands and smooth muscle of the head.


THE CERVICAL SYMPATHETIC TRUNK 355

In fact, the cephalic end of the sympathetic trunk consists ahnost exclusively of fine medullated fibers, most of which vary in size from 1.5/^ to 3.5/x. These fibers degenerate in an ascending direction after section of the nerve. In pyridine silver preparations no unmyelinated fibers can be distinguished in the normal sympathetic trunk at this level except for some fine branches of distribution from the superior cervical ganglion which happen to be included for a short distance in the same sheath with that nerve. Our observations along with those of Langley show that the superior cervical and stellate ganglia are not connected by myelinated commissural fibers and that unmyelinated commissural fibers if present are very few in number. Physiological experiments conducted by Langley failed to show any evidence of commissural fibers joining these two ganglia. Physiological and histological evidence is also against the presence of afferent fibers in the cervical portion of the trunk.

The nervus caroticus internus in the cat contains, in addition to great numbers of unmyelinated fibers, a very considerable number of fine myelinated fibers, mostly from 1.5m to 5.5^ in diameter. The fibers in this nerve do not degenerate after section of the sympathetic trunk in the neck ; all or nearly all of them are postganglionic fibers with their cells located in the superior cervical ganglion.

The dendrites of the cells in the superior cervical ganglion are of two kinds, intracapsular and extracapsular. The intracapsular dendrites are rarfe in the sympathetic ganglia of mammals but abundant in the human superior ganglion. Here they give rise to the complicated subcapsular formations that have been designated as dendritic crowns and glomeruli. A glomerulus may be formed from the dendrites of a single cell or from those of two or more cells and is designated accordingly as an unicellular, bicellular, tricellular, or multicellular glomerulus.

The extracapsular dendrites are long branched processes which run in every direction among the ganglion cells. In pyridine silver preparations it is not possible to follow them to their true terminations. We have summarized Michailow's account of the termination of these dendrites in preparations stained with methylene blue and illustrated them in figure 8. The dendrites


356 S. W. RANSON AND P. R. BILLINGSLEY

of one cell may form baskets or other special endings about neighboring cells, but these dendritic endings seem to be always outside the capsule of the second cell and therefore could not transmit impulses to it.

Sensory neurones with long dendrites have been described in sympathetic ganglia by Dogiel, but a review of the literature on this point shows that his interpretation of these structures has received little support from the observations of others. It is also doubtful if the axons of cells in the sympathetic ganglia run to spinal ganglia to form baskets about the cells located there.

The axons of sympathetic ganglion cells may acquire myelin sheaths, but usually do not. A study of the literature would indicate that they usually run, without giving off collaterals, into one of the branches of distribution arising from the ganglion. Some run through a connecting nerve to another ganglion, but there is no evidence to show that they ever end there. It would seem more likely that these fibers merely run through this second ganglion to join the nerve to which they are distributed. Some postganglionic fibers give off collaterals either in the original ganglion or in a second ganglion through which they pass, but these collaterals have been shown by Michailow to have endings not well adapted for the transference of nerve impulses.

Between the cells is a rich plexus of fine axonic ramifications which is formed by the branching of the preganglionic fibers. This disappears when the preganglionic fibers degenerate. It is probable that many of the fibers of the intercellular plexus form synapses with the dendrites of the sympathetic ganglion cells.

In pyridine silver preparations of the superior cervical ganglion of the cat it is possible to trace the darkly stained preganglionic fibers from the sympathetic trunk and to see that they undergo repeated branching and take a large part in the formation of the intercellular plexus. The postganglionic fibers, which are more lightly stained, and for the most part devoid of branches, take only a minor part in the formation of this plexus, but become grouped into bundles of parallel fibers which run toward the branches of distribution of the ganglion.

There is no evidence for the existence of synapses, either commissural or sensory-motor, between the neurones located in the


THE CERVICAL SYMPATHETIC TRUNK 357

ganglion and there appears to be no mechanism for a diffusion of incoming nerve impulses to all of the cells nor to all of the cells of a given function within the ganglion.

Evidence furnished by nicotine and degeneration experiments shows that all the synapses between the pre- and post-ganglionic neurones on the pathways through the superior cervical ganglion are located in that ganglion. There are no ascending postganglionic fibers in the cervical portion of the sympathetic trunk and no preganglionic fibers are continued through' the superior cervical ganglion into the branches of distribution. The pre-postganglionic synapses seem to be of two kinds : 1) pericellular networks and 2) relations established between the dendrites and axons in the intercellular plexus. One preganglionic fiber activates several post-ganglionic neurones. The dendrites of the postganglionic neurones serve .to increase the complexity of these relationships and may aid in bringing two or more neurones under the influence of a single axon as indicated in figure 15.

LITERATURE CITED

Arnstein, C. 1887 Die Methylenblaufarbung als histologische Methode.

