Skip to main content

Full text of "Journal of comparative neurology"

See other formats 


‘2 
a 
F 
rH 
* 
i! 
‘ate 
ee 
* 
aa 
‘ 
S 
. 


a] 
eee 


‘ 
¢ 
‘ait 
‘ 
- 
“> o ‘ 
4 2-6 2-4 
26 8 ~ 


CG 
> 
a2 
“s 
» i 


* ui 


7% . 
Pe oe ee & 
i) 


* 22 @¢ 2s 
enetet 
AG 
=é. 
* 


ee 
. 


> - 2 
x 
+ 


fc 


a 

J 

* 
* 


aK 
> 


yt ee 


. 
eee 


> 


* 
wn 


oe a ¢ 


vane 
.? 2 


Se 


e 
o*# 
e 


+ ee 


rs < 
ast Te 


* 
<4 
2 
-« 
ta 


4 6 a a we 
¢ 


“sr aes 


. 


—- 
Ce ee 
“2 ¢ 
* 
’ 
+ 
4 
Ca 


e-s-@ 6 


. 
-". « 


. 
. 

Sides 
« 


Pu 
% 2a 
te Pd 
+ + 
s*-e 
* 
* aa 
(a MARA 
sae 
Sete 


oo 2 * 
a 
* 


te 

4 
oes 
i 
ae? 

- 

ae 

ee 


ane 
4 


** 


A 
4-4 2 


ie ke 


of 4-¢ 
* 


ee 


* 


«4 i 
%% t~ 


¢ 
* 
+ 
-* 
*- 
s 


6 
ae 
e 
« 


¢ 


* 
or ne 
a 


2 4 
ofa 
*? 2-22 
2+ 2? 
ree 
ie 
* «of @ 
er @ 
* 
aie 86 ** 
* 
* 
as 
ee ee 
ee 2 4-82-44 ¢ 


24-2 © 


erGoetaerr ah? 2 os 
« 


* 
err eres eee 
“o 4 


i 
72 8 6-o 
22-6 
2&8 a 
Pa 


* 
eed 
t 


2-2. ¢-¢ 
o-* 
ae 2 a« 
ona 
« 
e 


- 


Ct 


. 


teat 
?-2-<¢ 
2e- 2 ¢ 
+26 
a4 ‘ 
eee #2 © & 
a 


a ea ee 
ee ew ee en 


* 
> 
Ses 24a 6 98 


een 


++ 


. 
© 
* 
OS RRS 
* 


>? 


im 
gener 
a 
ate 
te as 
4<@e 
or 2 
, 
4a-nneeesat 
~ - 
“« 
2 
2 
a 
2 2 
i 
*-ee 


Ne & 


hed 


= 


ers 
e 
© 
2 
* 
* 
o-* 


- 
e 


x 


<: 
-” 


~@ 2-0 @ 
ee oon 


“ea et 
24 
* 

* +2 +e ¢ ¢ 
A PRP PDAS 


*. 
ee *. 
ware ary 
hey 


' 
sy i 
“ b 
i 


A 


* 4 


a gtk 


THE JOURNAL 
OF 


COMPARATIVE NEUROLOGY 


EDITORIAL BOARD 


Henry H. DoNALDSON ApoLF MEYER 

The Wistar Institute Johns Hopkins University 
J. B. JoHNsTON Outver 8S. STRONG 

University of Minnesota Columbia University 


C. Jupson HERRICK, University of Chicago 
Managing Editor 


VOLUME 26 


1916 


PHILADELPHIA, PA. 


THE WISTAR INSTITUTE OF ANATOMY AND BIOLOGY 


COMPOSED AND PRINTED AT THE 
WAVERLY PRESS 
By THE Witiiams & WILKINS Company 
BatrmoreE, Mp., U.S. A. 


CONTENTS 


1916 


No.1 FEBRUARY 


W. J. Crozimr. Regarding the existence of the ‘common chemical sense’ in vertebrates 1 
Witu1am F. Auten. Studies on the spinal cord and medula of cyclostomes with special 
reference to the formation and expansion of the roof plate and the flattening of the 


spinal cord. Eighty-seven figures... St iidw tidee wie seek See side he ghk Ete Od 9 
PerctvaL Baitry. Morphology of the are plate of the fore-brain and the lateral choroid 
plexuses in the human embryo. Thirty-one figures... . ; ay eee 79 
No. 2 APRIL 
Leste B. Arry. The movements in the visual cells and retinal pigment of the lower 
vertebrates. Thirty-seven figures.... 4 Se BE 
H. Saxton Burr. Regeneration in the brain of Ambly stoma. I. The fore-brain. Four 
RR bs Warn are Sot cao alan a'ornhd do» Rx, «os >:s +> - geile di o 16s werha tls dec 208 


No.3 JUNE 


Lesuiz B. Arey. The function of the efferent fibers of the optic nerve of fishes. Twelve 
SAREE TU OMENS RRL ALG oo 5 oa 5 he a Savdsn « « « » «GUMS atau Clad sowie Mpls es SiS 213° 

G. E. Coeguinty. Correlated anatomical and physiological studies of the growth of the 
nervous system of Amphibia. II. The afferent system of the head of Amblystoma. 
Seventy-nine figures. . 8 On Ee oe eer ere ee. ee 247 

R. A. Kocurr. The Bieet ca activieea on phe histalamteal serhis ture of nerve cells......4 241 


No. 4 AUGUST 


Lesuiz B. Arey. The influence of light and temperature upon the migration of the ret- 
inal pigment of Planorbis trivolvis. Nine figures.................sccccsec eee r cree 359 

A. T. Rasmussen and J. A. Myers. Absence of chromatolytic change in the central nerv- 
ous system of the woodchuck (Marmota monax) during hibernation. Six figures 


BS PMAE ES hg tecerd eae eta 9 Siew Sign =/<) SRS other ti A Sw es pee ved > 2 ide RY 391 
J. J. Kenean. A study of a Plains Indian brain. Eight figures.................-....- 403 
Martin R. Coase. An experimental study of the vagus nerve. Four figures. ......... 421 
Lestiz B. Anny. Changes in the rod-visual cells of the frog due to the action of light. 

Two figures. AE ee aa Se Ee ee ak 429 


iv CONTENTS 


H. H. Donatpson. A preliminary determination of the part played by myelin in reduc- 
ing the water content of the mammalian nervous system (albino fat). One chart.. 448 
W. J. Grozipr. The taste of acids. Two figuresitiis. hes t cn. 1 > > «ener ne peo 453 


No.5 OCTOBER 


Wiutiam A. Hitron. The nervous system of pyenogonids. Twenty-one figures....... 463 
J. B. Jonnston. Evidence of a motor pallium in the fore-brain of reptiles. One figure. 475 
J. B. Jounston. The development of the dorsal ventricular ridge in turtles. Twenty- 
OVO TOTITOR 2 ooh. si 8 oe ccc acackle-« 5 pine cre eerie Sete cae Ne Sento te ta a 481 
PercetvAL Barney. The morphology and morphogenesis of the choroid plexuses with 
especial reference to the development of the lateral telencephalic plexus in Chry- 


semys marginata. ‘Twenty-seven figures... 000.5 oa cee i ae 507 
Sumner L. Kocu. The structure of the third, fourth, fifth, sixth, ninth, eleventh and 
twelfth cranial nerves. Five figures: 22.053. 625 hls - oe ee ee 541 


CAROLINE BurtING THompson. The brain and the frontal gland of the castes of the 
‘white ant,’ Leucotermes flavipes, Kollar. Twenty-six figures...................... 553 


CONTRIBUTIONS FROM THE ZOOLOGICAL LABORATORY OF THE MUSEUM OF COMPARATIVE ZOOLOGY AT 
HARVARD COLLEGE. No. 268. 


REGARDING THE EXISTENCE OF THE ‘COMMON 
. CHEMICAL SENSE’ IN VERTEBRATES! 


W. J. CROZIER 


I. In a recent paper by Coghill (14) exception is taken to 
the view held by Herrick (’08), Sheldon (’09), Cole (’10), and 
Parker (’12), that there is in vertebrates a set of receptors (free 
nerve terminations) which are responsible for reactions to rather 
high concentrations of chemicals when applied to moist periph- 
eral surfaces. The theoretical significance attached to this 
‘common chemical sense’ (Parker, 712) makes it appropriate to 
bring forward certain facts which may to some extent serve to 
clear up the situation. 

The evidence adduced by Coghill is this: that in Amblystoma 
larvae, before the establishment of the definitive nervous system, 
tactile and chemical irritability appear concomitantly and, so 
far as studied, remain completely parallel in development; and 
that if larvae be placed in irritating solutions of hydrochloric 
acid (as dilute as jo 900), the epithelial cells become visibly 
disrupted, the reaction of the larvae being correlated with the 
disintegration of the skin and a gradual disappearance of reaction 
to tactile stimuli. 

His conclusion, that there is not present in the skin any normal 
irritability to acid, is probably correct for the larvae experi- 
mented upon; but it by no means follows that the same con- 
ditions obtain in the adult. Coghill argues, however, that the 
permanently embryonic cells of the germinative layer in the 
skin of amphibians and fishes when exposed to the high con- 
centrations of acid (,)) used by Parker and Sheldon act as do 
the superficial ciliated cells of the Amblystoma larval skin. 
Because they are bound down by a ‘‘thick, less sensitive, and 


1 Contributions from the Bermuda Biological Station for Research. No. 40. 
1 


THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL. 26, NO. 1 
FEBRUARY, 1916 


2 W..J. CROZIER 


more impervious layer of cells,’’ the disintegration of the sensi- 
tive cells would cause exceedingly violent disturbances, pre- 
sumably affecting tactile and pain terminals; and further, that 
it has not been satisfactorily demonstrated that it is possible 
(as claimed by Sheldon and by Cole) to effect a separation 
of tactile and ‘common chemical’ irritability in fishes and 
amphibians. . 

This general argument is somewhat strengthened by the fact 
(to which Coghill does not refer) that aquatic vertebrates pos- 
sessing a soft slimy skin—cyclostomes, eels, catfish, Necturus— 
are known to react to local irritation by chloroform and other 
substances by the expulsion of mucus and even of entire gland 
cells. The reaction time of this response is not known accu- 
rately, though it is judged to be much greater than that, for 
example, of the catfish as a whole when stimulated by hydro- 
chlorie acid applied to its trunk region; still, it is not incon- 
ceivable that mechanical. deformations brought about in this 
way might provide stimuli for tactile nerve endings, provided 
the mucus response were sufficiently sharp and prompt. 

II. The exact manner in which solutions are to produce the 
hypothecated effects upon the cells of the germinative layer 
is left entirely untouched by Coghill. The rapidity of the re- 
sponses given by the catfish and Necturus immediately negatives 
the idea that osmotic transfer of water is the agency of stimu- 
lation. In the case of the spinal frog, as studied by Braeuning 
(04), Loeb (’05), and Cole (10), though the reaction times are 
rather long, there is abundant evidence that other than osmotic 
factors are at work. The most important point which arises 
for consideration is the extent to which chemical stimulants 
actually penetrate the skin, 1.e., the degree to which the cells of 
the germinative layer are exposed to the action of the stimu- 
lant. In addition to the rapidity of the reactions under discus- 
sion, it is to be remembered that they are excited by acids, salts, 
alcohols, and a variety of other-substances. 

That the skin of aquatic animals is not to any appreciable 
degree damaged by the agents responsible for ‘common-chemical- 
sense’ reactions, is clearly indicated by such facts as the following: 


COMMON CHEMICAL SENSE 3 

1. High concentrations of acids (#4), salts (2N), alkalies ( 3) 
and alkaloids (;5)), which evoke prompt and vigorous reactions 
from the earthworm, Eisenia foetida, cease to cause any dis- 
turbance the moment the stimulant is washed off by immersing 
the animal in water. If any serious disintegration were produced 
by these solutions, it would be reasonable to expect the continu- 
ance of activity after the external supply of the stimulant had been 
removed.’ 

2. Holothurians, such as H. surinamensis (Crozier, ‘15 ¢) 
and Stichopus moebii (Crozier, 15 b), possess skin pigments 
whose loss is a fairly good indicator of changes in permeability. 
Yet in no case on the application, by the pipette method, of the 
stimuli used by Parker (’12) was there any indication of pigment 
loss accompanying reaction.* This is true also of the nudi- 
branch, Chromodoris zebra. 

III. The value of the intracellular indicator of Chromodoris 
in the study of cell penetration by acids has been pointed out 
elsewhere (Crozier, °15b). When small volumes of acids, 
even in +; solution, are used to stimulate Chromodoris, there is 
no visible evidence that they penetrate the skin. In fact the 
time required for the penetration of the acids is quite high 
(see also Harvey, ’14), the most rapid rate observed being with 
3» 1so-valeric acid, where penetration requires 1.5 minutes. 


? Experiments with earthworms also disclose certain important deficiencies 
in’ the method whereby the animals to be tested are entirely immersed in the 
stimulating solution. If the reactions are to be studied quantitatively, there 
are features undoubtedly obscured by this procedure. For example, the time 
elapsing between the complete immersion of an earthworm (Hisenia foetida 
in a >) lithium chloride solution and the instant the first writhing movement 
appeared, was found to be at least ten times greater than the reaction time of the 
same worm when part of it (the anterior end) was stimulated. The speed of 
reaction is in part conditioned by the number of receptors affected, but this 
does not analyze the situation completely. It is quite probable that the re- 
sponses observed by Coghill are not at all due directly to sensory stimulation by 
acid. 

3T have observed that under certain circumstances Ptychodera sp.—a Ber- 
mudan enteropneust—reacts to chemical irritation by extruding a yellowish 
pigment. There is ground, however, for believing that this is an instance com- 
parable to that of Lumbriconereis (Kschischkowski, ’12), in which Ix salts have 
apparently a more or less specific action in producing the response. 


| W. J. CROZIER 

For sy HNO; the penetration time found (Chromodoris tissue) 
was 3.5 minutes; for 4, 4.3 minutes; whereas Necturus, under 
water, was found to react to 7 HNO; (0.5 ee. of the solution 
being applied with a pipette) in 1.5 seconds when stimulated on 
the dorsal surface of the head, 5 seconds on the lateral mid-body 
surface, and 2 seconds on the tail; with HNO; the reaction 
times for corresponding locations were 5, 10, and 6 seconds 
respectively.!. Even here, though the pipette tip was held within 
approximately a centimeter of the animal’s surface, the concen- 
tration of the solution in the pipette does not represent the con- 
centration which actually reaches the stimulated surface, as 
already noted by Parker (12, p. 222). The conclusion must 
therefore be, that acids do not penetrate the skin with sufficient 
rapidity to affect the cells of the germinative layer. This con- 
clusion must also be extended to alkalies, since Harvey (10) 
and others have shown the high impermeability of cells to 
strong hydroxides. 

IV. In the light of the evidence just discussed, it seems im- 
probable that the high concentrations of irritants employed by 
Parker and others in studying reactions attributed to the com- 
mon chemical sense produce any violently disruptive effects 
when applied to the skin of aquatic animals. Indeed, so far as 
concerns the skin as a whole, they do not penetrate at all,° 
and the cells of the germinative layer cannot be held to be ex- 
posed to their action. 

This conclusion was verified, in the case of the spinal frog, by 
experiments of the following type: 

1. The reaction time for the withdrawal of the frog’s foot 
from 7 CuSQ,, is about 7 seconds (at 24°9). After being 
withdrawn from the solution, the feet continue for some time 
to be spasmodically contracted. These subsequent contractions 
are entirely inhibited the instant a foot, just retracted from CuSO,, 


* 


4 The reaction times were measured at 20°, while the penetration of the acid 
was studied at 27°; the speed of penetration decreases markedly with falling 
temperature. 

> The mode of action of the stimulating agent on the individual receptor is 
entirely another question. 

‘ 


COMMON CHEMICAL SENSE 5 


is dipped into a weak solution of K,Fe(CN)., which precipitates 
the copper held by the mucus of the foot. A similar result was 
obtained with copper acetate. Washing the foot with distilled 
water does not lead to a cessation of the contractions, because, 
after exposure of the foot to certain solutions (Loeb, 05), water 
stimulates. 

2. After two to four stimulations of the frog’s foot by relatively 
strong solutions of either copper acetate or ferric sulphate, the 
stimulated area was sectioned .and tested microchemically for 
the penetration of the copper or iron. The metals were found in 
the mucus of the surface of the foot, and in several instances 
doubtful traces were discovered between cells of the extreme 
outer portion of the epidermis. No evidence was found of any 
disruption of the germinative layer. 

3. As a stimulating agent whose penetration would readily 
be visible, the action of picric acid was studied. The outcome 
of experiments with this substance at several concentrations may 
be illustrated by the records here copied: 

Experiments Hand L. Picrie acid, %. R. T.=Reaction time in seconds. 


Successive tests at 5 min. intervals. N. R.=No reaction. 


R.T 
NO, os NOTES 

H I 
iL sate oa Bee 7.8 ay No staining. 
BSAA DN «6 seer d the 12.8 6.6 ’| Slight staining. 
153 SRe, ee: 16.8 7.8 

20.9 

Oe SRR 30.2 11.8 Staining progressively 
re 40.0" 24.6  Drtenee 
ee 
ee aes 45.0 | 43 .2* ties ce 
ER Seed devas bees N.R. | N.R. No reaction to ¥ eshte acid. 


* Not reactive to pinching beyond this point. 


From these and similar tests, it was concluded that the pene- 
tration of the stimulant rapidily renders the frog’s foot less re- 


6 W. J. CROZIER 


active by killing the superficial cells. In the case of picric acid 
it is probably the H ion which is mainly concerned in stimulation, 
since 3; ammonium picrate is entirely ineffective, neither does 
it easily stain the frog’s foot. This does not, however, signify 
that other acids behave as does picric, since as many as 20 suc- 
cessive reactions may be obtained from * HCl. Moreover, the 
staining is not directly correlated with the stimulating effect, 
since the skin of the frog’s foot was stained by immersion for 
3 minutes in 39) picric acid without producing any reaction. 
The skin of Necturus and the catfish may be stimulated with 
3) picric acid without producing any visible stain. 

V. I have repeated on the frog’s foot the experiment of Shel- 
don (’09) and Cole (10) regarding the separation of tactile 
irritability and responsiveness to irritating solution, by treat- 
ment with cocaine. A 0.5 per cent solution of cocaine hydro- 
chloride was used, and the tests were made by comparing at 
intervals the reactions of the cocainized foot with those of the 
untreated one. After about 20 minutes’ immersion, the reaction 
time of the cocainized leg to + formic acid (chosen as a powerful 
stimulant) was usually twice that of the normal foot; after about 
an hour, varying in some tests to an hour and a half, the cocainized 
foot no longer reacted to pinching, but gave good responses to 
the acid with reaction times of 10-15 seconds, still about twice 
the reaction time of the non-narcotized leg. 

It is therefore possible, I believe, to effect experimentally a 
separation of sensitivity to mechanical and to chemical stimu- 
lation in the frog’s foot. 

There can be no question of the distinctness of the human 
sensation attributed to the common chemical sense (Parker, 712; 
Parker and Stabler, ’13) as compared with any tactile sensation; 
and from tests made upon cocainized areas of my own mouth,® I 
am certain that the two sets of receptors are not only qualita- 
tively distinct as regards the sensations with which they are 
connected, but also may be separated by the use of cocaine. 


* These tests concerned mostly the inner surface of the cheek. 


COMMON CHEMICAL SENSE 7 


SUMMARY 


1. There is positive evidence, in the case of alkalies, acids, 
and certain salts, that solutions supplying stimuli for the com- 
mon chemical sense do not penetrate the skin of aquatic animals, 
nor when applied from a pipette do they damage the skin to any 
extent. 

2. There is consequently no ground for Coghill’s assumption 
that the cells of the germinative layer of the epithelium of fishes 
and amphibians are exposed to the action of the stimulating 
agent and thereby disrupted; and there is no histological evi- 
dence of disruption. 

3. The reactions attributed to the common chemical sense 
depend upon a group of sense organs distinct from those sensi- 
tive to mechanical stimulation. 


REFERENCES 


Bragunina, H. 1904 Zur Kenntnis der Wirkung chemischen Reize. Arch. 
ges. Physiol., Bd. 102, p. 163-184. 


Coauitt, G. E. 1914 Correlated anatomical and physiological studies of the 


growth of the nervous system of Amphibia. I. The afferent system 
of the trunk of Amblystoma. Jour. Comp. Neur., vol. 24, p. 161-233. 


Coz, L. W. 1910 Reactions of frogs to chlorides of ammonium, potassium, 
sodium, and lithium. Jour. Comp. Neur., vol. 19, p. 273-311. 


Crozier, W. J. 1915 a The behavior of an enteropneust. Science, N. S., vol. 
41, p. 471-472. 
(Note—Lines 5 and 6 [p. 472] have been accidentally interchanged by 
the printer. ) 
1915 b The rhythmic pulsation of the cloaca of holothurians. Ibid., 
p. 474. 


1915¢c The sensory reactions of Holothuria surinamensis. Zool. 
Jahrb., Abt. Physiol. [Not yet published]. 


1915 d On cell penetration by acids. Science, N.8., vol. 42, p. 735- 
736. 


Harvey, E. N. 1910 Studies on the permeability of cells. Jour. Exp. Zodl., 
vol. 10, p. 507-556. 


1914 Cell permeability for acids. Science, N. 8., vol. 39, p. 947-949. 


Herrick, C. J. 1908 On the phylogenetic differentiation of the organs of taste 
and smell. Jour. Comp. Neur., vol. 28, p. 157-166. 


8 W. J. CROZIER 


Kscuiscuowskl, KX. 1912 Neue Beitrige zur Pigmentabsonderung bei Anne- 
liden. Zentralbl. f. Physiol., Bd. 26, p. 528-582. 


Lorn, J. 1905 Production of hypersensitivity of the skin by electrolytes. In: 
Studies in General Physiology, pt. II, Chicago. pp. 748-765. 


Parker, G. H. 1912 The relation of smell, taste, and the common chemical 
sense in vertebrates. Jour. Acad. Nat. Sci. Phila., Ser. 2, vol. 15, p. 
221-234. 


Parker, G.H., ano Stasier, EH. M. 1918 On certain distinctions between taste 
and smell. Amer. Jour. Physiol., vol. 32, p. 230-240. 


Suetpon, R. E. 1909 The reactions of the dogfish to chemical stimuli. Jour. 
Comp. Neur., vol. 19, p. 273-311. 


STUDIES ON THE SPINAL CORD AND MEDULLA OF 
CYCLOSTOMES WITH SPECIAL REFERENCE TO 
THE FORMATION AND EXPANSION OF THE ROOF 
PLATE AND THE FLATTENING OF THE SPINAL 
CORD! 

WILLIAM F. ALLEN 


Department of Anatomy, University of Minnesota 


EIGHTY-SEVEN FIGURES 


CONTENTS 

SP AUNOUUCHLON: aed eerie s cies ta ways ena oso e a «<P § 0d oe ee a Le ag 9 
Material and its preparation. . ‘ 10 

Il. Hereditary and mechanical causes ‘underigi ing “the formation of ‘the 
fourth ventricle and the tela chorioidea....................... des 
1. Roof plate of the medulla oblongata.......................00008: 12 
A. Amphioxus.. , ys da ace WOR aatas wahee eal ete 
B. Pistoia: (Sdeliostom: B) 5 iondit eee ee HR at Ree 13 
C. Petromyzon. = Lt. gS ok» aeitadle been . 15 
D. Selachians and Auvphibia. na hss - ne ees sade Gene ee. 23 
RESEND NO Gy ci tuentieay ot <Ada. vce: ober ttae vee wee Meee ToS 25 
HSMN SMM OP VOB aes kie has a= « «SERED. 2 lores kha tema ein. 's te 27 

2. Descriptions of three roof plate expansions of the spinal cord in the 
ZU GH euTOLABTGHeRIA SELICH. .... <j. ee peice bub ceive evreee Be. « 28 
IiI. Causes underlying the flattening of the spinal cord in oyclostomes: sare See 
iV. General oonsiderapions Snd summary. ..:. ees. ss oe pe den elee ve 35 
Ns aL RPC CUE Ste ie eg. <5 33 ML § ple A aicia » Mean eb oD a.nd 40 


INTRODUCTION 


This paper has grown out of the study of the caudal heart and 
the spinal cord and nerves related to it in Polistotrema. Parts 
of that study which are not yet completed will be published 
later and will deal with the origin, distribution, and phylogeny 
of the spinal nerves; the origin of muscle sense organs in con- 
nection with the specialized muscles of the caudal heart; and 


1A thesis submitted in partial fulfilment of the requirements for the degree of 
Doctor of Philosophy, Department of Anatomy, The University of Minnesota. 


9 


10 WILLIAM F. ALLEN 


certain ganglion cells, possibly sympathetic, which are found 
along the course of the spinal and vagus nerves. The present 
paper includes a study of the mode of origin of the fourth ven- 
tricle in several groups, an account of a structure similar to the 
fourth ventricle found in the caudal part of the spinal cord of an 
adult Polistotrema, and a study of the development of the 
central canal, and the causes underlying the flattening and the 
ventral indenting of the spinal cord in Cyclostomes. 


Material and its preparation 


Since the material for this paper came from such diverse 
sources, and such different modes of technique were employed, 
too much time and space would be consumed if a detailed de- 
scription were given of each method employed. With the 
exception of the Petromyzon material, which was fixed in Flem- 
ming, corrosive-acetic, and Perenyi’s fluid, my own material 
was fixed either in Tellyesnicky’s or Bouin’s fluid. It was 
sectioned after paraffin or the combined celloidin-paraffin 
method of imbedding. The latter method gave by far the bet- 
ter results, since it appears to have all of the advantages of 
celloidin in causing almost no shrinkage and its ability to hold 
yolk granules and blood corpuscles intact; besides allowing 
the sections to be cut as thin as paraffin alone, and causing no 
more difficulty in manipulation. For the most part the sections 
were stained in Heidenhain’s iron hematoxylin and counter- 
stained in an alcoholic solution of orange G plus a little acid 
fuchsin. Ina few instances, in very young Petromyzon embryos, 
where all the tissues were filled with yolk, carmine and hemalum 
were used to advantage. 

Acknowledgments are due to Prof. T. G. Lee for the gift 
of a very complete series of Petromyzon embryos, which were 
obtained from the Connecticut River. Also to Prof. R. E. 
Scammon for the loan of his very complete serial collection of 
Squalus embryos, and to Prof. J. B. Johnston for the use of a 
similar serial collection of Amblystoma embryos. The splendid 
series of human and pig embryos belonging to the Institute of 


SPINAL CORD AND MEDULLA OF CYCLOSTOMES 11 
Anatomy of which Prof. C. M. Jackson is director, were espe- 
cially valuable. 

It is a great pleasure to the writer to have this opportunity 
of expressing his obligations to Prof. J. B. Johnston for his 
valuable suggestions and friendly criticism of this work. 

A model of the caudal end of the spinal cord from a region a 
little in front of the caudal hearts to its extreme tip was pre- 
pared in four sections from the 20 em. Polistotrema series. 
Also a model in two pieces was prepared of the cavity of the so- 
called first roof plate expansion and the enlarged central canal 
of the above mentioned series of Polistotrema. These models 
were constructed out of blotting paper after a modification of 
the Born method. Tracings of each section were made on 
ordinary writing paper with the aid of an Edinger-Leitz draw- 
ing apparatus, using a magnification of 100 diameters, and 
afterward each tracing was carefully checked up for accuracy 
with a higher magnification. At the very outset two base or 
projection lines were drawn on the first tracing, one following 
the median longitudinal plane, and the other a horizontal line 
drawn at right angles to the first line, passing along the ventral 
border of the notochord. These projection lines were added by 
pencil to all of the succeeding tracings after the following manner: 
The second tracing was carefully fitted over the first, after plac- 
ing both over a plate glass covered box, containing an electric 
light reflected upward, and in like manner these lines were added 
to the third tracing from the second, and so on to the end of the 
series. Then with the aid of carbon paper these tracings and 
projection lines were transferred to sheets of blotting paper, 
having a definite thickness, previously determined after the 
following method. When thoroughly cooled after an immersion 
in melted paraffin, the blotting paper selected should have a 
thickness, equal to the thickness of the section multiplied by 
the magnification used, which in this case totalled 1.5 mm. The 
sheets of blotting paper containing the transposed tracings were 
then immersed in melted paraffin, drained and cooled. After 
which the outlines of the tracings were cut out with a sharp 
scalpel, and the sections were built up in regular order. In 


12 WILLIAM F. ALLEN 

. 
part to maintain this regular order and in part to add strength 
to the model, a copper wire was inserted through each section, 
passing through the point of intersection of the two projection 
lines. After a certain number of sections had been strung on 
this wire they were securely fastened to each other by pins. 


HEREDITARY AND MECHANICAL CAUSES UNDERLYING THE 
FORMATION OF THE FOURTH VENTRICLE AND THE 
TELA CHORIOIDEA 


To the writer this problem appears to be primarily phylo- 
genetic rather than ontogenetic; consequently this study begins 
with the lowest vertebrates, and is approached as a problem of 
evolution. 


1. Roof plate of the medulla oblongata 


A. Amphioxus. Unfortunately I have not had access to 
any embryological Amphioxus material so that my inferences 
will have to be drawn entirely from adult material.2 At the 
outset it can be maintained with a considerable degree of safety 
that the adult Amphioxus brain contains nothing which can be 
homologized with the fourth ventricle of higher vertebrates. 
Figure 67, which pasess through the highest point of the anterior 
ventricle (V.,) demonstrates clearly that there is nothing here 
comparable with the fourth ventricle. For the dorsal portion 
of the cavity is much narrower than the ventral portion, and 
there is absolutely nothing in the way of a thin and expanded 
roof plate. In fact, the dorsal portion of the cavity is fairly 
filled up with processes from ependymal cells. 

A few sections behind the first ventricle there appears a small 
isolated cavity in the roof plate region of this series, designated 
as the posterior ventricle (fig. 68, V..), and a considerable dis- 
tance behind this cavity, there is a second dorsal cavity, also 
designated as the posterior ventricle (fig. 69, V.2). This is some- 
what larger than the previous cavity, and can be readily located 

2 In order to eliminate minor details of description from the text, very com- 


plete and detailed descriptions of the figures have been given at as end of the 
paper, to which the reader’s attention is directed. 


SPINAL CORD AND MEDULLA OF CYCLOSTOMES 13 


from its position immediately above a dorsal group of large nerve 
cells (M’.C’.). From the figures by Hatschek, Willey, Sterzi, 
and others it is evident that these isolated roof cavities described 
above, were once a part of a common central cavity of the 
embryonic brain, which later in development, became isolated 
‘through an invasion of ependyma, and it is entirely possible 
that this region of the central cavity in the embryo was much 
more suggestive of the fourth ventricle. Judging from the adult 
alone they may be looked upon merely as vestiges of the embryonic 
central canal. 

B. Polistotrema (Bdellostoma) My embryonic material of 
Polistotrema confirms the statements of Price, von Kupffer, 
and Dean that Polistotrema possesses well-developed ventricles 
in the embryo; the expansion being fully as great as in a similar 
stage of Petromyzon. As development proceeds the lateral 
plates increase in thickness from additions of fibers and cells 
until the fourth ventricle becomes reduced to a canal, but little 
larger than the central canal of the spinal cord. 

Sanders, Holm, Miss Worthington, Sterzi, Cole, and Nicholls 
describe and figure the fourth ventricle about as it is shown in 
figure 63. There is some little discrepancy in the terminology 
used, due largely to the differences of opinion as to whether or 
not Polistotrema has a cerebellum. If the posterior lobes of the 
mesencephalon should in the light of future investigation turn 
out to be a cerebellum, then the boundaries of the metencephalon 
will have to be carried further forward than we have indicated, 
and the so-called sinus mesoccelicus of Nicholls (S..) will 
have to be called the anterior dilation of the fourth ventricle of 
Miss Worthington’s description. 

My transverse series through the brain of Polistotrema show 
the condition of the ventricles to be almost identically the same 
as Sterzi and Nicholls found them. Until the embryological 
and functional areas of the brain have been better worked out 
it seems advisable to the writer to let the posterior border of 
Nicholls’ sinus mesoccelicus (fig. 63, S.M.) mark the boundry 
line between the mesencephalon and the metencephalon, and to 
regard his isthmie and ventricular canals (figs. 63 and 65, A.V.) 


14 WILLIAM F. ALLEN 


as dorsal and ventral portions of the anterior end of the fourth 
ventricle. For the reason that in Polistotrema they extend 
some distance behind the posterior tip of the mesencephalon, 
and the dorsal or posterior canal is shown in transverse section 
(fig. 65) to lie close to the dorsal surface, which is then the only 
part of the fourth ventricle to retain the characteristic dorsal 
position of the higher vertebrates. These two canals appear in 
this series about as Nicholls has described them, the dorsal 
(isthmic) is the larger and contains Reissner’s fiber. Although 
probably subject to a considerable variation, these canals 
apparently extend further caudad in this series than Nicholls 
represents them. Also in this series the dorsal canal (isthmic) 
gives off one or two branches at the level of the posterior tip of 
the mesencephalon (cerebellum of Miss Worthington), which 
run parallel to the main canal, but a little to one side and below 
(fig. 65, A..V.). After travelling side by side for some little 
distance in a mass of spongy ependymal tissue close to the roof 
plate, they reunite in the dorsal canal, and soon afterward 
both the dorsal and ventral (isthmic and ventricular) canals 
unite in a common canal, which is little if any larger than the 
central canal of the spinal cord. This constricted portion of the 
fourth ventricle (fig. 63) continues caudad, rather deep-seated 
in a mass of loose vascular ependyma, until the posterior end of 
the medulla is reached, where it expands into a much larger 
vesicle or sinus, designated as the posterior dilation of the 
fourth ventricle (figs. 63 and 66, P.4V). Behind this it soon 
tapers down into the ordinary central canal of the spinal cord. 

A glance at figures 63 to 66 suffices to show that the fourth 
ventricle of Polistotrema is greatly reduced as compared with 
that of Petromyzon. This is due probably to the rapid increase 
of fibers and cells in the lateral plates. Notwithstanding this 
reduction in size and general alteration in appearance and struc- 
ture, the walls of the fourth ventricle in Polistotrema, although 
representing a greatly modified chorioid plexus, are unquestion- 
ably capable of producing cerebro-spinal fluid. The posterior 
dilation of the fourth ventricle (fig. 66, P.4V.) contains cerebro- 
spinal fluid (S.C.F.) in the form of a deeply staining feltwork, 


SPINAL CORD AND MEDULLA OF CYCLOSTOMES 15 


which is not ependymal cilia or a tangled Reissner’s fiber. Also 
throughout its entire length, as was noted by Sterzi, the fourth 
ventricle is enveloped by a rather thick layer of spongy and very 
vascular ependyma, which would be distinctly favorable for 
infiltration and possibly for secretion into the ventricle. This 
rich blood supply is from the blood vessels and sinuses travers- 
ing the meningeal membranes, and especially from two large 
arteries (A. rhombencephalica of Sterzi), one of which appears 
in figure 64 (M.A.). Also it would be quite possible in the 
anterior part of the fourth (A.,V.) for the various canals, which 
run close to the dorsal, to receive infiltration direct from the 
outer meningeal blood and lymph sinuses. 

Coagulated cerebro-spinal fluid is also to be seen in reduced 
amounts in the mesencephalic ventricles designated as the 
posterior mesoccele and the sinus mesoceelicus. They are also 
surrounded by a layer of vascular ependyma, which, while much 
thinner than the corresponding layer of the fourth ventricle, 
doubtless functions as a cerebro-spinal fluid forming organ. 

A careful examination of this peculiar modification of the 
chorioid plexus of the fourth ventricle in Polistotrema leads 
one to believe that this is not as efficient an organ for the pro- 
duction of cerebro-spinal fluid as the more expanded tela chori- 
oidea of Petromyzon and higher vertebrates. 

C. Petromyzon. Petromyzon is apparently the best type that 
could be selected for obtaining definite information concerning 
the early history of the formation of the fourth ventricle and its 
expanded roof plate. 1) It is the lowest living vertebrate that 
possesses a well-developed fourth ventricle and expanded tela 
chorioidea in the adult. 2) At no times does the medulla have 
a pontine flexure. 3) Its central nervous system remains a 
solid cord of ectoderm until after the cranial and spinal ganglia 
are well-differentiated. For these and other reasons the early 
history of the fourth ventricle in Petromyzon has been studied 
in the effort to determine the essential factors involved in its 
appearance and growth. 

In order to arrive at the fundamental factors involved in the 
anlage and development of the fourth ventricle in Petromyzon 


16 WILLIAM F. ALLEN 


it is necessary to go back in the ontogeny of the central nervous 
system to the time when it was a solid cord. Of my embryos 
which represent stages killed at 5, 7, 9, 10, 11, 12, 14, 16, 18, 20, 
and 26 days after fertilization, all of the 5 day series revealed 
the central nervous system as a solid cord or keel of ectoderm 
cells, formed according to Calberla by a process of delamination. 
In places neural crest cells can apparently be seen budding off 
from the dorso-lateral surface of the brain and spinal cord al- 
most identically as described and figured by von Kupffer (’90). 
It is obvious that the preponderance of yolk granules in all 
tissues makes accurate observation in these stages very difficult. 
The majority of the 7 day embryos disclosed a central canal 
either formed or in the process of formation. Out of a great 
number of series of 10 day embryos, three were found in which 
the opening of the central canal had been considerably retarded. 
In one series the central nervous system was still a solid cord of 
cells, and in the other two the central canal was in the early 
stages of formation. These series were selected in preference 
to similar stages of 7 day embryos for the reason that less yolk 
was present to obscure the various structures. 

To facilitate comparison of similar sections of successive 
stages, figures 33 to 38 and figures 40 to 53 have been so arranged 
as to bring several regions of series in a horizontal row, while 
successive stages of the same region are placed in a vertical row. 
The conditions at the level of the trigeminal, auditory and vagal 
ganglia are thus readily compared. 

In figure 32 we have a section through the medulla in the region 
of the VIII ganglion, which portrays very well an extremely 
early stage in the formation of the central canal. Here the 
medulla will be seen to be composed of a mass of nuclei imbedded 
in a syncytium of protoplasm. Excepting in the roof and floor. 
plates, the nuclei have migrated some little distance to either 
side of the median sagittal plane, presenting the appearance of 
a rather broad light line surrounded by nuclei. The center of 
this light streak of protoplasm discloses a very conspicuous 
furrow or seam (C.C.S.), which appears in every section of the 
central nervous system of this series, very much as it looks in 


SPINAL CORD AND MEDULLA OF CYCLOSTOMES 17 


this section. This furrow or seam represents the position of the 
future central canal. The protoplasm adjacent to this seam is 
sufficiently granular to suggest a secretory function at this early 
stage. 

Figures 33 to 35 also taken from a 10 day Petromyzon series 
exhibit a slightly later stage in the formation of the central 
canal. Conspicuous cavities (C.C.) have appeared in the dorsal 
and ventral portions of this furrow, which are not only visible 
throughout the entire medulla, but are continuous throughout 
the spinal cord. It should be recorded that a few sections through 
the medulla possessed slightly larger dorsal and ventral cavities 
than were represented by these figures, but in no case had they 
approached each other close enough to unite. In other respects, 
excepting possibly for a few more fibers in the marginal layer, 
the structure of the medulla has remained about the same as 
in the previous series. Emphasis should be made of the fact 
that each of these cavities, when examined with a higher magni- 
fication, reveals a certain amount of fine granules, which in some 
cases may have completely filled the cavity, while in others 
they are confined to the outer edge, leaving a clear space in the 
center. The presence of these granules here suggests two proc- 
esses: 1) disintegration of the central protoplasm, and 2) prod- 
ucts of secretion. The lateral migration of the nuclei could be 
utilized to support either inference. That such a migration 
of nuclei would be favorable for disintegration is self evident, 
and secretory cells are usually characterized by having their 
nuclei somewhat remote from their lumina. 

In the 11 day series (figs. 36 to 38) we find that the two iso- 
lated dorsal and ventral cavities of an earlier stage have not 
only united and formed a cleft-like cavity, which may now be 
designated as the typical embryonic central canal, but that the 
original dorsal and ventral cavities of this canal, especially 
the dorsal, have increased notably in size. The narrow central 
portion (figs. 37 and 38) indicates that this is the place where the 
protoplasm was last to separate. It should not be confused 
with a similar condition that occurs later, when the central and 
ventral portions of the lateral plates migrate inward, fuse, and 


THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL. 26, NO. 1 


IS WILLIAM F. ALLEN 


completely obliterate that portion of the embryonic central 
eanal. Aside from the change in the central canal the size of the 
medulla has increased in all directions, but especially laterally, 
due doubtless to an increase in the number of nerve fibers and 
eells in the lateral plates. Both roof and floor plates are very 
thin, comprised of about one layer of nuclei each. The ventral 
plate may be slightly thicker, due to the addition of a few nerve 
fibers to the outer layer. Absolutely no stretching of the roof 
plate has occurred, indicating that it is not under any marked 
internal pressure. In figure 37 (6.V.) branches of the inter- 
segmental blood vessels have stretched out toward the roof of 
the brain, but at this stage they are too remote from the roof 
plate to be very active in infiltration. Figure 39 makes clear 
that the spinal cord has made equal progress with the medulla 
in developing a cleft-like or typical embryonic central canal. 
In the cord the dorsal and ventral enlargements of the central 
canal are shown to be of about equal size. 

An additional day (figs. 40 and 41) discloses considerable 
expansion of the central canal throughout, with the exception 
of the extreme ventral portion. Very noticeable is the increase 
in size of the dorsal portion, the future fourth ventricle (cavita 
della tela coroidea of Sterzi). The roof plate exhibits no signs 
of stretching. Up to this stage the expansion appears to be due 
to the migration of the ependymal cells upward and outward, 
rather than to pressure within. The floor plate will be seen to 
have increased considerably in thickness through an addition 
of fibers to its marginal layer, which would obviously tend to 
make the floor plate more resistant than the roof plate to pressure 
from within from this time on. Both lateral plates disclose a 
remarkable growth in thickness due to an addition of both fibers 
and cells, some of which, however, must be attributed to the 
fact that the head was sectioned somewhat obliquely (note in 
fig. 40 that the V and VIII ganglia appear in the same section). 
But little progress has occurred in the development of the inter- 
segmental blood vessels, so that Sterzi’s conjecture, that the 
embryonic cerebro-spinal fluid does not differ in any way from the 
ordinary intercellular fluids, would probably hold, if it could be 


SPINAL CORD AND MEDULLA OF CYCLOSTOMES 19 


established that the central protoplasm in disintegration con- 
tributed nothing to it. The spinal cord was found to be in about 
the same condition as the 11 day series. 

Moreover an 18 day series (figs. 42 to 44), some 6 days older 
than the last stage compared, shows practically no change in 
the shape and size of the central canal or expansion of the roof 
plate, despite the fact that a notable increase of fibers had oc- 
curred in the ventral and median portions of the lateral plates. 
It should also be recorded that the intermediate stages, as rep- 
resented by the 14 and 16 day series, exhibited a like state of 
the central canal. A further interesting condition is revealed 
from a section of the spinal cord of this series (fig. 45). Here the 
lateral plates have expanded to such an extent from the for- 
mation of fibers in the marginal layer, that the inner surfaces of 
their central areas have nearly formed a complete concrescence 
at the center of the original central canal. In other words, the 
inner surfaces of the lateral plates have met, and are about to 
fuse, leaving dorsal and ventral cavities (C.C.). The ventral 
one of these cavities will persist as a permanent central canal. 
In Petromyzon this partial closure of the original embryonic 
central canal of the spinal cord may have considerable bearing 
on the increase of internal pressure of cerebro-spinal fluid on the 
walls of the medulla, during this and later stages. We may 
regard the period from about the 12th to about the 18th day as a 
period of rest in the formation of the fourth ventricle in Petromy- 
zon. During this time the processes which would tend to expand 
the central canal have been met with equal counter forces, which 
would make for closing it up. 

Transverse sections through the medulla of a 20 day Petromy- 
zon series (figs. 46 to 48) demonstrate a marked change in the 
medulla and its central cavity, which cavity has now assumed 
the form of a fourth ventricle. A marked increase has taken 
place in the number of fibers in the ventral plate and in the 
ventral portion of the lateral plates. This together with pres- 
sure from the growing auditory vesicles (Aud.V.) and the noto- 
chord, has produced a nearly complete concrescence of the 
corresponding middle and ventral portions of the inner surfaces 


20 WILLIAM F. ALLEN 


of the lateral plates. Had a similar increase of nerve fibers 
occurred in the dorsal plate and in the dorsal portion of the 
lateral plates, the central canal of the medulla might have been 
completely obliterated. On the contrary, the dorsal cavity 
and the roof plate have been slightly expanded, apparently 
through an increased internal pressure from the embryonic 
cerebro-spinal fluid. One evidence that increased pressure 
may exist was found in the marked decrease in the size of the 
central canal of the cord and of the ventral part of the canal in 
the medulla region. 

A second evidence is found in the changes that the walls of 
the fourth ventricle are undergoing preparatory to becoming a 
functional organ for the production of cerebro-spinal fluid. In 
connection with the expansion of the roof plate it will be seen 
that the dorsal mesenchyme has become decidedly vascular 
(fig. 46, B.V.), making easy an infiltration process into the 
ventricle. Also ependymal cells, surrounding the ventricle, are 
taking on form and are probably assuming a secretory function, 
if they have not acquired one previously. From the recent 
work of Dandy and Blickfan based on the action of certain drugs 
and on the chemistry of cerebro-spinal fluid, it is evidence that 
cerebro-spinal fluid must be a product of secretion as well as of 
infiltration and diffusion. It is a well-established physiological 
fact that certain secretory cells, as for example the salivary 
glands, may assume a definite polarity, and produce a secretion 
against a very strong pressure, even stronger than that of the 
blood. Some investigators hold that lumina in glands are the 
result of pressure from secretion. 

In the light of these facts, it is fair to assume that the marked 
lateral expansion of the fourth ventricle and its roof plate 
exhibited in the 26 day series (figs. 50 to 52) are the direct result 
of internal pressure caused by the marked increase in cerebro- 
spinal fluid. The general increase in size of the ventricle to- 
gether with the marked convexity of the roof plate and the 
concavity of the internal surface of the lateral plates certainly 
suggest internal pressure from the cerebro-spinal fluid. It 
should be noted in figures 50 to 52 that this expansion of the 


SPINAL CORD AND MEDULLA OF CYCLOSTOMES | eT 


fourth ventricle has produced a considerable secondary splitting 
in the concrescence of the lateral plates recorded for the pre- 
vious series. Apparently this fissure did not penetrate so deeply 
in the region of the auditory vesicles (fig. 51, Aud.V.) as it did 
in front of them (fig. 50) or behind them (fig. 52), which probably 
indicates that the growth of the auditory vesicles in some way 
operated against the further splitting of medulla. It is signifi- 
cant that the expansion of the roof plate and the development 
of its mechanisms for infiltration and secretion occur at a time 
prior to the entrance of blood vessels into the wall of the brain 
and cord. The large production of cerebro-spinal fluid at this 
time is evidence that it serves some nutritive function. 

In figure 54 we have a median sagittal section through the 
brain of a 26 day Petromyzon, representing a stage similar to 
that of figures 50 to 53. Especial attention is called to the 
fact that figure 54 demonstrates conclusively that the pronounced 
roof expansion displayed in figures 50 to 52 has occurred without the 
aid of a pontine flerure. The slight convexity of the floor of the 
medulla can be attributed to the increase of fibers. Also earlier 
and later series revealed that Petromyzon possesses no pontine 
flexure. Attention should be called to the fact that the marked 
convexity of the cephalic end of the thin roof plate of the fourth 
ventricle (C.C.) in figure 54 gives every appearance of being 
under internal pressure from cerebro-spinal fluid. This section 
also makes clear, as Sterzi has previously shown, that the fourth 
ventricle is formed from the dorsal portion of the embryonic 
central canal; while the central canal of the cord is formed from 
the ventral portion. 