Anat. Anz., 2, p. 125. Aronson 1886 Beitrage zur Kenntniss der centralen und peripheren NerA^en endigungen. Inaugural Dissertation, Berlin, 1886. Carpenter, F. W., and Conel, J. L. 1914 A study of ganglion cells in the sympathetic nervous system with special reference to intrinsic sensory

neurones. Jour. Comp. Neur., vol. 24, pp. 269-281. Cajal, S. Ramon. 1911 Histologic du systeme nerveux de I'homme et des

vertebres, vol. 2, p. 891. Paris, 1911. Chase, M. R., and Ranson, S. W. 1914 The structure of the vagus nerve.

Jour. Comp. Neur., vol. 24, p. 31. DoGiEL, A. S. 1895 Zur Frage liber den feineren Bau des S5mipathischen Ner vensystems bei den Saugethieren. Arch. f. mikr. Anat., vol. 46, p. 305.

1S96 Zwei Arten sympathischer Nervenzellen. Anat. Anz., 11, pp.

679-687. Ehrlich 1886 Uber die Methylenblaufarbung der lebenden Nervensubstanz.

Deutsche Med. Wochensch., Bd. XII, p. 49. HoFMANN, F. B. 1904 Die neurogene und myogene Theorie der Herzthatigkeit

und die Funktion der inneren Herznerven. Schmidt's Jahrb. d. ges.

Medicin, Bd. 281, p. 113. HuBER, G. C. 1899 A contribution on the minute anatomy of the sympathetic

ganglia of the different classes of vertebrates. Jour. Morph., vol. 16,

pp. 27-90.


358 S. W. RANSON AND P. R. BILLINGSLEY

HuBER, G. C. 1913 The morphology of the sympathetic system. XVII International Congress of Medicine, London, 1913, Section I, p. 211.

V. KoLLiKER, A. 1894 tJber die feinere Anatomie und die physiologische Bedeutung des sympathetischen Nervensystems. Gesellschaft Deutscher Naturforschen und Aerzte. Verhand. 1894, AUgemeiner Theil. 1896 Handbuch der Gewebelehre des Menschen. Bd. 2, pp. 854-871. Leipzig, 1896.

KuNTZ, A. 1913 On the innervation of the digestive tube. Jour. Comp. Neur., vol. 23, p. 173.

L.\NGLEY, J. N. 1892 On the origin from the spinal cord of the cervical and upper thoracic sympathetic fibers with some observations on white and grey rami communicantes. Phil. Transact. Roy. Soc. Lond., vol. 183, Series B, p. 85.

1896 Observations on the medullated fibers of the sytapathetic system and chiefly on those of the grey rami comlnunicantes. Jour, of Physiol., vol. 20, p. 55.

1900 a Reanarks on the results of degeneration of the upper thoracic white rami cotnmunicantes, chiefly in relation to commissural fibers in the sympathetic system. Jour, of Physiol., vol. 25, p. 468. 1900 b The sympathetic and other" related systems of nerves. Schafer's Text-book of Physiology, vol. 2.

1903 The autonomic nervous system. Brain, vol. 26, p. 1.

1904 On the question of commissural fibers between nerve cells having the same function. Jour, of Physiol., vol. 31, p. 244.

v. Lenhossek, M. 1894 Beitrage zur Histologie des Nervensystems und der

Sinnesorgane. Wiesbaden, 1894. M.'i.RiNESEO, M. G. 1906 Quelques rechereches sur la morphologie noi;male et

pathologique des cellules des ganglions spinaux et sympathiques de

I'homme. Le Nevraxe, t. 8, p. 9. MicHAiLow, S. 1908 Die Neurofibrillen der sympathischen Ganglienzellen bei

Saugetieren. Folia Neuro-biologica, Bd. 1, p. 637.

1911 Der Bau der zentralen sympathischen Ganglien. Internat.

Monatschrift f. Anat. u. Physiol., vol. 28, pp. 26-115. MtJLLER, R. L. 1909 Studien iiber die Anatomie und Histologie des sympathischen Grenzstranges insbesondere iiber seine Beiziehungen zu dem

spinalen Nervensysteme. 26. Kongr. innere Med. Wiesbaden, p. 658.

Ref. in Jahres. Anat. u. Entwick., 15 III, p. 731. Retzius, G. 1889 Zur Kenntniss der Ganglien Zellen des Sympathicus. Ver handlungen d. biolog. Vereins in Stockholm. Bd. 2, 1889. Cited

after Huber. Sala, L. 1893 Sur la fine, anatomie des ganglions du sympathique. Archiv.

Ital. de Biol., vol. 18, p. 439. Van Gehtjchten, A. 1892 Les cellules nerveuses du sympathique chez quelques

mammiferes et chez I'homme. La Cellule, t. 8. Waller and Budge Cited after Langley.