The following observations were noted in connection with the 
appearance of the fourth ventricle in Petromyzon, beginning at 
a stage when the medulla was a solid cord of undifferentiated 
nuclei in a syncytium of protoplasm: 1) The ependymal nuclei 
migrated a short distance to either side of the median sagittal 
plane leaving a narrow strip of granular protoplasm in the center. 
2) A median sagittal furrow or seam appeared in the central 
protoplasm, extending from the roof plate to the floor plate. 
3) An isolated cavity appeared at the dorsal and ventral ends 


aa WILLIAM F. ALLEN 


of this furrow. 4) The two cavities became connected, forming 
a cleft-like canal, designated as the typical embryonic central 
eanal. This canal seems to be formed by a disintegration of the 
central ends of the ependymal cells, now in the form of a syn- 
eytium. 5) A considerable increase in the size of the dorsal 
portion of the central canal occurred through the upward and 
outward migration of some of the roof plate nuclei and a disinte- 
gration of the inner protoplasm. No similar expansion of the 
ventral portion of the central canal of the spinal cord took place 
because its marginal layer became reinforced very early by the 
addition of nerve fibers. 6) Following the formation of the 
central canal of the medulla there was an increase of cells and 
fibers in the lateral walls, but for the space of about six days 
there was little change in the size of the central canal. 7) Next 
a very pronounced increase of fibers occurred in the median 
and ventral portions of the lateral plates, which brought about 
a complete concrescence of the corresponding inner margins of 
the lateral plates. 8) A sufficient amount of cerebro-spinal fluid 
was formed by infiltration and secretion to produce a marked 
expansion of the fourth ventricle. This expansion pushed apart 
the dorsal portion of the lateral plates, which had not been 
thickened by an addition of fibers, and stretched the roof plate 
to a much greater width. 9) Along with this expansion a 
secondary splitting of the concrescence noted in (7) took place. 
This fissure did not penetrate so deeply in the region of the 
auditory vesicles on account of the mechanical obstacle offered 
by these vesicles. 10) This expanded roof plate apparently 
assumed the function of producing cerebro-spinal fluid at a time 
previous to the entrance of blood vessels into the central nervous 
system, when its nutritive function would be of importance. 
From a review of the main points in the development of the 
fourth ventricle in Petromyzon we are warranted in concluding: 
1) That a well-developed fourth ventricle and tela chorioidea 
were formed in one of the lowest living vertebrates, Petromyzon, 
without the aid of a pontine flexure. 2) The best suggestion 
that has been given for the appearance of such an organ in the 
medulla rather than elsewhere in the central nervous system is 


SPINAL CORD AND MEDULLA OF CYCLOSTOMES 23 


that the roof plate has been weakened more in the medulla region 
through a greater migration of neural crest cells. This explana- 
tion, however, is not entirely satisfactory; since it was shown for 
Petromyzon (fig. 32) that the cranial ganglia were well-differ- 
entiated while the medulla was still a solid cord of cells. If then 
this were the determining factor it must have cast its shadow a 
long way ahead. Also it is apparent when the difference in size 
is considered between the medulla and the spinal cord that there 
are relatively no more neural crest cells extruded from the 
medulla. 3) The roof expansion appeared at the same time in 
the region of the X ganglion as in the region of the VIII and V, 
and developed at a uniform rate throughout the medulla. 4) 
Two factars are evident in the formation of the fourth ventricle; 
first, an upward and outward migration of the roof plate nuclei 
followed by a disintegration of the inner protoplasm; second, 
internal pressure exerted by the rapidly increasing cerebro-spinal 
fluid, infiltrated and secreted by the roof plate itself. 

D. Selachians and Amphibia. As types of these classes I had 
access to very complete sets of serial sections of Squalus and 
Amblystoma embryos, and to one transverse series of a 15 mm. 
Necturus. <A careful examination of this material contributed 
nothing new to the ontogeny of the fourth ventricle. It was 
possible, however, to confirm in these forms many of the con- 
clusions arrived at in Petromyzon. In both Squalus and Ambly- 
stoma the much earlier appearance of the central canal brings 
about a much earlier and more extensive expansion of the fourth 
ventricle. The sections shown in figures 72 and 73, while pos- 
sessing enormous expansions of the roof plate, exhibit less progress 
in the differentiation of the structure of the medulla than was 
revealed in a 12 day Petromyzon (fig. 41), in which very little 
or no expansion had taken place in the roof plate. 

In Squalus and Amblystoma the typical embryonic central 
canal, in the form of a vertical cleft, is a comparatively late 
production. The canal appears soon after the neural folds co- 
alesce as a horizontal cleft, its presence often being indicated 
only by a layer of pigment. This cleft becomes elliptical, then 
more or less circular, and finally changes to a vertical or dorso- 


24 WILLIAM F. ALLEN 


ventral cleft. These changes in the shape of the central canal 
suggest a migration of cells, rather than a disintegration process 
or a splitting apart of the walls as a result of pressure from the 
secretion of cerebro-spinal fluid. The earliest Squalus series 
showed an elliptical canal with its longest axis horizontal, this 
changed to circular, and finally to the so-called typical embryonic 
central canal (dorso-ventral cleft canal). For a time in the 
medulla of Amblystoma the roof plate is much thicker than the 
floor plate, a condition which can be attributed to the presence 
of a number of neural crest cells in the roof plate, easily distin- 
guishable from the other cells by their large size and spherical 
form. After the neural crest cells had been entirely separated 
from the roof, the roof and floor plates are found to.be about 
equally thick. The same thinning out of the roof plate occurs 
in Squalus in part through the giving off of neural crest cells, 
which in this species are indistinguishable from the other cells 
of the medulla. It is evident in Amblystoma (figs. 74 and 83, 
R.P.) and also in Squalus that the roof plate of the medulla 
has become no thinner than the roof plate of the spinal cord 
through the throwing off of the neural crest cells. There is no 
evidence that the roof plate in the medulla is rendered any 
weaker or any more susceptible to expansion than is that of the 
spinal cord, through the migration of ganglion cells. 
Apparently the first expansion of the roof plate in Ambly- 
stoma (fig. 74, R.P.) and in Squalus is produced as in Petro- 
myzon by an outward and upward migration of the roof plate 
cells. The later, more pronounced expansion and stretching of 
the roof plate can also be ascribed, as in Petromyzon, to internal 
pressure due to a decided increase of cerebro-spinal fluid (see 
fig. 72 for Squalus and 73 for Necturus). In both species the 
dorsal and middle portions of the embryonic central canal of the 
spinal cord are obliterated by an inward growth of the lateral 
plates through an addition of fibers. Pressure on the cerebro- 
spinal fluid might be increased from that source. Also a similar 
effect would be produced through a marked proliferation of fibers 
in the medial and ventral portions of the lateral plates of the 
medulla, which brings about a coalescence of the ventral por- 


SPINAL CORD AND MEDULLA OF CYCLOSTOMES ay 


tions of the lateral plates in Squalus (fig. 72, C.C.C.) similar 
to the condition in Petromyzon. In Necturus (fig. 73) and in 
Amblystoma the roof plates are expanded to such an extent that 
no fusion of the lateral plates takes place; nevertheless, the 
thickening of the walls tends to bring about a reduction in the 
size of the fourth ventricle and consequent pressure on the 
cerebro-spinal fluid. The probability of a decided increase of 
internal pressure from the cerebro-spinal fluid resulting from the 
reduction in caliber of the embryonic central canal of the spinal 
cord by at least two-thirds, is more evident, if attention is directed 
to the fact that the length of the central canal of the spinal cord 
is fully twenty times that of the fourth ventricle. In both 
forms the above changes took place before the blood vessels had 
reached the dorsal surface of the roof plate or entered the 
medulla. A median sagittal section through a Squalus embryo 
of the same stage as figure 72 shows that no pontine flexure has 
appeared. 

FE. Pig embryos. For making observations on the development 
of the tela chorioidea of the fourth ventricle in mammals the 
writer had access to a very complete set of frontal sections of 
pig embryos from a stage of 4 or 5 mm. up to 14 mm. embryos. 

The earliest section (fig. 75), which is from a 4 or 5 mm. pig, 
discloses that the fourth ventricle has only begun to expand. 
A large part of this expansion could be attributed to an outward 
and upward migration of the roof plate cells, and a part to in- 
ternal pressure from cerebro-spinal fluid, of which traces are 
beginning to appear as a coagulum (S.C./.). It should be noted 
that the first blood vessels are appearing above the basal por-~ 
tions of the roof plate, while none have at this stage entered the 
medulla. Also the ependymal cytoplasm is sufficiently granular 
to suggest a secretory process, and finally the roof plate has 
begun to expand before any nerve fibers have appeared in the 
marginal layer. 

A transverse section of the medulla of a 6 mm. pig (fig. 76) 
revealed a considerable expansion of the roof plate without the 
aid of a pontine flexure. It is apparent that an increase in the 
amount of cerebro-spinal fluid is the main factor in bringing 


« 26 WILLIAM F. ALLEN 


about this pronounced expansion of the roof plate. Since the 
closure of the dorsal portion of the embryonic central canal of 
the spinal cord occurs much later in embryonic life, no increase 
in pressure from cerebro-spinal fluid could take place from that, 
source. Also no increase was shown in the number of blood 
vessels outside the roof plate, and no blood vessels had entered 
the medulla. Consequently the only means of an increase of 
cerebro-spinal fluid would be through secretion and a slight 
infiltration from the blood vessels. This section demonstrates 
very strikingly, even more so than figure 75, that the roof ex- 
pansion begins very early in the development of the pig’s medulla, 
as is evident from the fact that nerve fibers have only begun 
to appear in the marginal layer. 

Between the 6 mm. and 7 mm. stages there occurs a marked 
increase in the blood vessels in the mesenchyme above the roof 
plate and a few blood vessels are entering the substance of the 
medulla. There can be no question that the roof plate is now 
an efficient organ for the production of cerebro-spinal fluid, and 
there is a noteworthy increase in the amount of coagulum in the 
cavity (figs. 75, 76, and 77). The first embryonic cerebro-spinal 
fluid, probably a mere intercellular fluid, showed little or no 
coagulum, from the method of fixation and staining used. This 
may indicate a relative increase of the elements formed by secre- 
tion. Figure 77 shows that a marked expansion of the roof 
plate has occurred in the 7 mm. pig. Since the 7.5 and 8 mm. 
-pig embryos have very small pontine flexures, it is evident that 
this pig embryo has developed a well-expanded roof plate and 
chorioid plexus without the influence of a pontine flexure. 

The roof plate in the 10 mm. pig (fig. 78, R.H«.) and in the 
14 mm. pig (figs. 79 and 80, R.Hz.) has undergone a decided 
expansion, especially in a dorsal direction. This final expan- 
sion of the roof plate is unquestionably due to the action of the 
pontine flexure upon a fourth ventricle filled with cerebro-spinal 
fluid already under moderate pressure. It would have been 
impossible for a pontine flexure acting alone on an empty fourth 
ventricle, as would be implied from His’ experiments with bend- 
ing an empty rubber tube slit dorsally, to have brought about 


SPINAL CORD AND MEDULLA OF CYCLOSTOMES 27 


the dorsal expansion of the roof plate exhibited in figures 79 and 
80. Apparently this expansion of the roof plate in the pig has 
been gradual, for absolutely no stretching of the roof plate has 
occurred, except in one place, namely in its central anterior 
portion (fig. 79). 

A study of these sections has disclosed a direct relationship 
between the expansion of the roof plate and the amount of visi- 
ble coagulum in the ventricle. Since coagulum does not appear 
in sections of the early fourth ventricle, but does appear after 
the tela chorioidea has attained the function of producing cerebro- 
spinal fluid (as is indicated by its vascularity and the granular 
appearance of the cells) it is fair to assume that the non-coagu- 
lable cerebro-spinal fluid found in sections of the early embryos 
is an embryonic cerebro-spinal fluid, which differs in no way 
_ from the ordinary intercellular fluid of other tissues. On the 
other hand, the coagulum seen in sections after the roof plate 
has reached the stage of a functional chorioid plexus is evidence 
of a chemical change in the fluid, which, if a product of secre- 
tion, is capable of exerting considerable internal pressure and 
consequent expansion of the roof plate. 

It is apparent that the greater expansion of the roof plate in 
the pig is produced by the same factors as were recorded for 
Petromyzon, namely, an early migration outward of the roof 
plate cells followed by an expansion from within due to the for- 
mation of cerebro-spinal fluid, plus the action of a conspicuous 
pontine flexure on a fourth ventricle filled with cerebro-spinal 
fluid already under moderate pressure. 

F. Human embryos. For this study an 8 and a 15 mm. trans- 
verse series and a 23 mm. frontal series were available. These 
embryos were too far advanced to show the earliest stages of 
the roof expansion of the fourth ventricle. If, however, the 
extreme posterior end of the roof plate of the fourth ventricle 
is examined in the 23 mm. embryo (figs. 26 and 27, R.Fx.), in 
the 15 mm. embryo (figs. 28 and 29, R.E£x.), and in the 8 mm. 
embryo (fig. 31, R.Lx.), which represents a region of the medulla 
little affected by the pontine flexure, it is apparent that the roof 
expansion was caused by identically the same factors as was 


28 WILLIAM F. ALLEN 


recorded for the pig embryos. In view of the facts, this expla- 
nation of the formation of the roof expansion seems more ten- 
able to the writer than to attribute all, as His has done, to the 
action of a pontine flexure. For beyond question, a considerable 
expansion of the roof plate in the human medulla takes place 
before the pontine flexure appears. 


2. Description of three roof plate expansions of the spinal cord in 
the 20 cm. Polistotrema series 


An interesting variation (abnormality) was found in the spinal 
cord of a single specimen of Polistotrema (Bdellostoma), other- 
wise unusual. Certain structures appeared in the roof of the 
cord that were very similar to the tela chorioidea of the fourth 
ventricle, and will be described because of the light that they 
may throw on the origin of the tela chorioidea of the fourth 
ventricle. 

What has been designated as the first roof plate expansion of 
the 20 em. Polistotrema series is shown in the photographs of 
models 1 and 2 (figs. 1 and 2, 4 and 5, R.Ex.) to be an immense 
outcropping of the roof plate ependyma. Most unfortunately 
the anterior portion of this specimen, from which the series 
through the tail was taken, has been lost, so that it is impossible 
to state how much further cephalad this expansion of the roof 
plate extended, or whether there were other outcroppings of 
the roof plate in front of it as there are behind. It appears in 
the first model (fig. 2, R.Hx.) and in the first transverse section 
(fig. 10, R.P.Ex.) as a median mass, covering about one-half of 
the dorsal surface of the spinal cord; it then shifts gradually 
over to the right side (figs. 2 and 11); then gradually attains a 
median position. In this position it continues as far as the 
middle of model 2 (fig. 5), covering a vue part of the dorsal 
surface of the spinal cord. 

For the most part the roof plate expansion contains a cavity 
of considerable size, which is shown anteriorly in the cast (fig. 
3, C.C.Ex.) and in transverse section (figs. 10 to 12) to be in 
direct communication below with the central canal. Posteriorly 


SPINAL CORD AND MEDULLA OF CYCLOSTOMES 29 


such a connection is wanting (fig. 13). The fact that in various 
places (fig. 13) there is more or less of a string of ependymal 
cells between the walls of the central canal and the roof plate 
expansion suggests that in a more embryonic state an open com- 
munication existed between the roof plate cavity and the central 
canal, which became closed in the region of the posterior end of 
the roof expansion, and persisted in the anterior end. Another 
variation to be noted in the cavity of the posterior end of the 
roof plate expansion is that it contains numerous islands and 
promontories of ependymal cells and connective tissue, shown 
as white spaces in figure 6. These probably represent portions 
of the roof plate that have not been completely excavated to 
form a continuous cavity. 

Transverse sections 11 to 13 show the cavity of this first roof 
plate expansion to be larger than the fourth ventricle in Polisto- 
trema (figs. 64 to 66), and the whole structure more nearly 
resembles a typical fourth ventricle than does the fourth ven- 
tricle itself in this animal. For the most part the walls of the 
roof plate consist of true ependymal cells, differing in no way 
from those surrounding the central canal, except for their shorter 
peripheral processes. Posteriorly connective tissue takes the 
place of many of these cells. One of these cells is shown in 
figure 13A to be sufficiently granular to suggest a secretory func- 
tion. Everywhere the walls of both the roof plate and the cen- 
tral canal are very vascular, suggesting a modified choroid 
plexus. Figure 13A will demonstrate the ease with which infil- 
tration and diffusion could take place between the blood vessels 
of the roof plate and its cavity. In figure 11 a fold of the roof 
expansion, containing a blood vessel, will be seen extending 
into the cavity, and about it there is collected a mass of coagu- 
lated cerebro-spinal fluid (S.C.F.). 

Of still greater interest are the two posterior outcroppings of 
the roof plates, designated as roof plate expansions 2 and 3 
(figs. 4 and 5, R.Ex. 2 and 3), Since these two roof plate expan- 
sions are considerably smaller than the first, they can be com- 
pared directly with the roof plate of the rhombencephalon of 
any embryo. 


30 : WILLIAM F. ALLEN 


The so-called second roof plate expansion appears in model 
2 (figs. 4 and 5, R.Ex. 2) immediately behind the first. This is 
a much smaller outcropping of the roof plate. It contains a 
cavity (figs. 16 and 17, C.C.Ez.), which spreads out a little lat- 
erally and caudally, and communicates below with the central 
canal. This cavity is filled with a fine fibrillar feltwork that 
stains deeply with orange G, and which for the most part is 
coagulated cerebro-spinal fluid. The ependeymal walls of both 
the roof expansion and the central canal are sufficiently vascular 
to suggest that we have here as in the previous roof expansion, 
a modified chorioid plexus which is producing cerebro-spinal 
fluid. 

In model 2 (figs. 4 and 5, R.Kx. 3) the third outcropping of the 
roof plate is some little distance behind the second, about equal- 
ing it in size. This roof expansion has not been figured in trans- 
verse section, but from an examination of a graphic reconstruc- 
tion of the central canal, three small isolated cavities were seen 
extending in a cephalo-caudal direction. The middle cavity was 
found to be in communication below with the central canal. It 
held cerebro-spinal fluid, and its walls apparently functioned in 
the production of the same. 

It is evident that these three expansions of the roof plate are 
independent of one another. The arrangement of the three 
small isolated cavities in the third expansion seems to be an 
embryonic condition, and suggests that the larger cavities may 
have been produced by the union of several smaller roof expan- 
sions. It may be supposed that these expansions were formed 
by the multiplication of roof plate cells, which were pushed up 
in solid masses, in which vacuolization and confluence of adja- 
cent cavities produced the larger cavities seen in the adult. 

It is of interest to note that in this individual a posterior sinus 
(fig. 20, S.T.), probably representing the sinus terminalis of a 
normal individual, was isolated by the complete occlusion of the 
central canal by ependymal tissue. The ependyma surrounding 
this cavity is very vascular, and the cavity is distended with 
cerebro-spinal fluid. This sinus is much larger than the fourth 
ventricle in a normal individual. 


SPINAL CORD AND MEDULLA OF CYCLOSTOMES 31 


When the roof expansions in the spinal cord of this series of 
Polistotrema are compared with the fourth ventricle of the higher 
vertebrate embryos, it is evident that the similarity is only super- 
ficial, for the later stages in the medulla oblongata (action of 
pontine flexure on a thin-roofed neural tube full of cerebro-spinal 
fluid under moderate pressure), are dependent on factors which 
are not present in the spinal cord of Polistotrema. The extreme 
caudal end of the roof plate of the fourth ventricle in the higher 
vertebrates, however, has not been affected by these later fac- 
tors, and presents a condition where comparison is made possible. 
From a comparison of the roof plate expansion in the 23 mm. 
human embryo (figs. 26 and 27, R.£x.), in the 15 mm. human 
embryo (figs. 28 and 29, R.#2x.), and in the 8 mm. human embryo 
(fig. 31, R.Ex.) with the second roof plate expansion of the 
spinal cord inthe 20 em. Polistotrema series (figs. 4 and 5, R. 
Ke. 2, and figs. 15 to 17, R.P.Ex. 2) and with the third roof plate 
expansion of the same series (figs. 4 and 5, R./x. 3), it is evident 
that the same main factor is present in all, namely, a migration 
outward and upward of certain roof plate ependymal cells to 
form an enlarged dorsal cavity. In each case the purpose of 
this structure is to form an organ for the production and storing 
of cerebro-spinal fluid. As soon as these structures assumed 
the function of infiltrating and secretory organs their walls 
became further expanded from internal pressure of the cerebro- 
spinal fluid. 

From a review of the early stages of the formation of the roof 
expansion in man, the pig, Amblystoma, Squalus, and Petro- 
myzon, it is fair to assume that the source and early development 
of the roof expansion of the medulla are identical to the three 
similar roof plate expansions, described for the spinal cord of a 
20 em. Polistotrema. It is possible that both had their phylo- 
genetic anlage as mutations, the former from some primitive 
vertebrate, and the latter from a normal Polistotrema, that both 
were useful and dominant characters, and in case of the medulla, 
where the animal was allowed to reproduce, this character became 
preserved for the race. 


Ww 
bo 


WILLIAM F. ALLEN 


CAUSES UNDERLYING THE FLATTENING OF THE SPINAL 
CORD IN CYCLOSTOMES 


If our attention is first directed to a transverse section through 
an adult Polistotrema (figs. 10 and 71) it might be inferred, since 
there is ample room in the membranous neural canal for a well- 
rounded spinal cord, that the flattening of the spinal cord might 
be attributed entirely to internal factors. A glance at a trans- 
verse section through a developing Polistotrema spinal cord 
(fig. 57) suffices to show that there is proportionately much less 
room within the membranous neural canal, and that a mechani- 
eal force in the form of a rapidly growing notochord is at work 
immediately below the spinal cord. 

At the outset it seems advisable to establish arbitrarily a typi- 
cal state of an embryonic spinal cord, by which a direct com- 
parison of one form can be made with another. The examina- 
tion of very early stages of the spinal cord in a large number of 
embryos of Squalus, Amblystoma, the chick, and the pig, in 
all of which the neural tube is formed by the rolling up of the 
neural plate, shows that the neural tube passes through three 
stages: a) A depressed tube with the central canal in the form 
of a horizontal cleft; b) a cylindrical tube with the canal circu- 
lar in cross section; and c) a laterally compressed tube with the 
canal in the form of a vertical cleft (figs. 81 to 86). The exist- 
ence of this series of changes in Squalus has been shown in a 
table of developmental stages compiled by Seammon. The third 
stage may be selected as the typical embryonic spinal cord. 
This stage is reached at about the time of the first appearance 
of nerve fibers in the marginal layer. 

As a result of a comparison of the typical embryonic stages of 
the spinal cord in the following transverse sections’ (fig. 39 for 
Petromyzon, fig. 55 for Polistotrema, fig. 81 for Squalus, fig. 
83 for Amblystoma, fig. 84 for the turtle, fig. 85 for the chick, 
and fig. 86 for the pig) it is clear that we have all gradations from 


* It is obvious that this comparison would have no value unless the sections 
were truly transverse sections. To avoid selecting oblique transverse sections, 
these figures were always drawn from anterior trunk sections if the embryo 
showed any flexures. 


SPINAL CORD AND MEDULLA OF CYCLOSTOMES 33 


the nearly cylindrical spinal cords of the Cyclostomes (Petro- 
myzon and Polistotrema) to the very much compressed (flat- 
tened laterally) cords of the chick and pig. Were it not for the 
intermediate stages of Amblystoma and the turtle we might be 
justified in establishing two distinct types of the embryonic 
spinal cord: type 1 cylindrical, and type 2 compressed. We 
could even go further and classify the Cyclostome embryonic 
cord under type 1 and the Gnathostome cord under type 2. 

While the equal expansion of the spinal cord through the addi- 
tion of nerve cells and fibers to the typical embryonic stage in 
the pig would tend to produce a cylindrical spinal cord, and in 
the Cyclostome would tend to produce a depressed cord, the 
internal structure shows that the origin of the differences between 
the spinal cord of a Cyclostome and a mammal is not so simple. 
The neural axis of Petromyzon in an early stage, corresponding 
to stage (a) above, is decidedly compressed instead of depressed. 
The problem then is to explain the change from a laterally com- 
pressed cord in the early Petromyzon embryos to the gradually 
depressed, ribbon-like, spinal cord of the adult. 

A careful examination of figures 55 to 57, which are taken from 
practically the same region from three different Polistotrema 
embryos, shows very clearly that the growing notochord is 
bringing about the marked flattening (depression) and ventral 
indenting of the spinal cord. In connection with figure 55 it 
should be noted that the spinal cord is in the so-called typical 
embryonic stage, and amid surroundings peculiarly favorable for 
undergoing a depression from a rapidly growing notochord. It 
will be seen that the spinal cord is closely enveloped by a men- 
ingeal membrane (P.M.), approximating a layer of connective 
tissue, the future membranous neural arch, which is firmly at- 
tached below to the growing notochord. Directly above, the 
mesenchyme is proliferating and migrating toward the center 
to form a median dorsal cartilage (W7.D.C.); while there is appar- 
ently but little lateral resistance in the way of massing of mesen- 
chyme and the formation of myotomes, In a somewhat later 
stage (fig. 56) some growth has taken place in the notochord. 
producing a slight indentation on the spinal cord. In a much 


THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL. 26, NO. 1 


34 WILLIAM F. ALLEN 


later stage (fig. 57) a median dorsal cartilage has been formed. 
The much stronger membranous neural arches are firmly attached 
above to this dorsal eartilage and below to the notochord. The 
soft plastic spinal cord is thus closely confined by the dorsal 
cartilage above and less closely by the neural arches laterally. 
The notochord beneath it has grown very rapidly and its enor- 
mous increase in size has brought about a decided flattening 
(depression) of the spinal cord and indentation of its ventral 
surface. 

It is equally clear in Petromyzon also that the growing noto- 
chord is to be looked upon as a direct cause for bringing about 
the flattening of the spinal cord. The external conditions sur- 
rounding the spinal cord are shown in figure 58 to be equally 
favorable for assisting the notochord in this process, with the 
possible exception that the membranous neural arch is attached 
above to a membranous neural spine instead of a cartilage, which, 
however, may be compensated for by an increased dorsal growth 
of the myotomes. 

The hypothesis that the flattening of the spinal cord in Cyelo- 
stomes is largely brought about by the upward growth of the 
notochord after the manner set forth in the previous paragraph 
is considerably strengthened by the fact that a certain relation- 
ship exists between the size of the notochord and the amount 
of flattening of the spinal cord. 

This is clear in the 20 em. Polistotrema series from sections 
of the posterior end of the spinal cord (figs. 19, 21, and 22) and 
from the photographs of the model of the same region (figs. 7 
to 9); the flattening (depression) of the spinal cord becoming 
less evident as the notochord decreases in size. More striking 
is a similar relationship shown in the tail region of the 70 mm. 
Polistotrema (figs. 59 and 60) for the reason that these two sec- 
tions were only one-fourth of a millimeter apart. The above 
relationship between the size of the notochord and the depres- 
sion of the spinal cord can be demonstrated fully as conelu- 
sively in the medulla region (see figs. 61 and 62). It should be 
recorded for these two sections that their structure is the same 
as that of the spinal cord and that they are less than one-half 


SPINAL CORD AND MEDULLA OF CYCLOSTOMES 35 


a millimeter apart. Also the same relationship could be shown 
in sections anterior to figure 62, and from similar sections of a 
Petromyzon larva. A possible objection to applying this argu- 
ment in the tail region might arise from the fact that the extreme 
posterior end of the spinal cord is non-nervous, consisting en- 
tirely of ependyma and undifferentiated embryonic cells. In 
reply to this we would invite comparison of figures 21 and 22, 
where the structure is non-nervous in both cases, and where the 
effect of the notochord is obvious. 

That the spinal cord of the higher vertebrates has not been 
depressed by pressure from the notochord is due obviously to 
the fact that the notochord is an embryonic structure, which 
never attains sufficient size to have any influence on the spinal 
cord. This is clear from figures 85 and 86. In Cyclostomes, 
however, the notochord is a very important structure, develops 
sxarly and grows for a long period of time, and serves as the 
skeletal axis of the adult. In fishes, Amphibia, and reptiles the 
growing notochord may have some slight effect in flattening of 
the adult spinal cord. Im Amphioxus the ventral surface of the 
spinal cord clearly shows the indenting effect of the growing 
notochord (fig. 70), and in the trunk region where the diameter 
of the notochord is greatest, the spinal cord is most depressed. 


GENERAL CONSIDERATIONS AND SUMMARY 


From the foregoing facts the following conclusions seem fully 
warranted : 

1) In the development of the roof plate expansion (tela 
chorioidea) in the medulla oblongata of most vertebrates three 
separate stages or epochs of expansion should be recognized: a) 
A first enlargement of the dorsal portion of the embryonic cen- 
tral canal took place from a migration outward and upward of 
certain of the roof plate cells, or as was the case in Petromyzon 
from the migration of the nuclei and probable disintegration of 
the cytoplasm. b) The second stage in the expansion of the 
roof plate was the direct result of an increase of pressure from 
the cerebro-spinal fluid, produced from at least two possible 


36 WILLIAM F. ALLEN 


sources. First in the lower vertebrates, as a consequence of con- 
siderable embryonic cerebro-spinal fluid being forced into the 
fourth ventricle from the closure of the dorsal and central por- 
tions of the embryonic central canal in the spinal cord and the 
ventral portion of the embryonic central canal in the medulla, 
through the union and fusion of the corresponding portions of 
the lateral plates. Second, in all vertebrates, through the pro- 
duction of cerebro-spinal fluid by the walls of the fourth ventricle 
assuming the role of infiltration, diffusion and secretion. That 
the ventral portion of the embryonic central canal of the spinal 
cord was not expanded by the same cause is explained by the 
fact that it was reinforced at a very early stage by nerve fibers 
and supported by a growing notochord. c) A third and final 
stage in the expansion of the roof plate in the higher vertebrates 
was brought about by the appearance of a pontine flexure acting 
on a thin-roofed medulla filled with cerebro-spinal fluid, itself 
under moderate pressure. 

2) It is the opinion of the writer that the most important 
factors in bringing about the expansion of the roof plate of the 
medulla are those concerned in the second stage described above. 
For a considerable period of time these factors apparently work 
in conjunction with the forces of the first stage in resisting counter 
ingrowths of the lateral plates which would tend to close up the 
central canal. Since it was shown that no pontine flexure oc- 
cured in Petromyzon, it can be concluded that the rather exten- 
sive expansion of the roof plate in the medulla of this genus was 
accomplished solely through the factors entering into the first 
and second stages. For the second stage it was recorded in 
most cases, especially well shown in the pig, that the size of the 
fourth ventricle and the expansion of its roof plate bear a close 
relationship to the amount of coagulum seen in sections. Since 
this coagulum did not appear in sections until after the roof 
plate had attained the function of producing cerebro-spinal fluid, 
it probably indicates a chemical change in the cerebro-spinal 
fluid, the products of secretion now being added to the tissue 
fluids which entered the ventricle by infiltration. As a product 
of secretion, the cerebro-spinal fluid would be capable of exerting 


- 


SPINAL CORD AND MEDULLA OF CYCLOSTOMES By: 


considerable pressure on the thin and plastic roof plate. The 
tela chorioidea is differentiated as an organ for the production 
of cerebro-spinal fluid before blood vessels have entered the 
central nervous system, at a time when the nutritive function 
of this fluid is important. 

3) The appearance of a third stage in the roof expansion of 
the medulla, due to a pontine flexure, is of little significance 
save in the higher vertebrates, where it was held that without 
cerebro-spinal fluid confined in the ventricle there would be no 
reason for maintaining that a further expansion of the roof plate 
would take place from the action of a pontine flexure; more than 
likely, the roof plate would have been folded up within the ven- 
tricle. In His’ experiment with the bending upward of a dor- 
sally-slit piece of rubber tubing, the elasticity of the rubber 
tubing, which forced apart the cut surfaces, would be comparable 
to the action of the cerebro-spinal fluid under moderate pressure 
within the ventricle, which factor His has apparently disregarded. 

4) In the adult Amphioxus there is nothing which for a cer- 
tainty could be homologized to the fourth ventricle and its ex- 
panded roof plate. Two isolated cavities in the region of the 
roof plate, which might be taken for the anlage of the fourth 
ventricle, appear to the writer to be nothing more than vestiges 
of a much larger embryonic central canal. If Amphioxus pos- 
sesses no fourth ventricle in the adult we may safely conjecture 
that more primitive vertebrates had a central nervous system 
in which there was no distinction between medulla and spinal 
cord. 

5) In an attempt to trace the phylogenetic history of the roof 
expansion of the fourth ventricle in living vertebrates, the pecu- 
liar modification of the fourth ventricle in the adult Polistotrema 
(Bdellostoma) should be recorded here even though it has been 
accurately described by Sanders, Holm, Miss Worthington, 
Sterzi, Cole, and Nicholls. From the adult it is evident that the 
well-formed fourth ventricle of the embryo has become trans- 
formed through a process of centralization to a deep-seated canal, 
for the most part no larger than the central canal of the spinal 
cord. Of especial interest is the fact that its anterior and pos- 


38 WILLIAM F. ALLEN 


terior portions has developed into specialized organs for the pro- 
duction of cerebro-spinal fluid. Notwithstanding this speciali- 
zation, the fourth ventricle is thought to be decidedly inferior 
to the tela chorioidea of Petromyzon as an organ for the produc- 
tion of cerebro-spinal fluid. 

6) In the spinal cord of one individual of Polistotrema there 
oecurred at least three expansions of the roof plate which re- 
semble the roof of the fourth ventricle in other vertebrates. 
From the fact that these expansions were very vascular and 
their cells granular it is inferred that they functioned as cho- 
roid plexuses for the formation of cerebro-spinal fluid. The 
writer presents the hypothesis that the fourth ventricle in ances- 
tral vertebrates may have originated as a mutation, similar to 
this sport plexus in the spinal cord of Polistotrema; that such 
sport expansions may have occurred at various places such as 
the diencephalic segment, the roof of the mesencephalon where 
a choroid plexus still exists in Petromyzon, and in the hind 
brain and spinal cord. Such mutations, proving to be useful 
have been preserved in the vertebrate race. 


Concerning the flattening of the spinal cord in Cyclostomes 


7) A great variation in the shape of the so-called typical 
embryonic spinal cord is to be recorded. In Petromyzon it was 
found to be nearly cylindrical, to be moderately compressed in 
Squalus, Amblystoma, and in the turtle, and decidedly com- 
pressed in the chick and the pig. 

8) To obtain this typical stage the original compressed spinal 
cord of Petromyzon must have undergone a marked depression, 
and the early depressed neural tubes of Squalus, Amblystoma, 
turtle, chick, and pig must have undergone a decided com- 
pression. The main factor causing this depression in the former 
was thought to be ventral pressure from a growing notochord, 
and the compression of the latter was attributed to lateral pres- 
sure from the growing myotomes. 

9) Transverse sections immediately before and during the time 
that the greatest depression of the spinal cord is taking place 


SPINAL CORD AND MEDULLA OF CYCLOSTOMES 39 


in Polistotrema and Petromyzon show conclusively that the 
main factor involved is the pronounced growth of the notochord. 
It was further established that the embryonic spinal cord was 
not only in a very plastic condition, but that the general environ- 
ment was decidedly favorable for bringing about a depression 
of the spinal cord through this agency. 

10) The conclusions outlined in (9) were considerably strength- 
ened by the fact that a direct relationship was established, in 
both the medulla and tail region, between the size of the noto- 
chord and the amount of depression exhibited in the spinal cord. 
This was shown in late embryos in both Polistotrema and Petro- 
myzon, and in adult Polistotrema. 

11) That a similar depression did not take place in the higher 
vertebrates from a growing notochord was explained by the 
fact that the notochord is relatively a transitory and insignificant 
structure; while in the Cyclostomes it is not only formed early 
in embryonic life, but grows rapidly and continuously. 


40 WILLIAM F. ALLEN 
LITERATURE CITED 


Autporn, F. 1882 Zur Neurologie der Petromyzonten. Géttinger Nachrich- 
ten. 
1883 Untersuchungen iiber das Gehirn der Petromyzonten. Zeitschr. 
f. wiss. Zool., Bd. 39. 

Ayers, H. 1892 Vertebrate cephalogenesis. II. Contribution to the mor- 
phology of the vertebrate ear. Jour. Morph., vol. 6. 
1906 The unity of the gnathostome type. Amer. Nat. 
1907 Vertebrate cephalogenesis. III. Amphioxus and Bdellostoma. 
Cincinnati. 
1908 The ventricular fibers of the brain of myxinoids. Anat. Anz., 
Bd. 32. 

Ayers, H. and Worruineton, J. 1908 The finer anatomy of the brain of 
Bdellostoma dombeyi. Am. Jour. Anat., vol. 8. 

Batrour, F. M. 1885 Memorial edition of the works of Francis Maitland 
Balfour. Vol. I. London. 

Beccart, N. 1909 Le cellule dorsali o posteriori dei Ciclostomi. Ricerche 
nel Petromyzon marinus. Monit. zool. Ital. 

Catperua, E. 1877 Zur Entwicklung des Medullarrohres und der Chorda 
dorsalis der Teleostier und der Petromyzonten. Morph. Jahrb., 
Bd. 3. 

Cots, F. J. 1918 A monograph on the general morphology of the myxinoid 
fishes, based on a study of Myxine. Pt. V. The anatomy of the 
gut and its appendages. Trans. Royal Soc. Edinburgh. 

Danpy and Buickran. 1914 Internal hydrocephalus. An experimental, clin- 
ical, and pathological study. Amer. Journ. Diseases of Children. 

Davenport, ©. B. 1895 Studies in morphogenesis. IV. A preliminary cata- 
logue of the processes concerned in ontogeny. Bull. Mus. Comp. 
Zool. Harvard Univ., vol. 27. 

Davin, J. 1894 On the histological structure of the medulla of Petromyzon. 
Jour. Comp. Neur., vol. 2. 

Dean, B. 1899 On the embryology of Bdellostoma stouti. Festchrift zum 
70 Geburtstage von Kupffer. Jena. 
1906 Chimaeroid fishes and their development. Carnegie Inst., 
Washington, D.C. 

Doaiet, A. S. 1902 Das periphere Nervensystem des Amphioxus. Anat. 
Hefte., Bd. 66. 

Epineer, L. 1906 Ueber das Gehirn von Myxine glutinosa. Anhang. Abh. 
d. k. preuss. Acad. d. Wiss. 

Encet, E. A. 1908 Ueber die Secretionserscheinungen in den Zellen der 
plexus chorioidei des Menschen. Arch. Zellforsch., Bd. 2. 

Franz, V. 1912 Beitrag zur Kenntnis des Ependyms im Fischgehirn. Biol. 
Centralbl., Bd. 32. 

GasKELL, W. H. 1890 On the origin of vertebrates from a crustacean-like 
ancestor. Quart. Journ. Micr. Sci., vol. 31. 

Gorrre, A. 1890 Entwicklungsgeschichte des Flussneunauges, Petromyzon 
fluviatilis. Leipzig. 


SPINAL CORD AND MEDULLA OF CYCLOSTOMES 41 


GREENE, C. W. 1900 Contributions to the physiology of the California hag- 
fish, Polistotrema stouti. 1. The anatomy and physiology of the 
caudal heart. Am. Jour. Physiol., vol. 3. 

Hasse, C. 1876 Zur Anatomie des Amphioxus lanceolatus. Morph. Jahrb., 
Bd. 1. 

HarscHek, B. 1881 Studien tiber Entwicklung des Amphioxus. Arb. Inst. 
Wein. 
1888 Ueber den Schichtenbau von Amphioxus. Anat. Anz. 

Hertwie, O. 1892 Embryology of man and mammals, Eng. Trans. by E. L. 
Mark. 

His, W. 1865 Ueber ein perivasculires Canalsystem in den nervésen Central- 
organen und iiber dissen Beziehungen zum Lymphsystem. Zeitschr. 
feowise», Zool.,.Bd. 15. 
1874 Unsere K6érperform und das Physiologische problem ihrer Ent- 
stehung Leipzig. 
1904 Die Entwicklung des menschelichen Gehirns wihrend der ersten 
Monate. Leipzig. — 

Horm, J. T. 1901 The finer anatomy of the nervous system of Myxine glutin- 
osa. Morph. Jahrb., Bd. 29. 

Jutin, C. 1887 Recherches sur l’appariel vasculaire et le systéme nerveux 
périphérique de l’Ammoceetes. Arch. de Biol., T. 7. 

Jounston, J. B. 1905 The cranial and spinal ganglia and the vicero-motor 
roots in Amphioxus. Biol. Bull., vol. 9. 

K6uturkER, A. 1896 Handbuch der Gewebelehre des Menschen. Bd. II. 

Kotmer, W. 1905 Zur Kenntniss des Riickenmarks von Ammoceoetes. Anat. 
Hefte., Bd. 29. 

Kowauevsky, A. 1867 Entwicklungsgeschichte des Amphioxus lanceolatus. 
Mem. Acad. St. Petersburg. 
1867 Weitere Studien iiber die Entwicklungsgeschichte des Amphi- 
oxus lanceolatus. Arch. mikr. Anat. 

IKtuprrer, C. von 1890 Die Entwicklung von Petromyzon planeri. Arch. 
mikr. Anat. 
1899 Zur Kopfentwicklung von Bdellostoma. Gesellschaft f. Morph. 
u. Physiol., Miinchen. 

Kurcutn, Miss H. L. 1913 Studies on the peripheral nervous system of 
Amphioxus. Proc. Amer. Acad. Arts. Sci., vol. 49. 

LANGERHANS, P. 1876 Zur Anatomie des Amphioxus lanceolatus. Arch. 
mikr. Anat., Bd. 12. 

LankestEeR, E. Ray 1899 Contributions to the knowledge of Amphioxus 
lanceolatus. Quart. Jour. Mier. Sci., vol. 29. 

MarsuHaty, A. M. 1893 Vertebrate embryology. London. 

Meek, W. J. 1907 A study of the choroid plexus. Jour. Comp. Neur., vol. 
1%. 

M’Intosu and Prince 1888 On the development and life-histories of the teleos- 
tean food- and other fishes. Trans. R. Soc. Edin. 

Mier, J. 1834-1843 Vergleichende Anatomie der Myxinoiden. Abhandl. 
Akad. Berlin, Bd. 3. 


42 WILLIAM F. ALLEN 


Nansen, I. 1886 Fore lobig Meddelelse om Underségelser over Centralnerve- 
systements. Bergens Museum Arberetning. 
Nicuouuts, G. BE. 1912 The structure and development of Reissner’s fiber 
and the sub-commissural organ. Quart. Jour. Micro. Sei. 
OwssANNIKOow, P. 1867 Ueber das Centralnervensystem des Amphioxus 
lanceolatus. Bull. Acad. St. Petersburg, Bd. 6. 

Parker, G. H. 1908 Sensory reactions of amphioxus. Proc. Amer. Acad. 
Arts and Scei., vol. 43. 

Prarr, Miss J. B. 1892 Fibers connecting the central nervous system and 
chorda in Amphioxus. Anat. Anz., Bd. 7. 

Price, G. C. 1896 Zur Ontogenie eines Myxinoiden. Sitzungsb. b. bayer 
Acad. d. Wissenshaften, Bd. 26. 
1896 Some points in the development of a myxinoid (Bdellostoma 
stouti Lockington). Verhandlungen d. Anat. Gesellschaft a. d. 
zehnten Versammlung in Berlin. 

Rerzius, G. 1891 Zur Kenntniss des centralen Nervensystems von Myxine 
glutinosa. Biol. Untersuch., N. F., Bd. 2. 
1893. Das Gehirn und das Auge von Myxine. Biol. Untersuch., N. F., 
deeb: 
1893 Studien iber Ependym und Neuroglia. I. Ependym und Neurog- 
lia bei den Cyclostomen. Biol. Untersuch., N. F., Bd. 5. 
1895 Das hintere Ende des Riickenmarkes und sein Verhiltniss zur 
Chorda dorsalis bei Amphioxus lanceolatus. Biol. Untersuch., N. F., 
Bd. 7. 
1895 Ueber das hintere Ende des Riitckenmarkes bei Amphioxus, 
Myxine, und Petromyzon. Biol. Untersuch., N. F., Bd. 7. 

Rouper, EK. 1888 Histologie Untersuchungen iiber das Nervensystem vom 
Amphioxus lanceolatus. Schneider’s Zool. Beitrage. 

Rouon, J. V. 1882. Untersuchungen iiber Amphioxus lanceolatus. Denkschr. 
Akad. Wien. 

Rouen, W. 1876 Untersuchungen iiber den Bau des Amphioxus lanceolatus. 
Morph. Jahrb. 

Ryper, J. A. 1884 <A contribution to the embryography of osseous fishes with 
special reference to the cod (Gadus morrhua). U. 8. Fish. Comm. 
Report for 1882. 

Sanpers, A. 1894 Researches in the nervous system of myxine glutinosa. 
London. 

SarGeNT, P. E. 1904 The optic reflex apparatus of vertebrates for short- 
circuit transmission of motor reflexes through Reissner’s fiber. Bull. 
Mus. Comp. Zool. Harvard, vol. 45. 

Scammon, R. E. 1911. Normentafeln zur Entwicklungsgeschichte der Wirbel- 
tiere. XII. Normal plates of the development of Squalus acanthias. 
Jena. 

Scorr, W. B. 1881 Preliminary account of the development of the lampreys. 
Quart. Journ. Micro. Sci. 
1888 Notes on the development of Petromyzon, Jour. Morph., 
voll, 1: 


SPINAL CORD AND MEDULLA OF CYCLOSTOMES 43 


ScHNEIDER, A. 1880 Ueber die Nerven von Amphioxus, Ammoccetes und 
Petromyzon. Zool. Anz. 

SHILtinGc, K. 1907 Ueber das Gehirn von Petromyzon fluviatilis. Abh. 
Senkenberg nat. Ges., Bd. 30. 

SuipLey, A. E. 1886 On the development of the nervous system in Petromy- 
zon fluviatilis. Proc. Cambridge Phil. Soc., vol. 5. 

Srerzi1, G. 1907 II sistema nervosa centrale die vertebrati. I. Ciclostomi. 
Padova. 

SrupnicKa, F. K. 1895 Ueber die terminale Partie des Riickenmarkes. Seit- 
zungsber. d. K. Boh. Gesellschaft d. Wissenshaften. 
1900 Untersuchungen iiber den Bau des Ependyms der Nervésen 
Centralorgane. Anat. Hefte. 

Tretsakorr, D. 1909 Das Nervensystem von Ammoccetes. I. Das Riicken- 
mark. II. Gehirn. Arch. mikr. Anat., Bd. 73. 
1913 Die zentralen Sinnesorgane bei Petromyzon. Arch. mikr. 
Anat., Bd. 83. 

Tuyne, F. W. 1911 Theanatomy of a7.83mm. pigembryo. Anat. Rec., vol. 5. 

Vocr und Yuna 1894 Lehrbuch der praktischen vergleichenden Anatomie. 
Braunschweig. Bd. II. 

WIEDERSHEIM, R. 1880 Das Gehirn von Ammoccetes und Petromyzon planeri. 
Jena zeitschr., Bd. 14. 

Wine, J. W. van’ 1893 Ueber Amphioxus. Anat.. Anz., Bd. 8. 

Witiey, A. 1891 The later larval development of Amphioxus. Quart. Journ. 
Micro. Sci., vol. 32. 

Wourr, M. 1907 Bemerkungen zur Morphologie und zur Genese des Amphi- 
oxus-Riickenmarkes. Biol. Centralbl., Bd. 27. 

WortHINGTON, Miss J. 1905 The descriptive anatomy of the brain and cran- 
ial nerves of Bdellostoma dombeyi. Quart. Journ. Micro. Sci., 
vol. 49. 


PLATE 1 
EXPLANATION OF FIGURES 


lL to 9 represent photographs from models of the extreme posterior end of 
the spinal cord from a 20 em. Polistotrema (Bdellostoma), illustrating a series 
of three extensions of the central canal into roof plate expansions. These models 
were prepared at a magnification of 100 diameters, and were reduced one-half 
in photographing. Ina few cases where certain outlines were somewhat indistinct 
in the photographs they were strengthened with pen and ink. The roof plate 
expansions in these models were painted lighter and appear the same in the 
photographs. Except at the posterior end of the last model, all of the models 
of the spinal cord exhibit a marked depression and in ventral views they show 
a pronounced indentation at the center immediately above the notochord. As 
indicated by an arrow, the caudal direction in figures 1 to 3 is toward the left; 
while in figures 4 to 9 it is toward the right. 

1 The most anterior of the four models seen from the right side. It includes 
a distance of about two segments. Observe that the roof plate expansion, which 
covers a large portion of the dorsal surface of the spinal cord, is rather low. It 
is not known where it begins, or whether there are other outcroppings in front of 
it as there are behind. It extends caudad some little distance on this model. 
On the dorsal surface of the cord one dorsal or sensory root is shown, and on the 
ventral side one ventral root is seen in entirety, being composed of several root- 
lets. X 50. 

2 Dorsal view of the same model shown in figure 1, seen from above. Note that 
the roof expansion covers a large portion of the spinal cord. At the anterior end 
it occupies a large central portion, then becomes gradually smaller, and at the 
same time is confined largely to the left side, after which it gradually increases 
in size until the posterior end of the model is reached, where it is decidedly wider 
than at the anterior end, and extends farther over to the right side than to the 
left. X 50. 

3 Ventral view of the cast of the central canal and extension of the same up and 
out into the roof expansion of the same region of the cord as is shown in figures 
land 2. It will be seen that the central canal and cavity of the roof plate expan- 
sion are connected throughout. In certain places on the right side, shown in 
white, the roof plate is solid, consisting of ependymal and connective tissue in 
place of a cavity. For some distance posteriorly on the right side, where the roof 
expansion is widest, there is no cavity in the roof expansion. The knob-like 
projections from the right side of the cast of the central canal represent diver- 
ticula, and in sections throughsuch a region there would appear to be two central 
canals. X 50. 

ABBREVIATIONS 


C.C., central canal or cast of the same M.R., motor or ventral spinal nerve 


C.C.Ex., central canal extension into root 
roof plate expansion or cast of the R.Ex., roof plate expansion 
same S.R., sensory or dorsal spinal nerve 
root 


44 


SPINAL CORD AND MEDULLA CF CYCLOSTOMES PLATE 1 
WILLIAM F. ALLEN 


7 ¢ ' 
hala 7 a 
PERN ot 
TN Fale a a 


‘fear “ 


PLATE 2 
EXPLANATION OF FIGURES 


4 Lateral view of the left side of the second model of the Polistotrema spinal 
cord, which should follow figure 1. Observe the continuation of the first roof 
plate expansion noted in figure 1 and two additional outcroppings of the roof 
plate (R. Er. 2 and 3). Between these outcroppings the spinal cord is perfectly 
normal. Three motor and sensory roots are shown in the figure. X 50. 

5 Dorsal photograph of the same model as figure 4. The extent and posi- 
tions of the three roof plate outcroppings previously mentioned in the description 
of figure 4 are well portrayed here, as are also the sensory roots. 50. 

6 Dorsal view of a cast of the more caudal portion of the central canal and 
cavity of the first roof plate expansion. This cast differs from the more anterior 
cast in that there are no connections between the central canal and the cavity 
of the roof expansion. Evidence from transverse sections favors the view that 
a connection once existed, which has been cut off in later development. In the 
two posterior outcroppings of the roof plate there is a connection between the 
central canal and the cavity of the roof expansion. The light places in the photo- 
graph are indicative of places in the roof expansion where it is solid and con- 
tains no cavity. They are more numerous and decidedly larger than was shown 
in the more anterior model (fig. 3). X 50. 

Between models 2 and 4 (figs. 4 and 7) comes model 3, the photograph of which 
has not been included as a figure. It is about equal in length to models 2 and 4, 
simply connecting the two, without presenting any peculiarities in roof plate 
expansion, etc. 

ABBREVIATIONS 


M.R., motor or ventral spinal nerve R.F#x.(3), third roof plate expansion 


root in the Polistotrema cord 
R.Ex., roof plate expansion S.R., sensory or dorsal spinal nerve 
R.Ex.(2), second roof plate expansion root 


in the Polistotrema cord 


46 


PLATE 2 


SPINAL CORD AND MEDULLA OF CYCLS ISTOMES 
WILLIAM F. ALLEN 


RE x. 


mit m | rpnVal 


PLATE 3 
HXPLANATION OF FIGURES 


7, 8, and 9 represent lateral, dorsal, and ventral photographs of the fourth 
model, which includes the extreme posterior end of the spinal cord. Note es- 
pecially the swelling above and below (S.7’.) caused probably by an abnormal 
sinus terminalis. From this point caudad two factors occur, which may greatly 
modify the shape of the spinal cord. First, the notochord gradually decreases 
in caliber and ends at (Ne. 1); second, the spinal cord has not developed a nervous 
structure, consisting solely of ependyma and round undifferentiated embryonic 
cells. As a result the spinal cord will be seen to gradually become rounded, 
and after extending past the notochord it ends in dorsal and ventral processes 
that become lost in the surrounding connective tissue. The posterior motor 
roots exhibit a reduction in the number of rootlets, and they approximate each 
other more closely. The posterior sensory roots become greatly reduced in size, 
and the last left one has no corresponding roots on the opposite side. 50. 


ABBREVIATIONS 


M.R.(1), last motor or ventral spinal S.R. (1), last sensory or dorsal spinal 
nerve root nerve root 

Ne. (1), posterior end of Polistotrema  S.7., sinus terminalis 
notochord shown in model 


PLATE 4 
EXPLANATION OF FIGURES 


10 to 23, show 14 transverse sections through the posterior spinal cord region 
of a 20 cm. Polistotrema (Bdellostoma), the same as was modelled and shown in 
figures 1 to9. The outlines were all drawn with the aid of a Leitz-Edinger draw- 
ing apparatus, using a magnification of 140 diameters and were reduced one- 
half in reproduction. They are numbered consecutively from anterior to posterior. 
Figures 10 to 14 pass through what has been designated as the first roof plate 
expansion; figures 15 to 17 through the second roof plate expansion; figure 20 
through the abnormal sinus terminalis; and figures 21 to 23 at various intervals 
through the extreme posterior end of the spinal cord, which has failed to develop 
any nervous structures, and has not to any extent been flattened in its develop- 
ment by the growth of the notochord. The enormous concavity seenon the ven- 
tral surface of spinal cord in figures 11 to 14 is to a large extent an artifact due to 
the preparation of the series. It should be noted that the cavities of the roof 
plate expansions are full of a fibrillar feltwork, for the most part coagulated 
cerebro-spinal fluid, that their walls are moderately expanded by it, and their 
cells are sufficiently granular to suggest a secretary function. 

10 is from the most anterior section of this series. The entire neural arch, 
median dorsal cartilaginous bar, and a portion of the notochord are included in 
this figure. The so-called first roof plate expansion covers a large area of the 
central portion of the spinal cord; its cavity is in communication with the cen 
tral canal. The ependymal walls of both the cavity and central canal are com- 
posed of several layers of cells. Motor roots, motor cells, substantia gelatinosa 
cells, and blood vessels are to be seen in transverse section. X 70. 


(Continued on page 50) 
48 


SPINAL CORD AND MEDULLA OF CYCLOSTOMES PLATE 3 
WILLIAM F. ALLEN 


THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL. 26, NO. 1 


(Continued from page 48) 


11 A transverse section taken 840 microns behind figure 10. It would pass 
through about the center of the first model (figs. 1 and2). Note that the roof ex- 
pansion, while containing an enormous cavity, is confined almost entirely to the 
right side. It is decidedly suggestive of a chorioid plexus forming cerebro- 
spinal fluid. A blood vessel lies in a fold in its wall and about this fold there is 
a great mass of coagulum. Throughout the first roof plate expansion, the cav- 
ity, and in many places the central canal itself, is larger than either the third 
or the fourth ventricle. > 70. 

12 From a section 645 microns behind figure 11, and not far from the caudal 
end of the first model (figs. 1 and 2). At this point the first roof plate expansion 
and its enclosed cavity attain their greatest width, which cavity is broadly con- 
nected with the central canal. X 70. 

13 Transverse section S80 microns behind figure 12 and at the very beginning of 
model 2 (figs.4and5). The dorsal wall of the first roof expansion is very wide and 
extremely vascular, and the convexity of its walls and the arch of the cavity bear 
evidence of moderate internal pressure from cerebro-spinal fluid. The wall, 
while still quite thick, contains a lesser number of ependymal cells, but more 
connective tissue. Note especially the absence of any direct connection between 
the cavity of the roof expansion and the central canal. A chain of ependymal 
cells still connects the two, and may be indicative of a former embryonic con- 
nection that has been lost. From this region, caudad, there is no communica- 
tion between the cavity of the roof expansion and the central canal. That a 
former connection occurred may be indicated by the fact that at various inter- 
vals ependymal cells are scattered between the two. A sensory root can be seen 
sending its fibers inward toward what is believed to be the substantia gelatinosa. 
xX 70. 

13 A, represents a small portion of the first roof plate expansion and its in- 
closed cavity from a section taken 410 microns behind figure 13. Observe es- 
pecially the rich blood supply for the dorsal wall and the ease by which diffusion 
could take place between a blood vessel and the cavity. A large roof plate cell 
is drawn separately, highly magnified, directly to the left of 13 A. Note the 
fine granules in cytoplasm, which gives evidence of being secretory. 125. 


ABBREVIATIONS 


B.V., blood vessel M.R., motor or ventral spinal nerve 
C.C., central canal root 
C.C.Ex., central canal extension into N.A., membranous neural arch 
roof plate expansion N.C., notochord 
Ep.N., layer of ependymal nuclei R.P.Ex., roof plate expansion 
M.C., motor or effective cells S.C.F., cerebro-spinal fluid 
M.D.C., median dorsal cartilaginous S.G., substantia gelatinosa 
bar S.R., sensory or dorsal spinal nerve 
M.F., Miillerian or giant fiber root 


50 


SPINAL CORD AND MEDULLA OF CYCLOSTOMES PLATE 4 
WILLIAM F. ALLEN 


At 


Wey 
we 


” 


LATE 
PLATE 


EXPLANATION OF FIGURES 


14 From a transverse section taken 19 microns behind figure 138A. It repre- 
sents a condition found throughout the entire posterior end of the first roof 
expansion, where a series of cavities are connected with one another. 
root is seen issuing from the cord. X 70. 

15 to 17 Three transverse sections, anterior, middle, and posterior through 
what has been designated as the second outcropping of the roof plate, which 
in model 2 (figs. 4 and 5) appears immediately behind the first roof expansion 
already figured. In figure 15, which is through the anterior end of this roof 
plate expansion, the outcropping is one of cells only. In figure 16 a distinet cay- 
ity containing coagulum is visible, connected below with the central canal. Ob- 
serve the vascularity of its walls, of the ependyma surrounding the central canal 
and the diverticula of the same, which doubtless serve as a modified chorioid 
plexus for producing cerebro-spinal fluid. In figure 17, which is through the 
posterior end of ths roof expansion, a cavity is still present. Its walls, which 
are mainly connective tissue, are rich in blood vessels. <A few scattered ependy- 
mal cells and fibers represent a possible connection with the central canal. A 
third outcropping of the roof plate has been described, but not figured in trans- 
verse section. It is located behind the second expansion, and the fact that no 
ependymal cells are encountered between the first and the second, and the second 
and third roof expansions, except immediately surrounding the central canal, 
favors the view that these separate outcroppings of the roof plate in the adult 
were never connected in an embryonic state. X 70. 

18 and 19 Two sections some distance caudad of figure 17 and some little 
distance apart. They show the spinal cord to be perfectly normal, but to be 
gradually tapering down in size. Figure 18 is from that part of the cord repre- 
sented by model 3. The ependymal area in the center has maintained its origi- 
nal size, the reduction that has occurred in size is to be found in the nervous 
portion. Not only does the ependymal area broadly divide the spinal cord into 
two halves, but it has here as elsewhere in a number of places, entirely obliter- 
ated the central canal for a space of a few microns. Figure 19 is some distance 
caudad of figure 18, passing through the sensory nerve roots one segment anterior 
to the abnormal sinus terminalis (model 4, fig. 9, S.7’.). The spinal cord will 
be seen to be fast losing its nervous structure, for no motor cells will be seen in 
this section or in any further caudad. X 70. 


A\ motor 


ABBREVIATIONS 


B.V., blood vessel 

C.C., central canal 

C.C.Ex., central canal extension into 
roof plate expansion or cast of the 
same 

Ep.C., ependymal cells 

Ep.N., layer of ependymal nuelei 


M.V.C., median ventral cartilaginous 
bar 

N.A., membranous neural arch 

Ne., notochord 

R.P.Ex., roof plate expansion 

R.P.Ex. (2), second roof plate expan- 
sion 


M.D.C., median dorsal cartilaginous S.G., substantia gelatinosa 

bar S.R., sensory or dorsal spinal nerve 
M.R., motor or ventral spinal nerve root 

root 


SPINAL CORD AND MEDULLA OF CYCLOSTOMES PLATE 5 
WILLIAM F, ALLEN 


~ STS 
“ST BAB, 
2 9 3) 


— 


PLATE 6 
EXPLANATION OF FIGURES 


20 Through the center of the abnormal sinus terminalis, showing the last 
pair of sensory roots (not last sensory root) in section. Observe especially the 
large abnormal sinus together with the rich vascular supply for the tissue immedi- 
ately surrounding it. Note also the dorsal and ventral swelling caused by this 
sinus and compare with photograph of the model (figs. 7 and 9). They show every 
evidence of being expanded by the cerebro-spinal fluid. From this point caudad 
no nervous structures have developed in the spinal cord. Also from this point 
caudad the notochord gradually decreases in caliber. > 70. 

21 to 23 represent three transverse sections taken through the extreme pos- 
terior, non-nervous, end of the spinal cord, which is composed solely of support- 
ing tissue and undifferentiated embryonic cells. Figure 21 is the most cephalic, 
and passes through the spinal cord a short distance behind the abnormal sinus 
terminalis (figs. 9 and 20, S.7'.). The spinal cord is still flattened here, but not 
indented ventrad, and contains a normal central canal. In figure 22 there has 
occurred a marked reduction in the size of the notochord. Observe that the 
spinal cord has not become flattened as it has more anteriorly, where the noto- 
chord is massive. With figures 21 and 22 compare figures 59 and 60, which are 
taken from a similar region of the spinal cord from a 70 mm. Polistotrema embryo. 
Figure 23 is the last of this series of drawings. It passes through the extreme 
posterior end of the spinal cord, some 45 microns caudad of the last trace of the 
notochord. Note the presence of a central canal, and that the cord has separated 
into several processes, which further on become lost in the surrounding con- 
nective tissue. X 70. 


PLATE 7 
EXPLANATION OF FIGURES 


24 to 31 From transverse frontal sections through the medulla and spinal 
cord of several human embryos, drawn with the aid of an Edinger-Leitz draw- 
ing apparatus, and reduced 3 diameters in reproduction. 

24 and 25 From transverse sections through the rhombic brain (frontal 
through the embryo) of a 23 mm. human embryo; figure 25 passes through the V 
root and posterior end of the cerebellar rudiments (lateral lobes), while figure 24 
is from a more anterior section. The roof expansion (chorioid plexus) is shown 
as a conspicuous black line in these figures. Everywhere within the boundaries 
of the roof expansion, the cavity is filled not only with coagulated cerebro-spinal 
fluid, but with embryonic red corpuscles. Whether these entered through veno- 
lymphatic openings (C) or are the result of extravasations was not determined. 
Wherever mesenchyme borders the roof expansion it is very vascular. It is 
apparent that the roof expansion is under moderate internal pressure. At first 
glance the roof expansion will show resemblance to the so-called first roof plate 
expansion of the spinal cord of the 20 cm. Polistotrema, already figured, but its 
later mode of development was shown to be very different. 10. 


(Continued on page 56) 


54 


SPINAL CORD AND MEDULLA OF CYCLOSTOMES 


PLATE 6 
WILLIAM F. ALLEN 


ABBREVIATIONS 


B.V., blood vessel 


Myo., myotomes 
C.C., central canal 


Ne., notochord 

S.C.F., cerebro-spinal fluid. 

Sp.M., spinal cord 

S.R., sensory or dorsal spinal nerve 
root 


C.N.A., cartilaginous neural arch 

C.T., white fibrous connective tissue 

Ep., ependyma 

M.D.C., median dorsal cartilaginous bar 


M.V.C., median ventral cartilaginous bar S.T., sinus terminalis 


55 


(Continued from page 54) 


26 and 27 Represent two transverse sections through the medulla, passing 
through the posterior end of the fourth ventricle of the same series from which 
figures 24 and 25 were drawn. Note especially the expansion of the roof plate 
and compare with the so-called second roof plate expansion of the 20 em. Polis- 
totrema spinal cord (fig. 16). It is questionable whether the openings (C) are 
artifacts or not. As was noted previously, the mesenchyme outside the roof 
plate is very vascular and the roof plate has the appearnce of being under a moder- 
ate degree of internal pressure. > 10. 

28) A rather oblique frontal section through the medulla of a 15 mm. human 
embryo (Inst. of Anat., trans. series, H 23). In this region the contour of the 
rhombic brain is such that the posterior part of the roof expansion of the fourth 
ventricle is cut transversely; while the more anterior portion, seen below, appears 
more or less in frontal section. More anterior sections would show the roof 
plate to be continuous. The posterior end of the fourth ventricle will admit of 
direet comparison with the second roof expansion of the 20 cm. Polistotrema 
spinal cord (fig. 16). Had the fourth ventricle been empty, as was the case of 
the rubber tubing in His’ experiments, there would be absolutely no grounds 
for believing that the anterior portion of the roof plate would be expanded as 
it is by the appearance of a pontine flexure. It might on the contrary have been 
folded up within the ventricle. > 16.6. 

29 A transverse section through the extreme posterior end of the fourth 
ventricle of the same seriesas figure 28. There ishereaslight roof plate expansion 
containing no cavity. Compare with figure 15. X 16.6. 

30 Transverse section through the thoracic spinal cord, taken from the same 
series as figure 28. Note that the roof plate consists of ependyma only, while 
the floor is reinforced by white matter, and even at this late stage if any marked 
increase In pressure occurred from the cerebro-spinal fluid of this region, an 
expansion of the roof plate would have been entirely possible. X 16.6. 

31 Similar to figure 28, but from an 8 mm. human embryo (Inst. of Anat., 
series H4). In this plane the posterior end of the fourth ventricle is cut nearly 
transversely, and is directly comparable with the second roof expansion of the 
20 cm. Polistotrema spinal cord (fig. 16). The cavity was full of coagulum, its 
walls have the appearance of being under moderate internal pressure, and the 
adjacent mesenchyme is very vascular. X 46.6. 


ABBREVIATIONS 


B.V., blood vessel Ol., inferior olive 
C., an apparent communication be- F.Hx., roof plate expansion 
tween the veno-lymphatics and the &.L., rhombic lip 
fourth ventricle R.P., roof plate of the central nervous 
C.C., central canal system 
Ch n., chondrocranium S.C., semicircular canals 
C.P., choroid plexus of the fourth S.R., sensory or dorsal spinal nerve 
ventricle root 
Crb.L., lateral lobes of the cerebellum T.S., tractus solitarius 
C.T., white fibrous connective tissue VJ///.G., auditory ganglion 
Ep., ependyma V.L.S., veno-lymphatie sinus 
Ex.Ar., external arcuate fibers V.R., trigeminal root 
I.L., inner or ependymal layer of nuclei W.R.F., white reticular formation 
Mar.L., marginal layer XI.R., accessory nerve root 


M.L., mantle layer 


SPINAL CORD AND MEDULLA OF CYCLOSTOMES 
WIL' IAM F. ALLEN 


5) 


7 


PLATE 7 


PLATE 8 
EXPLANATION OF FIGURES 


32 to 538 A series of transverse sections through the region of the V, VIII, 
and X ganglia in embryos of Petromyzon of ages varying from 10 to 26 days. It 
will be seen from these sections that Petromyzon develops an extensive roof 
expansion without the aid of a pontine flexure, and the cranial and spinal ganglia 
are well-formed while the central nervous system is a solid cord. All of the 
figures were drawn with the aid of an Edinger-Leitz drawing apparatus. With 
figure 54 a magnification of 76.6 diameters was used, while 250 diameters was used 
for the others. In reproduction they were all reduced one-half. 

32 Transverse section through the medulla in the region of the auditory 
vesicle from a 10 day Petromyzon. This is my oldest embryo in which the cen- 
tral nervous system has remained a solid cord. Ordinarily it becomes tubular 
during the seventh day. This section shows the medulla to consist of a syney- 
tium of protoplasm, consisting of a mass of round nuclei, much yolk, and a few 
fibers in the marginal layer. The nuclei have migrated a short distance to either 
side of the median dorso-ventral line. A seam (C.C.S.) has appeared here, which 
marks the position and beginning of the embryonic central canal. The proto- 
plasm bordering the central canal seam is finely granular and may be assum- 
ing a secretory function. The acustic ganglia and fibers are shown to be well- 
differentiated on both sides. 125. 

33 to 35 Three transverse sections passing through the medulla region of 
another 10 day Petromyzon embryo, in which the central canal has been somewhat 
retarded in development. These sections pass through the V, VIII, and X gan- 
glia respectively, and with the exception that the central canal furrow or seam 
(C.C.S.) has expanded into small dorsal and ventral cavities (no dorsal cavity 
has appeared in fig. 35) the general structure of the medulla is about the same 
as in figure 32. Later these cavities will become the dorsal and ventral ex- 
pansions of the embryonic central canal of the medulla. The protoplasm in the 
region of this seam is granular and may be secreting an embryonic cerebro- 
spinal fluid. From these figures it will be seen that the beginning of the central 
canal occurs at the same time throughout the entire rhombic brain. The anterior 
portion of the spinal cord, while not figured, contains a central canal furrow in 


the same stage. X 125. 
(Continued on page 60) 


ABBREVIATIONS 

Aud.V., auditory vesicle or otocyst Mes., mesencephalon 
Br.A., branchial arch M.L., mantle layer 
B.V., blood vessel Myo., myotomes 
C.C., central canal Ne., notochord 
C.C.C., central canal closure, caused  P.B., pineal body 

by fusion of lateral plates R.P., roof plate of the central ner- 
C.C.S., central canal seam or furrow, vous system 

in Petromyzon Sp.M., spinal cord 
Ep.N., layer of ependymal nuclei Sy.P., syneytium of protoplasm 
F.L.P., fused lateral plates of the  Tel., telencephalon 

spinal cord V.G., Gasserian or semilunar ganglion 
F.P., floor plate of the central nervous VJJI.VII., acustico-fascialis ganglion 

system; W.M., white matter 
G.C., germinal cell X.G., vagus ganglion (nodosum). 
Mar.L., marginal layer ’Y., yolk granules 


58 


PLATE 8 


SPINAL CORD AND MEDULLA OF CYCLOSTOMES 


WILLIAM F. ALLEN 


ecreky 


(Continued from page 58) 


36 to 38 From a Ll day Petromyzon embryo through the same regions as those 
shown in figures 33 to 35, and to facilitate comparison were placed directly un- 
der them. Considerable progress has occurred everywhere. Note 1) that 
the central canal seam with its small dorsal and ventral cavities in the 10 mm. 
series has given place to a typical embryonic central canal, which is much wider 
at the top and bottom than at the center. The constricted portion of course 
represents the last place for the protoplasm to give way or to be disintegrated. 
2) The floor plate is slightly thicker and less expanded than the roof plate, being 
reinforced on the outside by white matter and by a rapidly growing notochord 
below. 3) A few more nerve fibers and nuclei have appeared in the lateral 
plates. 4) The number of dividing germinal cells has increased while the num- 
ber of yolk granules remains about the same. 5) The first blood vessels have put 
in appearance directly outside the meningeal membrane (fig. 37, B.V. and 
nearer the roof plate on the opposite side). > 125. 

39 Transverse section through the spinal cord of the same 11 day Petromyzon 
series as figures 36 to 38, showing the so-called typical embryonic spinal 
(c@yeels — << Mey 

40 and 41 Somewhat oblique transverse sections from a 12day Petromyzon, 
passing in figure 40, through the V ganglionon one side and the VIII ganglion on 
the opposite side, and in figure 41 through the X ganglion on one side and a 
region behind the X ganglion on the opposite side. They can readily be compared 
with the 11 day series above (figs. 36-38). The lateral plates have apparently 
increased notably in the number of nerve fibers and nuclei, some of which, however, 
will have to be attributed to the fact that the sections are cut quite obliquely. 
Also the numbers of nerve fibers have increased in the floor plate. Throughout, 
the central canal has increased in width. Of especial interest is a small central 
mass of protoplasm (Sy.P.) in figure 40, which for a space of 50 microns per- 
sists as the last remnant of a once solid mass of protoplasm in the center of the 
medulla. It is obvious at this stage that some factor must have produced suf- 
ficient internal pressure to prevent the closing up of the ventricle on account 
of the rapid increase of cells and fibers in the lateral plates. It is fair to assume 
that this factor is internal pressure from cerebro-spinal fluid. X 125. 

54 of this plate will be described in its proper place, opposite the next plate. 


PLATE 9 
EXPLANATION OF FIGURES 


42 to 44. Three transverse sections through the medulla of an 18 day Petromy- 
zon embryo, passing through the V, VIII, and X ganglia, and for the sake of com- 
parison preserving the same order or arrangement as was used for the earlier 
embryos. A slight increase in the white matter is to be noted for the lateral and 
ventral plates over the 12 day series; but little, if any change has taken place in 
the central canal, unless possibly the central portion has increased slightly in 
width. Absolutely no further expansion of the roof plate has occurred. Since 
sections through the medulla of a 14 and 16 day series presented about the same 
appearance as figures 42 to 44 none were figured. X 125. 

45 Transverse section through the cephalic portion of the spinal cord from 
the same series as figures 42 to 44. Observe especially the beginning of the dorsal 
closure of the central canal (C.C.C.), showing the central canal to consist of dor- 
sal and ventral cavities and a central seam, strongly resembling the stage when 
it first appeared in the spinal cord. This dorsal closure of the embryonic central 
canal begins in the anterior portion of the spinal cord much earlier than the cor- 
responding ventral closure of the embryonic central canal in the medulla, occur- 
ing in my series of 14 and 15 days. It is obviously caused by the ingrowth of 
the lateral plates due to the great increase in the number of nerve fibers. & 125. 

46 to 48 Three transverse sections through the same region of the medulla of 
a 20 day Petromyzon as is shown above in figures 42 to 44 for the 18 day embryo. 
It will be seen that many noticeable changes have taken place, due primarily 
to a marked increase of nerve fibers in the lateral and ventral plates, and to a 
slight increase in the number of cells in the lateral plates. The shape of the 
medulla has become more compressed (flattened out in a dorso-ventral plane). 
Unquestionably the increase of fibers in the lateral plates has occurred largely 
in the median and ventral portions. Note the result on the embryonic central 
canal, which has been completely closed, except for a small dorsal triangular 
‘avity (C.C.) the early fourth ventricle. Its roof plate, however, has expanded 
somewhat. Observe that true ependymal cells are beginning ¢ > take on form 
about the ventricle, and those in the roof plate may soonassume a secretory func- 
tion, if they are not already active. Also the blood vessels have become more 
abundant outside of the medulla, especially in the region of the roof plate, which 
would make infiltration and diffusion through the roof plate into the ventricle 
sasy. X 125. ; 

49 From a transverse section through the anterior portion of the spinal cord. 
Observe the great increase in fibers in the lateral and ventral plates, and note 
that the cleft-like embryonic central canal has become entirely closed, but for 
a small portion, which will remain as the adult central canal. It is obvious that 
the two portions of the embryonic central canal which persist in the medulla 
and the spinal cord are the opposite, being the dorsal in the medulla and the ven- 
tral in the cord. X 125. 

50 to 53. ~Four transverse sections through the medulla of a 26 day Petromy- 
zon embryo, taken through the V, VIII, and X ganglia, and through a region 
behind the fourth ventricle where the central canal is passing ventral to assume 
its characteristic position in the spinal cord. A comparison with the 20 day 
series above (figs. 46 to 48) will demonstrate a marked increase in the size of the 
medulla and the fourth ventricle, and a greater expansion and convexity of the 


(Continued on page 62) 
61 


(Continued Jrom page 61) 


roof plate, in consequence of which the dorsal tips of the lateral plates are widely 
separated. There is a marked increase in the number of nerve fibers in the lateral 
plates, especially in the median and ventral portions. Of prime importance is 
the great expansion of the fourth ventricle and the roof plate, which apparently 
in Petromyzon can be explained only from internal factors, the most obvious 
of which is the mechanical expansion due to an increase in the cerebro-spinal 
fluid. It will be seen that these forces were sufficiently strong to more than 
offset the thickening of the lateral plates which would tend to obliterate the 
dorsal portion of the embryonic central canal as it has the ventral portion. It 
is apparent that this internal pressure has pushed the lateral wall apart in the 
dorsal region, where the lateral plates are thinnest and weakest. As was pointed 
out in the 20 day series the ependymal cells are becoming differentiated and 
probably have assumed a secretory function. Likewise the increase in the num- 
ber of blood vessels above the roof plate favors filtration and diffusion into the 
fourth ventricle. X 125. 

54 (See preceding plate.) Median longitudinal section through the head 
region of a 26 day Petromyzon embryo introduced for a comparison with the 
transverse sections in figures 50 to 53. Note especially that the marked convexity 
of the roof plate of the fourth ventricle is suggestive of expansion from an in- 
crease of cerebro-spinal fluid. Absolutely no pontine fiexure is to be seen, the 
little convexity that occurs in the floor plate can easily be attributed to an in- 
crease in the number of nerve fibers. Observe that the fourth ventricle (C.C.) 
is the remains of the dorsal portion of the original embryonic central canal, while 
the central canal of the spinal cord is the remains of the ventral portion. The 
ventral portion of the embryonic central canal of the medulla has been obliter- 
ated through the fusion of the ventral portions of the lateral plates. 38.3. 


ABBREVIATIONS 
Aud.V., auditory vesicle or otocyst Myo., myotomes 
B.V., blood vessel N.C., nerve cell 
C.C., central canal R.Ex., roof plate expansion 
C.C.C., central canal closure, caused R.P., roof plate of the central nervous 
by fusion of lateral plates system 
Ep., ependyma V.G., Gasserian or semilunar ganglion 
Ep.N., layer of ependymal nuclei VIIT.VII., acustico-fascialis ganglion 
G.C., germinal cell W.M., white matter 
G.M., white matter X.G., vagus ganglion (nodosum). 


M.L., mantle layer 
PLATE 10 


EXPLANATION OF FIGURES 


55to62 Taken from various transverse sections through embryonic and larval 
Polistotrema and Entosphenus (Pacific coast lamprey). Introduced to show the 
effect of the developing notochord on the spinal cord in Cyclostomes. They 
were drawn with the aid of an Edinger-Leitz drawing apparatus and reduced 
one-half in reproduction. 

55 Transverse section through the caudal region of a 20 mm. Polistotrema 
embryo. It will be seen at this stage that the notochord has produced very 
little visible effect on the spinal cord. Cyclostome embryos of this stage (com- 
pare fig. 39 for Petromyzon) presenta nearly cylindrical spinal cord; while that 
of all other vertebrates is more or less elliptical in cross section, the greater 


(Continued on page 64) 


62 


PLATE 9 


SPINAL CORD AND MEDULLA OF CYCLOSTOMES 


WILLIAM F. ALLEN 


(Continued from page 62) 


limmeter being dorso-ventral. It should be noted that the spinal cord is en- 
veloped tightly by a meningeal membrane, more or less fused with connective 
tissue outside that will form the neural arch, which is firmly attached to the 
notochord below. Immediately above, the mesenchyme is proliferating rapidly 
and migrating to the center where it will form the median dorsal cartilaginous 
bar. Little progress has occurred in the formation of the myotomes at the 
side, and elsewhere there is only loose mesenchyme. > 70. 

56 Similar transverse section to figure 55, but from a 27 mm. Polistotrema 
embryo. ‘This slightly later stage shows considerable growth of the notochord 
and a median indentation on the ventral surface of the spinal cord as the result. 
Note that the conditions surrounding the development of the notochord previous- 
ly enumerated under the description of figure 55 are instrumental in assisting the 
notochord in producing the gradual flattening (depression) of the spinal cord 
seen in the next figure. 70. 

57 Transverse section of the spinal cord of a 60 mm. Polistotrema embryo 
from the same region as figure 56. It will be seen that the spinal cord is en- 
closed in a membranous canal of dense connective tissue, attached below to the 
notochord and above to the median dorsal cartilaginous bar. Above this there 
are developing cartilaginous rays surrounded by dense connective tissue. The 
developing myotomes rest against the neural arches both laterally and dorsally. 
The notochord has increased greatly in size and, pushing up against the soft 
spinal cord, produces the depression and ventral indentation of the spinal cord 
exhibited in this figure. It should be noted that the roof plate is still ependyma 
and an expansion of the roof plate could take place even in this late stage if the 
mechanical factors enumerated for the medulla of Petromyzon were operative 
here. The thickening of the lateral plates has about obliterated the central 
portion of the embryonic central canal, leaving only the dorsal and ventral 
portions, in which there is a fibrillar feltwork, probably representing both cere- 
bro-spinal fluid and ependyma cilia. Reissner’s fiber is visible in the ventral 
or permanent central canal. X 70. 

58 ‘Transverse section through the tail region of a 20 mm. Entosphenus larva. 
It will be observed that the spinal cord is further developed than in the 27 mm. 
Polistotrema embryo (fig. 56). It 1s apparent that the same factors are involved 
in flattening the spinal cord as were enumerated for Polistotrema. The noto- 
chord has made fully as much growth and the structures surrounding the spinal 
cord are the same as in Polistotrema, with the exception that instead of a median 
dorsal cartilage for the attachment of the membranous neural arch there is a 
membranous neural spine. To some extent this may reduce the dorsal resistance, 
but on the other hand it may be compensated for by a greater development of 
the myotomes above the neural arch. X 125. 


ABBREVIATIONS 


C.A., caudal artery M.V.C., median ventral cartilaginous 
C.C., central canal bar 
C.H., caudal heart Myo., myotomes 
C.V., caudal vein N.A., membranous neural arch 
D.R., dorsal cartilaginous rays Ne., notochord 
Ep.N., layer of ependymal nuclei P.M., pia mater or meningeal mem- 
L.S., lateral veno-lymphatie sinus or brane of the younger stages 

anlage of the same R.P., root plate of the central nervous 
Mar.L., marginal layer system 


M.D.C., median dorsal cartilaginous Sp.G@., spinal ganglion 
bar V.7T., ventral veno-lymphatic trunk 
W.AL., white matter 
64 


PLATE 10 


SPINAL CORD AND MEDULLA OF CYCLOSTOMES 


WILLIAM F. ALLEN 


THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL. 26, No. 1 


PLATE il 
EXPLANATION OF FIGURES 


59 and 60 ‘Two transverse sections only 290 microns apart through the ex- 
treme posterior end of the spinal cord of a 70 mm. Polistotrema embryo. In 
the more anterior section (fig. 59) the spinal cord is depressed, contains develop- 
ing nervous elements, and the notochord is large proportionately. In the pos- 
terior section (fig. 60) the diameter of the notochord is much reduced and only 
supporting elements appear in the spinal cord. As a result no flattening of the 
spinal cord has taken place in this region. Compare with figures 21 and 22, which 
are similar sections through an adult Polistotrema. X 70. 

61 and 62 Taken from two transverse sections 480 microns apart, through 
the medulla oblongata of an adult Polistotrema. In both sections no nervous 
structures have appeared that were not present in the spinal cord. Note as you 
pass caudad (figs. 62 to 61) that the medulla becomes flattened ventrally and dor- 
sally in direct proportion to the increase in size of the notochord. This rela- 
tionship can be shown fully as marked in more anterior sections, and in sections 
taken from a similar region of larval Petromyzon. X 25. 


ABBREVIATIONS 


Aud.V., auditory vesicle or otocyst 

B.V., blood vessel 

C.C., central canal 

E.N., undifferentiated embryonic nu- 
clei 

Ep., ependyma 

Ep.N., layer of ependymal nuclei 

L.S., lateral veno-lymphatic sinus or 
anlage of the same 

M.D.C., median dorsal cartilaginous 
bar 

M.F., Miillerian or giant fiber 


M.V.C., median ventral cartilaginous 
bar 

Myo., myotomes 

N.A., membranous neural arch 

N.C., nerve cell 

P.M., pia mater or meningeal mem- 
brane of the younger stages 

P.P., parachordal plate 

Sp.G., spinal ganglion 

W.M., white matter 

X.G. in figure 62 should be Sp.G. 

X.N., vagus nerve 


M.L., mantle layer 
PLATE 12 


EXPLANATION OF FIGURES 


63, represents a diagrammatic reconstruction of the fourth ventricle from an 
adult Polistotrema series, and the planes from which the transverse sections were 
drawn for figures 64 to 66 are indicated by lines bearing those figures. Observe 
how the large fourth ventricle of the embryo has been reduced to a small central 
canal, having a posterior dilation (PyV), and how the anterior end breaks up 
into two or more small longitudinal canals that soon terminate in the sinus 
mesoccelicus. 

64 and 65, represent transverse sections, taken at different levels of the fourth 
ventricle of Polistotrema. These sections were drawn with the aid of an Edin- 
ger-Leitz drawing apparatus and were reduced one half. 

64 From a transverse section through the caudal portion of the mesencephalic 
lobes (cerebellum of Miss Worthington). Exact plane indicated by line 64 in 
figure 63. The section passes through the posterior mesoccele or cerebellar ven- 
tricle (7’’.) and the sinus mesoccelicus (anterior dilation of the fourth ventricle 
of Miss Worthington), a short distance behind the branching off of the posterior 
mesoccele. The cavity contains a fibrillar feltwork, which is in part coagulated 
cerebro-spinal fluid and in part ependymal cilia. The ependyma surrounding 
the fourth ventricle is rich in blood vessels, which derives its arterial supply from 


(Continued on page 68) 
66 


PLATE 11 


SPINAL CORD AND MEDULLA OF CYCLOSTOMES 


WILLIAM F. ALLEN 


(Continued from page 66) 


two medulla arteries (/.A.). Here as elsewhere, the ependyma surrounding 
the fourth ventricle doubtless functions as a modified chorioid plexus, discharg- 
ing cerebro-spinal fluid into the fourth ventricle. It will be seen that the cavity 
of the fourth ventricle is smaller than the peculiarly modified central canal and 
roof expansion cavity of the Polistotrema spinal cord portrayed in figures 10 to 
13. X 25. 

65 A more caudal section through the extreme tip of the posterior lobes of 
the mesencephalon (cerebellum of Miss Worthington), its exact plane being 
indicated by line 65 in figure 63. It will be seen that the fourth ventricle of the 
embryo has in this region of the adult become reduced to three small longitudinal 
canals (A,V.), which are imbedded in a rather large, dense, and vascular epen- 
dymal mass. The most dorsal of these canals contains Reissner’s fiber. Here 
again the ependymal walls are probably functional as a modified chorioiod plexus. 
Very shortly these canals reunite and continue some little distance caudad as 
a small central canal, no larger than the central canal of the spinal cord. X 25. 

66 Transverse section through the posterior end of the medulla of the same 
series as figure 64. The exact plane of the section is indicated by line 66 in figure 
63. It passes through what has been designated as the posterior dilation of the 
fourth ventricle (P;V), which is nothing more than a fair-sized centrally lo- 
cated cavity, the remains of a much larger embryonic fourth ventricle, surrounded 
by a great mass of vascular ependyma. The center of this cavity contains a 
fibrillar feltwork (S.C.F.) composed largely of coagulated cerebro-spinal fluid 
and some ependymal cilia. Here as more anteriorly we probably have a modi- 
fied chorioid plexus, the ependymal walls and their blood vessels secreting and 
filtering cerebro-spinal fluid into the fourth ventricle. > 25. 


ABBREVIATIONS 


A4V., anterior fourth ventricle M.A., medulla artery 
B.V., blood vessel Mes’., posterior lobes of the mesen- 
Ep., ependyma cephalon, cerebellum of Miss Worth- 
Hab. B., habenular body ington 
Inf., infundibulum P,V., posterior fourth ventricle 
M., mesoccele or mesencephalic ventri-  S.C.F., cerebro-spinal fluid 

cle S.M., sinus mesocelicus 


M’., anterior portion of the mesocele V.M.R., motor V root 
or sub-commissural canal of Nicholls  4V., fourth ventricle 

M"’., dorsal portion of the mesoccele VII/I.G., auditory ganglion 
or optocoel and posterior portion of | X.R., vagus root 
the optocoel of Nicholls 


PLATE 13 
EXPLANATION OF FIGURES 


67 to 69 Three transverse sections through the brain region of a 37 mm. 
Amphioxus. No blood vessels were seen in any of these sections, but the mem- 
branous neural canal is surrounded on three sides by enormous veno-lymphatic 
sinuses, and the structure of the central nervous system is to a considerable extent 
made up of rather coarse supporting tissue, making infiltration an easy method 
for nourishing the brain. Drawn with an Edinger-Leitz drawing apparatus and 
reduced one-half in reproduction. 


(Continued on page 70) 
68 


PLATE 12 


SPINAL CORD AND MEDULLA OF CYCLOSTOMES 
WILLIAM F. ALLEN 


(Continued from page 68) 


The most anterior section, figure 67, passes through the anterior ventricle at 
its highest point, which is a short distance behind the neuropore. This ventri- 
cle has no dorsal dilation suggestive of the fourth ventricle. What dilation occurs, 
is median and ventral. Cilia-like processes from the border of the cells enter 
the cavity. If the ependymal cells are not secretory it is possible that the cere- 
bro-spinal fluid of Amphioxus does not differ from the serum of the adjacent 
veno-lymphatie sinuses. If nerve cells occur in this region they are small, and 
in ordinary preparations indistinguishable from ependymal cells. > 125. 

68 60 microns behind figure 67. The large central canal of the embryonic 
brain has evidently become reduced in this region to a ventral central canal 
(C.C.) and a small dorsal isolated cavity (V 2.). This isolated dorsal cavity can 
not be compared with the fourth ventricle of higher vertebrates. It is rather 
to be looked upon as a vestigeal structure, which may aid in the infiltration of 
lymph from the outer veno-lymphatic sinuses. 125. 

69 Taken from a section 530 microns behind figure 68. It passes through 
that part of the brain in which there are accumulated a great number of giant 
cells (J/’.C’.) in the region of the roof plate. As in figure 68, there is an isolated 
cavity near the dorsal surface, which was probably a portion of the large embry- 
onic central canal, but which in the adult is separated from the central canal and 
from the more anterior isolated dorsal cavity by ependyma. It seems best to 
the writer to regard this and the preceding dorsal cavity as vestigeal structures. 
x 125. 

70 (See next plate.) Transverse section through the anterior spinal cord 
from the same series as the three previous figures. Observe that the Amphioxus 
spinal cord is not depressed as is the Cyclostome spinal cord, but is indented ven- 
trally by the notochord. The central canal, which in some places exists as a 
dorso-ventral cleft, is almost obliterated here by the ingrowth of ependymal 
tissue. X 125. 

71 Cephalic transverse section through a portion of the spinal cord, menin- 
geal membranes, neural arch, notochord, spinal ganglion,and sensory root of 
an adult Polistotrema. Observe the depression of the spinal cord, its ventral 
indentation, the ventral or permanent central canal (C.C.), which contains Reiss- 
ner’s fiber, and immediately above, the dorsal portion of the embryonic central 
canal, which is here more or less filled with ependymal cells and their processes. 
It will be seen that the gray matter is as much flattened out as is the cord it- 
self, and the ventral horn and motor cells are crowded laterad, while the dorsal 
horn, substantia gelatinosa (S.G.), is apparently median and dorsal. Within 
the neural arch there is abundant room for a spherical spinal cord. The cord is 
held in place by the usual meningeal membranes. X 70. 


ABBREVIATIONS 

Ar., Arachnoidea Ne., notochord 
C.C., central canal P.M., pia mater or meningeal mem- 
C.T., white fibrous connective tissue brane of the younger stages 
D.M., dura mater S.G., substantia gelatinosa 
D.S., dorsal veno-lymphatic sinus Sp.G., spinal ganglion 
Ep., ependyma S.R., sensory or dorsal spinal nerve 
L.S., lateral veno-lymphatie sinus or root 

anlage of the same Suba.S., subarachnoid cavities 
M’.C’., Miillerian or giant cells Subd.S., subdural spaces 
M.F., Miillerian or giant fiber W.M., white matter 
Myo., myotomes V.1., anterior ventricle Amphioxus 
N.A., membranous neural arch V.2., vestiges of the embryonic central 
N.C., nerve cell canal in Amphioxus 


70 


PLATE 13 


SPINAL CORD AND MEDULLA OF CYCL ISTOMES 
WILLIAM F. ALLEN 


PLATE 14 
EXPLANATION OF FIGURES 


72 to 80 represent a number of transverse sections through the medulla of shark, 
amphibian, and pig embryos for the purpose of demonstrating various stages of 
roof plate expansion. 

72 Rather oblique transverse section through the medulla of a 19 mm. Squa- 
lus embryo in the region of the VIII ganglion (from series No. 2 of Professor 
Scammon’s collection). Note the well-formed fourth ventricle, and the broadly 
expanded and very much stretched roof plate. Its collapsed appearance in this 
section is doubtless due to fixation or preparation. On account of a great pro- 
liferation of cells and nerve fibers the lateral plates have fused ventrally, as 
in Petromyzon, obliterating that part of the embryonic central canal. Attention 
should be called to the fact that the medulla roof plate in sharks begins to expand 
much earlier than it does in Petromyzon. This figure shows a well-expanded 
roof plate, while the cells in the mantle layer are no more differentiated and there 
are no more nerve fibers in the marginal layer than appear in a 12 day Petromy- 
zon-embryo (fig. 40), where there is no fourth ventricle and no expansion of the 
roof plate.  X 70. 

73 Transverse section through the medulla of a 15 mm. Necturus taken 
through the VIII ganglion. Observe the wide fourth ventricle and the broadly 
expanded and greatly stretched roof plate, and the coagulated appearance of 
the cerebro-spinal fluid in the ventricle. Like Squalus, the fourth ventricle 
begins relatively much earlier than in Petromyzon. It should be noted that no 
blood vessels have reached the level of the roof plate or entered the medulla; 
hence the coagulum in the ventricle must be largely a product of secretion. X 39. 

74 From a transverse section of a very young Amblystoma embryo in the 
region of the auditory vesicle (Professor Johnston’s series No. 50). Here there 
has occurred a dorsal and a smaller ventral excavation of the cleft-like central 
canal. It will be seen that the larger dorsal cavity, the beginning of the fourth 
ventricle, possesses no thinner roof plate than does the spinal cord (fig. 83). 
Also in this section (fig. 74) the roof and floor plates are about equally thick. 
What has taken place dorsally and ventrally throughout the embryonic central 
canal has been a migration of the cells outward. The fact that in this section 
of the medulla the roof plate is no thinner than in the section of the spinal cord 
(fig. 83), taking note that the spinal and cranial ganglia are well-formed, is evi- 
dence against the hypothesis, that the greater migration of the neural crest 
cells of the medulla was the prime cause of the thinning out of the roof plate of 
the rhombic brain. X 70. 


(Continued on page 74) 


ABBREVIATIONS 
Aud.V., auditory vesicle or otocyst M.L., mantle layer 
B.V., blood vessel Myo., myotomes 


C.C., central canal or cast of the same N.A., membranous neural arch 
C.C.C., central canal closure, caused JN.C., nerve cell 


by fusion of lateral plates Ne., notochord 
Ep., ependyma P.C., pigmented or eye cells of Amphi- 
Ep.N., layer of ependymal nuclei oxus 
F.P., floor plate of the central nervous R.F2z., roof plate expansion 

system R.P., roof plate of the central nervous 
G.C., germinal cell system 
L.S., lateral veno-lymphatic sinus or  8S.C.F., cerebro-spinal fluid 

anlage of the same V.G., Gasserian or semilunar ganglion 
Mar.L., marginal layer VITI.G., Auditory ganglion . 
M’.C’., Miillerian or giant cells W.M., white matter 
M.F., Miillerian or giant fiber 4 V., fourth ventricle 


72 


SPINAL CORD AND MEDULLA OF CYCLOSTOMES PLATE 14 
WILLIAM F. ALLEN 


(Continued from page 72) 


75 to SO represent five transverse sections through the developing fourth 
ventricle and roof plate expansion in pig embryos from 5 mm. up to 14 mm. 
With the exception of figure 75, which is from my collection, the remaining figures 
are from frontal series belonging to the Institute of Anatomy. That there is 
a direct relationship between the amount of visible coagulum in the form of a 
fibrillar feltwork and the expansion of the roof plate is evidenced by the fact 
that this coagulum does not appear in the early embryos before the roof plate 
has assumed the appearance of an organ capable of the production of cerebro- 
spinal fluid (as indicated by vascular supply and granular appearance of the 
cells). It may be inferred that the earliest non-coagulable cerebro-spinal found 
in the earliest stages is an embryonic fluid which differs in no way from the ordi- 
nary intercellular juices, but that the appearance of coagulum at the time when 
the roof plate has attained the appearance of a functional chorioid plexus is 
indicative of a chemical change in the fluid, which if a product of secretion is 
capable of producing a marked increase of internal pressure in the cerebro-spinal 
fluid and consequent expansion of the roof plate. 

75 From a transverse section through the medulla of a 5 mm. (or less) pig 
embryo, in the region of the auditory vesicle. It will be seen that the peripheral 
branches of the intersegmental blood vessels have about reached the roof plate, 
but no blood vessels have entered the medulla. The protoplasm of the inner 
margin of the ependymal cells is sufficiently granular to suggest a secretory 
function. The small amount of coagulum in ventricle is probably the result 
of secretion, but the cerebro-spinal fluid has probably not exerted much inter- 
nal pressure. X 70. 

76 Transverse section of a 6 mm. pig medulla through the widest portion 
of the fourth ventricle, namely, at the level of the V ganglion. This is the only 
portion of the roof plate to have undergone any stretching of its cells, and this 
is confined solely to the most centrally located cells. This section shows a con- 
siderable increase in the size of the fourth ventricle and expansion of the roof 
plate, together with some increase in the amount of coagulable cerebro-spinal 
fluid (S.C.F.) and an increase in the number of blood vessels above the roof plate; 
but no blood vessels have entered the substance of the medulla. At this stage 
the pontine flexure could not have been a factor in producing the roof expansion. 
The collapsed appearance of the roof plate at its center is not natural, but rather 
a result of the preparation of the material. 39. 


PLATE 15 


EXPLANATION OF FIGURES 


77 ‘Taken from a transverse section through the V ganglion of a 7 mm. pig 
embryo. Note the increase in the number of blood vessels above the roof plate, 
which together with the increase in the coagulable cerebro-spinal fluid suggests 
a functional chorioid plexus. The pontine flexure in this stage is too slight to 
have any effect on the expansion of the roof plate. It should also be recorded 
that a few blood vessels have entered the outer surface of the medulla. Asin 
the previous series the sections have suffered a collapse of the roof plate from 
fixation or later preparation of the material. X 39. 

78 From a transverse section of a 10 mm. pig embryo through the region of 
the V ganglion. The increased vascularity of the mesenchyme above the roof 
plate together with the enormous amount of coagulated cerebro-spinal fluid 
GS.C.F.) in the ventricle are evidences of the factors which have produced the 
increased expansion noticed in the roof of the ventricle. Also at this stage the 
pontine flexure has increased to such an extent that its action on a fourth ventri- 
cle full of cerebro-spinal fluid, itself under a moderate pressure, would produce 
a further expansion of the roof plate. As in the preceding sections the roof plate 
has suffered a collapse in the preparation of the material. X 39. 


74 


SPINAL CORD AND MEDULLA OF CYCLOSTOMES PLATE 15 
WILLIAM F. ALLEN 


3 — Fr = eens REx. 


RX BY. 
~ V. 


Ms pty. 
us Ri iA 
ie yr thee, 


79 and 80 Two transverse sections through the anterior and posterior ends 
of the fourth ventricle from a 14 mm. frontal series of a pig. A more advanced 
stage in the development of the chorioid plexus together with « more pronounced 
pontine flexure has produced a much larger fourth ventricle and expanded roof 
plate than is shown in the previous series (fig. 78). The thickening of the lateral 
walls of the medulla is taking place as it did in Petromyzon and Squalus, but the 
greater expansion of the ventricle in the pig (fig. 79) has prevented the walls 
from fusing ventrally. Nevertheless, the thickening of the ventral portion of 
the lateral plates would increase the pressure of the cerebro-spinal fluid. 25 


and 39. 
ABBREVIATIONS 
B.V., blood vessel R.Ex., roof plate expansion 
C.S.F., cerebro-spinal fluid S.C.F., cerebro-spinal fluid 
Ep.N., layer of ependymal nuclei V.G., Gasserian or semilunar ganglion 
Mar.L., marginal layer 4 V., fourth ventricle 
M.L., mantle layer X.R., vagus root 


R.C., embryonic red corpuscle 
75 


PLATE 16 
EXPLANATION OF FIGURES 


S81 to 87 Represent true transverse sections through what has been termed 
in the text, the typical embryonic spinal cord, from a number of different verte- 
brates, all of which have developed a tubular nervous system after the neural 
fold method. They were drawn with the aid of an Edinger-Leitz drawing appara- 
tus and reduced one half in reproduction. 

81 From a transverse section through the anterior portion of the spinal cord 
of a 10 mm. Squalus embryo (Professor Scammon’s series No. 16). This so- 
called typical embryonic spinal cord is decidedly compressed. An earlier stage 
possessed an elliptical cord with its greatest diameter from side to side. The 
floor plate is slightly thicker than the roof plate. Each contains a single layer 
of nuclei. The ventral portion of the cleft-like central canal is expanded into 
a cavity, which persists as the permanent central canal. <A well-formed spinal 
ganglion is seen to the left. > 125. 

82 Transverse section of the spinal cord of a 19 mm. Squalus embryo (from 
Professor Scammon’s series No. 2). Note that the dorsal closure of the lateral 
plates, due to fiber and cell proliferation, is the same as was figured for Petromy- 
zon. They meet in a seam, leaving dorsal and ventral cavities, of which only 
the ventral one persists. As in Cyclostomes this method of closure would tend 
to throw a large part of the embryonic cerebro-spinal fluid into the brain cavities. 
x 70. 

83 and 84 ‘Transverse section through the anterior portion of the spinal 
cord of an Amblystoma and a turtle embryo. The former (taken from Professor 
Johnston’s series No. 50) is a rather early representative of the so-called typi- 
cal embryonic stage; while the latter is a rather late representative of this stage. 
Both cords may be said to be compressed (elliptical, having its greatest diameter 
dorso-ventral), but only slightly so, when compared with birds and mammals 
(figs. 85 and 86). As a result, granting an equal proliferation of fibers and cells 
in the lateral plates, it would be expected that the adult cord in Amblystoma and 
the turtle would be more depressed, which is found to be the case. X 70 and125. 

85 and 86 From anterior transverse sections of the spinal cord of a 93 hour 
chick anda 5mm. pig. Both are good illustrations of the so-called typical em- 
bryonic stage, the pig being in a slightly more embryonic state. In these we have 
the most compressed of all embryonic cords examined, while the adults cord are 
nearly cylindrical. XX 125. 

87 Transverse section through the caudal end of the same spinal cord shown 
in figure 86. Observe spherical appearance which is indicative of an earlier phase 
in its development. X 125. 


76 


SPINAL CORD AND MEDULLA’ OF CYCLOSTOMES 
WILLIAM F. ALLEN 


ABBREVIATIONS 


C.C., central canal 

C.C.C., central canal closure, caused 
by fusion of lateral plates 

Ep.N., layer of ependymal nuclei 

F.P., floor plate of the central nervous 
system 

G.C., germinal cell 

Mar.L., marginal layer 


M.L., mantle layer 

M.R., motor or ventral spinal nerve 
root 

Ne., notochord 

R.P., roof plate of the central nervous 
system 

Sp.G., spinal ganglion 

Sy.P., syneytium of protoplasm 


MORPHOLOGY OF THE ROOF PLATE OF THE FORE- 
BRAIN AND THE LATERAL CHOROID PLEXUSES 
IN THE HUMAN EMBRYO 


PERCIVAL BAILEY 


From the Anatomical Laboratory of the University of Chicago 


THIRTY-ONE FIGURES 


CONTENTS 


IMAGCOCWUCELOMS «ss bien iea's wee TG ee ieee 15, MASE ey 
PAM E eectai aiken cine s Sid'y hd ues ao White ois wa tie WG bien ai Ee Oe ee Aaa 80 
Material and methods..... Ar tre OPE Neer etl toe th ea Seto 84 
PPDRGUIDWONS ¢ «<0 hk ss00is ies oh ao ie or er Park Soo a ew oe 
1. The 19 mm. embryo...... fry ree areeise rte ears te. Reta z 86 
Pe DEO INN OM DLT Osa < scum eho sxe was ale lee conte Meee eae anal 9] 
3. The 32 mm. embryo...... .y e eS ; ee Ror Mie Come Ce 96 
MIB EUB SION so w.¢ Medasos abe aves SPA? SCR Be err es Be 99 
1. Telencephalon...... a t Wyswaie ek, ac nlig atk are pat erp RU Ses Pere aks 99 
2. Diencephalon........... de ele tcsie le iS: ae. < Wis 9 ok OR RRO LS AG ton in et cal ea 108 
YANO LINO; te isin re ne eeia ee tre, <i SR cies a a | CIR he ws Palen atets 110 


INTRODUCTION 


The researches of Minot, von Kupffer, Burckhardt, G. Elliot 
Smith, C. Judson Herrick, J. B. Johnston and others, have 
sueceeded in homologizing with considerable certainty the 
structures in the roof of the prosencephalon of the lower verte- 
brates, and of some of the lower Mammalia. The results of 
these comparative studies Johnston extended to the human 
embryo and it seemed desirable to examine some other human 
embryos of different ages in the hope that additional light might 
be shed on the method of development of these structures. 

79 


SO PERCIVAL BAILEY 


HISTORY 


Concerning the structures with which this study is concerned, 
the writings of such authors as Faivre (’54), Luschka (’85), 
and Haeckel (’60), previous to the work of Wilhelm His, contain 
very little of value. 

Of the choroid plexus of the lateral ventricle, His writes (’04): 


Sein dem Thalamus angehefteter Randstreifen bleibt ependymal und 
in ihm bildet sich die Fissura chorioidea, von der aus die Epithelfaltun- 
gen des Corpus chorioideum in den Seitenventrikel sich einstiilpen. 


Minot (01) considers the lateral plexus to be developed from 
the velum transversum. 


In both birds and mammals the lateral portions of the velum, 1.e., 
the choroid plexus of the lateral ventricle is highly developed. It 
thus appears that as we ascend the vertebrate series there is first a 
broadening of the velum, and an increase in its lateral development, 
then occurs a further reduction and flattening out of the velum, and a 
much greater growth of the lateral plexus. 


G. Elhot Smith (03) attempted another explanation for the 
formation of the lateral plexus. 


Now, although in the whole of its extent the epithelial layer of the 
choroid plexus presents uniform features, it is difficult to admit a com- 
mon origin for the whole structure; with regard to that part of the 
plexus which is found in the region of the foramen of Monro, there 
can be little doubt of its origin from the primitive roof of the fore- 
brain. . . . . But the case is very different with that portion of the 
plexus which is not directly connected with the roof of theforebrain, but 
is attached to the stria terminalis. There is no evidence to show that 
this portion is derived from the roof, and all the facts of development 
point to the conclusion that its proximal attachment to the optic thala- 
mus 1s a primitive and not a secondarily acquired relation. Such being 
the case, the caudal extension of the epithelial choroidal fold in the 
mammalian hemisphere would appear to be derived from a stretching 
of the attachment of the labium caudale of the cerebral hemisphere to 
the optic thalamus. As a result of this, the connecting band becomes 
reduced to an epithelial lamina, which becomes invaginated and folded 
by an extension backward of the choroidal folding which begins farther 
forward in the region of the foramen of Monro. 


Johnston (09) after showing that the velum transversum is 
continued down the side-wall of the prosencephalon as the di- 
telencephalic groove, returns to a modification of Minot’s 
original idea. 


~ 


DEVELOPMENT OF THE CHOROID PLEXUS Sl 


In the angle between the [cerebral] vesicle and the diencephalon 
appears the choroid plexus pushing into the lateral ventricle. It 
appears as a folding of the anterior limb or wall of the velum trans- 
versum and its lateral prolongation [di-telencephalic groove]. 


Concerning the mesodermal portion of the lateral plexus, Meek 
(07) makes the following statement. 


The choroid plexuses of the lateral ventricles are due to an ingrowth 
of the pia mater pushing the mesial wal of the hemispheres into the 
ventricles. 


This is the current notion. 


The arachnoid is not supposed to be present, although the plexus 
is but a fringe of the velum interpositum, into the structure of which 
the arachnoid does enter. The neural wall is, of course, preserved, 
but consists only of a simple epithelium. The plexuses are then thin 
laminae covered with an epithelium, beneath which is a connective 
tissue stroma containing an extraordinarily rich network of blood- 
vessels. 


Findlay’s (’99) idea of a single membrane, the pia-arachnoid, 
removes the difficulty concerning the involvement of the 
arachnoid. 

Finally, Hochstetter (’13) has. written a purely descriptive 
account of the development of the lateral choroid plexus with no 
attempt to analyze its parts. He summarizes his work as 
follows: 


Fassen wir das bisher Mitgeteilte zusammen, so kénnen wir sagen, 
dass sich die Plexus chorioidei der Seitenventrikel ungefiihr in dersel- 
eben Richtung entwickeln wie die Hemisphiirenblasen selbst. Zuerst 
angelegt, wenn auch nicht gleich als Anlage der Plexus chorioidei 
kenntlich, ist ihr vorderster, im Bereiche der Decke des Cavum Monroi 
befindlicher Abschnitt. Er ersteht hier, wie wir gesehen haben, aus 
den die Sulei hemisphaerici [di-telencephalie grooves] bildenden Hirn- 
wandfalten, sowie aus dem diese beiden, in der Fortsetzung des 
Zwischenhirndaches verbindenden, vorerst kielférmig vorspringenden 
Wandtcile des Endhirns. Ein zweiter Abschnitt erscheint wesentlich 
spiiter in Form einer jederseits zuniichst einfachen gegen den Hohlraum 
der Seitenkammer zu vorspringenden Falte, der als Area chorioidea 
bezeichneten Wandplatte der Hemisphire. Diese Falte geht vorn in 
die Wandfalte des Sulcus hemisphaericus tiber, wiihrend sie sich nach 
riickwiirts etwas von ihr entfernt (fig. 4), noch weiter nach riickwarts 
aber bald verstreicht. So erscheint bei dem iltesten von den drei 


THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL. 26, NO. 1 


82 PERCIVAL BAILEY 


bisher besprochenen Embryonen (H. Sch. 2) der hinterste dem Zwischen- 
hirn anliegende Abschnitt der Area chorioidea nach vollkommen glatt 
und ungefaltet. 


All of his figures, with the exception of figure 6, appear to be 
through the diencephalon, back of the velum transversum. 

The development of the paraphysis in the human embryo has 
never been followed. In fact, its identification is very much in 
doubt. Francotte (94) claims to have found it in an embryo of 
twelve weeks. It is said to be an organ characteristic of all ver- 
tebrates, but becomes very rudimentary in birds and mammals. 
Its development has been followed by Dexter (’02) in the com- 
mon fowl, and it has been described by Selenka (’91) in the 
opossum. A good review of the literature is given by Warren 
(705). | 

Of the human embryo, Streeter in Keibel and Mall’s Human 
Embryology says, 

Orally this choroid roof [of the third ventricle] is continued into the 
telencephalon where it forms a pointed pouch overlapping the lamina 
terminalis and the contained commissures. . . . The anterior cho- 


roidal pouch has been homologized with the paraphysis of the lower 
vertebrates. 


It should be borne in mind. that in all vertebrates, the para- 
physis, if present, arises from the roof of the telencephalon just 
cephalad to the velum transversum. In view of this fact, the 
structure labelled paraphysis by Goldstein (’03) is obviously 
not so, since it lies behind a structure called the velum trans- 
versum and at the posterior end of the diencephalic roof. 

The epiphysis is a constant organ in the vertebrate series” 
(except in the alligator) but probably concerning no other organ 
has there been so much confusion and misinterpretation. For a 
review of the literature, reference may be had to Gaupp (’98). 
The development of the epiphysis has not been followed com- 
pletely in the human embryo, and so far as I know, an indication 
of the division into the two parts, epiphyseal stalk and pineal 
vesicle, which seems to be so characteristic of many vertebrates, 
has never been recorded. 

With the recognition of the velum transversum in Cyclostomes 
by Sterzi (07), the velum has been established as a constant 


DEVELOPMENT OF THE CHOROID PLEXUS 83 


morphological feature in the roof of the prosencephalon of 
vertebrates. 

The tela chorioidea diencephali has been the subject of an 
extensive anatomical, embryological and comparative study by 
Lachi (’88). It presents few features of interest. A number 
of names have been applied to parts of it (Zirbelpolster, dorsal- 
sac, post-paraphysis, post-velar arch, etc.), but it has generally 
been recognized as extending from the superior commissure to 
the velum transversum. 

With regard to the roof of the telencephalon, however, there 
is no such unanimity of opinion. I shall not attempt to review 
the observations on the lower’ vertebrates. Johnston ('13) 
reviews the literature fully and on the basis of this and his own 
extensive observations presents the following scheme as covering 
all the forms of the roof of the telencephalon, beginning with 
the preoptic recess in which the sulcus limitans ends: 

Lamina terminalis (containing the anterior commissure} 
Recessus neuroporicus 

Lamina supraneuroporica (containing the pallial commissures) 
Recessus superior 

Tela chorioidea telencephali 


, Paraphysis 
Velum transversum (anterior leaf) 


Instead of tela chorioidea telencephali, the more definite term, 
tela chorioidea telencephali medii, will be used in this paper. 

This plan may be completed for the diencephalon as follows: 

Velum transversum (posterior leaf) 
Tela chorioidea diencephali 
Commissura superior 

Epiphysis 

The posterior commissure belongs to the mesencephalon. 
The neuropore itself has never been followed through the suc- 
cessive stages of its development in the human embryo. 

The evidence for transferring the above scheme to the human 
embryo is not wholly conclusive, and the present work was under- 
taken primarily to determine in how far the morphology in this 
region in certain human embryos was compatable with the 
above scheme. 


S4 PERCIVAL BAILEY 


I wish here to express my indebtedness to Dr. C. J. Herrick, 
whose broad knowledge and mature judgment has been of in- 
valuable assistance, and especially to Dr. Geo. W. Bartlemez, 
at whose suggestion the work was undertaken and without whose 
kindly interest the work would have been impossible. Thanks 
are also due to Mr. A. B. Streedain for the care he has taken 
with the illustrative work. 


MATERIAL AND METHODS 


The material upon which this study was chiefly based con- 
sists of three very well preserved human embryos, cut in trans- 
verse series, stained in bulk with borax carmine, and counter- 
stained on the slide with orange G. Wax-plate reconstructions 
were made, the plates being stacked from aside view of the embryo 
drawn from a photograph, taken after fixation. The shrinkage 
after embedding was calculated and the outline reduced accord- 
ingly. Although the primary object of this study is the mor- 
phology and relations of the roof plate, in two cases the entire 
forebrain has been modeled. This was done because the embryos 
happened to fall in at opportune intervals between His’ embryo 
CR (13.6 mm.) and the embryo of 50 mm. also modeled by him, 
and also in order that the relations of the choroid plexuses might 
be seen more clearly. 

Embryo H 173 was obtained from an aborted ovum of 42 x 32 
x 19 mm., presented by Dr. N. R. Engels of Chicago. The only 
available history was that the patient had missed two menses. 
The intact ovum was placed in physiological salt solution and 
kept at about O°C. for 11 hours. It was then opened and fixed 
in formalin-Zenker for 24 hours, stained in bulk with borax 
carmine, imbedded by the celloidin-paraffin method and cut 
10u thick in a plane parallel to the hindbrain. The embryo 
measured 19.1 mm., crown-rump length after fixation. The 
sections were counterstained on the slides with orange G. The 
total shrinkage was about 20 per cent. There are frequent 
mitoses in embryo and chorion. The brain was modeled at a 
magnification of 50 diameters with the aid of the Edinger pro- 


DEVELOPMENT OF THE CHOROID PLEXUS 85 


jection apparatus. One millimeter plates were used and every 
other section drawn except in the region of the foramen inter- 
ventriculare, where half millimeter plates were used and every 
section drawn. The epiphyseal region was modeled at a magni- 
fication of 100 diameters. Millimeter plates were used and every 
section was drawn. 

Embryo H 91 was obtained from an aborted ovum 50 x 34 
x 30 mm. presented by Dr. G. C. Dittmann of Chicago, whose 
data indicate a clinical pregnancy of 60 days. The ovum was 
left unopened in physiological salt solution for 10 hours, then 
opened and fixed in an 8 per cent solution of formaldehyde, 
neutralized with magnesium carbonate. It measured 27.8 mm. 
erown-rump length after fixation in formalin and showed a 
shrinkage of 13.6 per cent after imbedding in paraffin. It was 
stained in bulk in borax carmine and on the slide with orange 
G. It was cut in 20y sections, and modeled at a magnification 
of 40 diameters. Millimeter plates were used and every fifth 
section was omitted. 

Embryo H 41 was obtained from an ovum of 71 x 39 x 32 mm., 
presented by Dr. L. A. Beaton of Chicago. The chorion was 
opened and the entire ovum fixed in formalin. The crown- 
rump measurement of the embryo after fixation was 32.1 mm. 
and it showed a shrinkage of 10 per cent after imbedding in 
paraffin. The staining was the same as that of H91. This 
embryo was sectioned 20 in paraffin, and modeled at a magni- 
fication of 25 diameters. Millimeter plates were used and every 
other section was drawn. The region around the foramen of 
Monro was modeled at a magnification of 100 diameters; 2 mm. 
plates were used and every section was drawn. 

Two points in the technical procedure are to be emphasized 
because they are in large measure responsible for the exception- 
ally good preservation of the form relations of the delicate roof 
plate of the brain. Both of the older embryos (H 91 and H 41) 
had the cranial cavity opened by an incision in the line of the 
sagittal suture. Distortions due to unequal shrinkage of the 
brain and overlying structures were thereby in great measure 
avoided. All three were passed from 95 per cent alcohol to 


S86 PERCIVAL BAILEY 


ether-alcohol, then through 0.5, 1, 2 and 3 per cent celloidin, 
hardened in chloroform-aleohol, cleared in benzol and imbedded 
in paraffin under the air pump. 

The plane of section in each case is shown in figures 29; 30 and 
al; 

Several other human embryos were studied, the most helpful 
being embryo H 44 of the Chicago collection, which measured 
60.4 mm. after fixation in formalin. A transverse series from 
a 25 mm. pig in the collection of Dr. F. R. Lillie was also used. 


DESCRIPTION 
1. The 19 mm. embryo (H. 173) 


The recessus preopticus is well marked (fig. 18, r-pre.). The 
roof plate stretches dorsad from this recess as a thickened lam- 
ina (fig. 18, l.t.) to about the level of the sulcus separating the 
medial and intermediate roots of the corpus striatum. Above 
this point the roof plate narrows and extends cephalad and dor- 
sad (fig. 18, l.s.?), forms a broad arch (fig. 18, r.s.), and then 
passes caudad and dorsad (fig. 18, ¢.c.t-m.) as a still thinner mem- 
brane toward the velum transversum. Just in front of the 
velum transversum, the roof plate forms a small narrow arch 
(fig. 18, p.a. and fig. 2, p.a.), from the sides of which arise the 
lateral choroid plexuses. 

(Throughout these descriptions, narrow and wide are used of 
dimensions tangential to the ventricular surface, and thick and 
thin of dimensions perpendicular to the ventricular surface. 
For example, in figure 12, the tela chorioidea diencephali is thin 
and wide.) 

The velum transversum is well marked (fig. 18, v.¢.), indicat- 
ing the boundary in the roof plate between the diencephalon and 
telencephalon. 

The roof of the diencephalon (fig. 18, t.c.d.) is still narrow 
throughout most of its extent. It is also relatively thick, with 
several rows of nuclei in cross-section. It is narrowest at its 
posterior end and remains narrow almost to the anterior end of 
the thalamus, where it suddenly widens (fig. 1, ¢.c.d.). The 


DEVELOPMENT OF THE CHOROID PLEXUS 87 


entire structure, when viewed from above, is somewhat trumpet- 
shaped, with the bell at the anteriorend. There is no indication 
of plexus formation. The entire roof plate of the diencephalon 
is perfectly smooth. The commissura superior is clearly indi- 
cated (fig. 18, com.s.). 

The epiphyseal evagination is a hollow outgrowth (figs? 18 
and 3, e.e.). The top of the evagination is cupped and in this 
cup lies a ball of cells (figs. 3, 10 and 18, e.v.). This ball of cells 
has an irregular lumen in its center (fig. 10, e.v.). The cells of 
the ball stain more lightly than the cells of the cup, and the line 
of separation is fairly distinct. It will be seen that this ball 
of cells, while not in actual contact with the epidermis, approaches 
it very closely (fig. 10, e.v.). The epiphyseal evagination lies 
some distance cephalad of the commissura posterior (fig. 18, c.p.). 

The extent and morphology of the lateral choroid plexus is 
shown in figure 19. It is clearly divisible into two parts, an 
anterior part (figs. 2 and 19, p.c.v.l., p.a.) attached to the lateral 
margin of the paraphysal arch along its entire length, and by the 
taenia‘fornicis (fig. 2, t.f.) to the medial hemisphere wall im- 
mediately above and lateral to it; and a posterior part (figs. 1 
and 19, p.c.v.l., p-p.) attached by the taenia chorioidea (fig. 1, 
t.c.) to the lateral thalamic wall and by the taenia fornicis (fig. 
1, t.f.) to the medial hemisphere wall immediately below the 
hippocampus. The fissura chorioidea is very wide throughout 
most of the extent of the posterior part of the lateral choroid 
plexus. 

If the angle between the taenia chorioidea and the thalamic 
wall be followed anteriorly, it is found to be continuous with the 
velum transversum; if the angle is followed posteriorly, it con- 
tinues backward between the lateral thalamic wall and the 
medial hemisphere wall, turns downward between them and 
passes anteriorly and downward behind the optic nerve on the 
lateral wall of the hypothalamus (fig. 19, d-t. gr.). This groove 
is more clearly marked on the lateral wall of the 32 mm. embryo. 

The ependymal portion of the plexus is still relatively thick 
(fig. 27). The mesenchymal tissue of the plexus is typical em- 


SS PERCIVAL 


bryonal connective tissue. 
numerous near the ependyma. 


BAILEY 


The blood capillaries are particularly 


If one examines now the ventricular surface, the corpus 
striatum (fig. 19, c.s.) appears at the posterior point of attach- 
ment of the lateral plexus as a single ridge in the floor of the 


lateral ventricle. 


REFERENCE 


a.c.a., area chorioidea anterior. 

a.c.p., area chorioidea posterior. 

a.p., anterior pouch (of the tela chori- 
oidea diencephali) 

aq.S., aqueduct of Sylvius 

cb., cerebellum 

c.o., chiasma opticum 

com.s., commissura superior 

c.p., commissura posterior. 

c.s., corpus striatum 

c.s.l.7., corpus striatum, intermediate 
root 

c.s.l.r., corpus striatum, lateral root 

c.s.m.r., corpus striatum, medial root 

d.p., deep pit (in telencephalic roof 
plate) 

d-t.gr., di-telencephalic groove 

e.e., epiphyseal evagination 

ep., epidermis 

e.r., epiphyseal ridge 

e.v., epiphyseal vesicle 

f.a., fissura arcuata 

f.c., fissura chorioidea 

f.int., foramen interventriculare 

f.r., fasciculus retroflexus (Meynerti) 

h.a., and hip., hippocampal area 

hem., hemisphere 

h.s-t.r., habenulo-subthalamic ridge 

hy., hypothalamus 

inf., infundibulum 

l.s., lamina supraneuroporica 

l.t., lamina terminalis 

mt., metathalamus 

n.h., nucleus habenulae 

n-p.a., neopallial area 


As it is followed anteriorly, this ridge is soon 
divided by a groove into two portions. 


The lateral portion is 


LETTERS 


n.post., recessus postopticus 

o.n., optic nerve 

p.a., paraphysal arch 

p.c.v.l., plexus chorioideus ventriculi 
lateralis 

p.c.v.l., p.a., plexus chorioideus ven- 
triculi lateralis, pars anterior 

p.c.v.l., p.p., plexus chorioideus ven- 
triculi lateralis, pars posterior 

p.o.hy., pars optica hypothalami 

r.m., recessus mamillaris 

r.n., recessus neuroporicus 

r.post., recessus postopticus 

r.pre., recessus preopticus 

r.S., recessus superior 

s.a., striatal area 

.l., sulcus lhmitans 

s.m., stria medullaris 

s.M., suleus Monroi 

s-t., subthalamus 

t.c., taenia chorioidea 

t.c.d., tela chorioidea diencephali 

t.c.t.m., tela chorioidea telencephali 
medii 

t.f., taenia fornicis 

tg., tegmentum 

th., thalamus 

th.1, th.2., parts of thalamus 

th., e.s., thalamus, ependymal surface 

th., p.s., thalamus, pial surface 

t.i-c., taenia infrachorioidea 

t.r-p., telencephalic roof plate 

t.s-c., taenia suprachorioidea 

t.t., taenia thalami 

v.t., velum transversum 


nH 


89 


The medial portion 


A ys 
LK 


Ne 


v 
vw 


xX 10. Slide 


DEVELOPMENT OF THE CHOROID PLEXUS 
at first in the floor of the lateral ventricle but anteriorly comes to 


lie in the lateral wall (figs. 2 and 23, c.s.l.r.). 


| 
L 
, 


Fig. 1 Section through the diencephalon of the 19mm.embryo, H 173. X 13}. 


Slide 21, Sect. 11. 


Compare the photograph of this section, figure 27. 


Fig. 2 Section through paraphysis and hypothalamus of the 19 mm. embryo, 
Slide 23, Sect. 13. 


H 173. X13}. 


g 
a 
J 


Fig. 3 Section through epiphysis of the 19 mm. embryo, H 173. X 10. Slide 


12, Sect. 4. 


Fig. 4 Section through epiphysis of the 28 mm. embryo, H 91. 


9, Sect. 148. Ependyma solid black. 


QO) PERCIVAL BAILEY 


continues to the foramen interventriculare where it divides 
into two parts, the lateral part passing forward in the floor 
of the lateral ventricle (the intermediate root of the corpus 
striatum, figs. 2 and 23, c.s.i.r.), and the medial part passing 
through the foramen interventriculare, forming its floor, and 
extending in the lateral wall of the third ventricle behind the 
lamina terminalis as far as the recessus preopticus (fig. 18, 
c.s.m.yr.). It is very likely that this elevation which I have just 
_ described as the medial root of the corpus striatum contains in 
its lower end other things besides striatal tissue. I shall con- 
tinue to describe the entire elevation as the medial root of the 
corpus striatum, following the usage of His, and shall not enter » 
into a discussion of its internal structure. 

Above the taenia fornicis on the ventricular surface of the 
medial hemisphere wall may be seen an elevation (fig. 19, hip.) 
extending from the posterior pole of the hemisphere just above 
the choroid plexus, and beyond it over the foramen interventri- 
culare. This is the anlage of the hippocampus, at least in part. 
There is corresponding to it on the outer pial side of the hemi- 
sphere wall, a shallow groove (figs. 1 and 2, f.a.). This groove is 
not due to the folding in of a thin weak place in the wall, for the 
wall at this point is very definitely thicker (fig. 27). 

Turning now to the wall of the third ventricle, we find just 
below the tela chorioidea diencephali a low ridge (fig. 18, n-h.) 
extending from the epiphyseal evagination almost to the velum 
transversum. This is the habenula. From its anterior end a 
sharp ridge runs backward and downward to the subthalamus 
(fig. 18, h.s-t.r.). Above this ridge, the wall is shrunken and 
thin up to the habenular thickening. Below the ridge, is an 
elevation (fig. 18, th.1) which extends upward and forward in 
front of the habenula toward the velum transversum. 

The sulcus limitans is indicated on figure 18 by a dotted line 
running above the tegmentum and subthalamus, below the 
elevation last mentioned (fig. 18, th.1) and behind the corpus 
striatum to the preoptic recess. Just back of the corpus stria- 
tum and between it and the hypothalamus, the sulcus limitans 
runs into a very deep recess (fig. 2). 


DEVELOPMENT OF THE CHOROID PLEXUS 91 


Below the sulcus limitans lie the subthalamus (fig. 18, s-t.) 
and hypothalamus (fig. 18, hy.). The elevation which marks 
this region is very long. The pars optica is only indistinctly 
marked off (fig. 18, p.o.hy.). At its posterior end, the elevation 
divides into two portions, one of which, subthalamus, continues 
upward into the tegmentum, the other, hypothalamus, backward 
into the mammillary recess. 

The floor plate is thin and somewhat widened (fig. 3). The 
infundibulum (fig. 18, inf.) lies a considerable distance back of 
the optic chiasm. The postoptic recess is but poorly marked 
(fig. 18, r. post.). 


2. The 28 mm. embryo (H. 91) 


The entire telencephalon is not modeled in this embryo. ‘The 
model was made primarily to show the lateral choroid plexus. 
Those portions not modeled differ in no essential respect from the 
corresponding portions of the 32 mm. embryo. 

Immediately cephalad of the velum transversum (fig. 20, 
v.t.), which is clearly indicated, the roof plate forms a small 
arch (fig. 20, p.a.) to the sides of which are attached the lateral 
choroid plexuses. In front of the arch, the roof plate becomes 
very thin for a few sections (fig. 5, t.c.t.m.). Then as we pass 
cephalad of this thin lamina, we come to a region where the roof 
plate thickens in a peculiar manner (fig. 6, ¢.7-p.). 

The hemisphere wall on each side of the roof plate is thin and 
in it two distinct zones are discernable, a broader zone next the 
ependymal surface, where the nuclei are very numerous, mantle 
layer, and a narrower zone next the pial surface which is relatively 
free from nuclei, marginal layer. In the roof plate, however, 
these two. zones are not discernable, the nuclei are equally 
numerous throughout from the pial to the ependymal surface, 
and at the ependymal surface are loosely arranged, so that the 
outline is indefinite and irregular. This is the characteristic 
arrangement, and such an arrangement I have found in this region 
in five human embryos of about this age in the collection of the 
Department of Anatomy and in a 25 mm. pig embryo in the 


92 PERCIVAL BAILEY 


collection of Dr. I’. R. Lillie. In the embryo under consideration 
toward the anterior end of this region (fig. 7, t.r-p.) the morphol- 
ogy is somewhat different from that shown in figure 6. In the 
mid-line again, the nuclei are evenly distributed from ependymal 
to pial surface and are very numerous. Immediately on either 
side, however, the wall is greatly thickened in such manner that, 


6) 6 


Fig.5 Section through tela chorioidea telencephali medii of the 28mm. embryo, 
H91. X 50. Slide 28, Sect. 362. 

Fig.6 Section through roof plate of telencephalon medium of the 28 mm. 
embryo, H 91. X 50. Slide 29, Sect. 367. 


although the pial surface still forms a regular curve, the ependy- 
mal surface shows a deep notch in the mid-line. The outline 
of the ependymal surface is again indefinite owing to the loosely 
arranged cells. The marginal layer approaches almost to the 
midline. At the extreme anterior end of this region the notch 
disappears; the ependymal outline becomes definite; and the 
roof plate thickens markedly in an undoubted lamina terminalis 
(fis: 2OW et.) 


pew 


DEVELOPMENT OF THE CHOROID PLEXUS 93 


The tela chorioidea diencephali is very broad and very thin 
(fig. 11, t.c.d.). It does not exhibit any folding except at the 
extreme anterior end. Toward the posterior end of the tela, a 
narrow strip of the alar plate is curved lateralward, resembling 
the rhomboidal lip of the rhombencephalon, and which we may 
term the thalamic lip (fig. 9, ¢./.). 


= 


7 8 


Fig. 7 Section through the roof plate of telencephalon medium of the 28 mm. 
embryo, H91. X 50. Slide, 29, Sect. 372. 

Fig. 8 Section through the roof plate of telencephalon medium in the 82 mm. 
embryo, H 41. X 50. Slide 32, Sect. 2. 


The thalamic lip carries with it lateralward the taenia thalami 
(fig. 9, ¢.t.) and the roof plate. Toward the anterior end of the 
tela, the thalamic lip bends laterally more and more until its 
pial surface comes into contact with the pial surface of the tha- 
lamic wall (fig. 11, ¢.J.). The ependymal surface of the roof plate 
is hereby brought into contact with the ependymal surface of 
the thalamic lip (fig. 11). At the apex of the angle between the 
thalamic lip and the thalamic wall lies the stria medullaris (fig. 
11, s.m.). The entire tela when viewed from above is wedge- 


Q4 PERCIVAL BAILEY 


shaped, the anterior end being very broad. Heuser (’13) has 
noted a similar condition in the pig. The commissura superior 
is well marked (fig. 20, com. s.). . 

It will be seen that the roof of the epiphyseal evagination 
(fig. 20, e.e.) becomes epithelial for a short space in its uppermost 
portion. Just in front of this epithelial region there is a small 
agerevation of cells (fig. 4, e.v.) which recalls the ball of cells 
described in the epiphysis of the 19 mm. embryo. The nuclei 


9 10 


Fig.9 Section through the diencephalon of the 28 mm. embryo, H 91. - xX 
62. Slide 15, Sect. 228. 

Fig. 10 Section through the epiphysis of the 19 mm. embryo, H 173. X 180. 
Slide 12, Sect. 4. 


of these cells stain densely like the nuclei of the ependymal 
cells, and lie very close together. The cytoplasm is stained a 
deep yellow like the cytoplasm of the ependymal cells. They 
are surrounded by more lightly staining cells, and are nowhere 
in connection with the ependymal cells. The lateral wail of the 
epiphyseal evagination is very massive (fig. 4, e.r.). A ridge 
arises from the postero-superior portion of the lateral wall of 
the diencephalon and extends upward and backward to the 
epiphyseal evagination (fig. 21, e.7r.). 


DEVELOPMENT OF THE CHOROID PLEXUS 95 


The lateral choroid plexus (fig. 21, p.c.v.l.) is of considerable 
size, but does not nearly fill the ventricle (fig. 11, p.c.v.l.). In 
antero-posterior extent the plexus measures 1.55 mm., the ven- 
tricle measuring 2.97 mm. The anterior end of the plexus is 
much the larger and is less folded. Thompson (’09) has de- 
scribed a similar condition in the cat. The attachment of the 
plexus to the roof plate is now much narrowed, owing to the 
relatively small size of the paraphysal arch (fig. 20, p.c.v.l., p.a.). 
The taenia fornicis is now closely approximated to the taenia 
chorioidea, the fissura chorioidea being reduced to a narrow 

povul~ 
ted 


tl 
Ite 


11 
Fig. 11 Section through the diencephalon of the 28 mm. embryo, H 91. 
< 6%. Slide 24, Sect. 318. 
Fig. 12 Section through the diencephalon of the 32 mm. embryo, H 41. 


x 6%. Slide 24, Sect. 2. 


slit (fig. 11, f.c.). In its growth, the plexus has extended some 
0.25 mm. anterior to its point of attachment to the mid-line of 
the telencephalon. The connective tissure resembles that of the 
19 mm. embryo closely, except that the cells are farther apart, 
but the ependymal layer has become much thinner and consists 
clearly of a single layer of columnar cells, with the nuclei in the 
ends of the cells next the ventricular cavity (fig. 28). The 
ependyma of the posterior part of the plexus resembles more 
closely that of the 19 mm. embryo, there being several layers of 
irregularly arranged nuclei. The posterior part of the plexus 
is much attenuated and much more folded than the anterior end. 


Q6 PERCIVAL BAILEY 


In the wall of the third ventricle, the habenula is well marked 
(fig. 20, n.h.) and from its posterior end a ridge extends downward 
to the tegmentum (fig. 20, f.r.). Just below the habenula, the 
wall of the diencephalon is much thicker than in the 19 mm. 
embryo, and extends anterior to the velum transversum to form 
the posterior wall of the foramen interventriculare. 


3. The 32 mm. embryo (H. 41) 


Just in front of the velum transversum, as in the 28 mm. 
embryo, the roof plate forms a low arch (figs. 22 and 25, p.a.) 
to the sides of which are attached the lateral choroid plexuses. 
Immediately anterior to this arch the roof plate becomes very 
thin (fig. 25, t.c.t.m.). Anterior to this thin lamina, the roof 
plate is thickened again in the manner noted in the case of the 
28 mm. embryo (see fig. 6, é.r-p.). At the anterior end of this 
region (fig. 25, l.s.?), there is an indication of the median notch 
on the ependymal surface (fig. 8, t.r-p). It is, however, not 
nearly so well marked as the similar notch in the 28 mm. embryo 
(fig. 7, tr-p.) At the extreme anterior end of this region is 
a structure not found in any of the other embryos. A deep 
narrow pit extends into the roof plate from the pial surface, 
which causes the roof plate to project into the ventricle (figs. 13, 
14, and 25, d.p). This pit is found only in one section, with 
indications of it in the two adjacent sections. Immediately 
anterior to the pit, the roof plate thickens markedly and extends 
as a thickened lamina (fig. 22, /.t.) to the recessus preopticus. 
There is, however, a short distance anterior to the pit a shallow 
notch on the ventricular surface of the lamina terminalis (fig. 
Doerene? ). 

The tela chorioidea diencephali shows only slight imdications 
of longitudinal folding. The outwardly curved and very promi- 
nent thalamic lip (fig. 12, ¢.l.) is not in contact with the lateral 
wall of the thalamus. At the anterior end, the tela chorioidea 
diencephali is very broad, and a pouch (fig. 25, a.p.) arises which 
extends forward over the velum transversum. The whole tela 
resembles very closely the same structure in the 28 mm. embryo. 


DEVELOPMENT OF THE CHOROID PLEXUS 97 


The commissura superior is very readily identified (fig. 22, 
comM.s.). 

The epiphyseal evagination (fig. 22, e.e.) is, in all essential 
respects, identical with that of the 28 mm. embryo. 

The lateral choroid plexus (fig. 12, p.c.v.l.) also resembles 
that. of the 28 mm. embryo. The posterior portion is more swol- 
len, resembling more nearly the anterior part. The ependyma 


peovl 


13 14 


Fig. 13 Section through the roof plate of the telencephalon medium of the 
32 mm. embryo, H 41. X 50. Slide 32, Sect. 4. 

Fig. 14 Section through the roof plate of the telencephalon medium of the 
32mm.embryo,H 41. X 5. Slide 32, Sect. 4. 


is almost entirely a single layer of cells except at the extreme 
‘aie end of the plexus. The plexus extends through 4.44 

; the ventricle through 7.2 mm. The plexus extends 1.6 
bik amtemen to its most anterior point of attachment. 

On the ventricular surface we find again that the corpus 
striatum exhibits three roots at its anterior end (fig. 24, c.s.). 
The most medial root, in the lateral wall of the third ventricle, 
is narrow, lies immediately back of the lamina terminalis, and 


THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL. 26, No. | 


OS PERCIVAL BAILEY 


forms more the anterior boundary than the floor of the foramen 
interventriculare (fig. 22, c.s.m.r.). The intermediate root is 
also small, extending beyond the foramen interventriculare in 
the floor of the lateral ventricle. The intermediate root unites 
at its posterior end with the medial root, the two being separated 
from the lateral root by a groove. The lateral root is the lar- 
gest, and lies more in the floor than in the lateral wall of the 
ventricle (fig. 24, c.s.l.r.). The two ridges, one formed by the 
lateral root and the other by the union of the medial and inter- 
mediate roots, extend backward some distance and then turn 
sharply downward as a single elevation (fig. 24). 

The hippocampus (fig. 12, hip.) is a broad, slightly thickened 
portion of the medial hemisphere wall, just above the taenia 
fornicis. The external sulcus is very shallow and broad. 

The habenular ridge is prominent, extending on the ventric- 
ular surface of the diencephalon just below the tela chorioidea 
diencephali from the region of the velum transversum to the 
epiphysis (fig. 22, n.h.). The ridge extending from the posterior 
end of the habenula to the tegmentum is also prominent (fig. 
DONE s 

The most striking feature, probably, is the great thickening 
of the posterior extremity of the thalamus. The anterior ex- 
tremity of the thalamus extends well into the foramen inter- 
ventriculare, some distance anterior to the velum transversum. 

The sulcus limitans is indicated on figure 22 by a dotted line. 
It is deepest below the posterior pole of the thalamus. Behind 
the lamina terminalis it is relatively shallow. 

The hypothalamic region is marked mainly by its great length. 
The mammillary recess is indicated (fig. 22, r.m.); the floor plate 
is wide and thin. The infundibular recess (fig. 22, inf.) is near 
the optic chiasm but is separated from it by an unmistakable 
postoptic recess (fig. 22, r.post.). There is no clear external 
division between hypothalamus and subthalamus. . 


DEVELOPMENT OF THE CHOROID PLEXUS 99 


DISCUSSION 
1. Telencephalon 


a. Recessus preopticus. In all three embryos there is no 
doubt about the identity of this recess. 

b. Velum transversum. Just as obvious is the location of the © 
velum transversum. It is marked by the groove running across 
the roof plate joining the anterior ends of attachment of the 
tela chorioidea diencephali to the thalamus. At the lateral 
end of the velum transversum the taenia thalami meets the 
taenia chorioidea (fig. 26) and at this point the velum trans- 
versum becomes continuous with the angle between the taenia 
chorioidea and the lateral thalamic wall. In figure 26 an arrow 
lies in the angle between the taenia chorioidea and the lateral 
thalamic wall and continues in the velum transversum. The 
position of the head of the arrow in the mid-line is shown in fig- 
ure 25. It was noted in the account of the 19 mm. embryo 
and appears more clearly in the 32 mm. embryo, that if this angle 
be followed backward it passes behind the attachment of the 
hemisphere to the thalamic wall, and then as a diagonal groove 
downward and forward across the lateral wall of the hypothala- 
mus and ends at the optic chiasm. This is the di-telencephalic 
groove (figs. 14 and 19, d-t.gr.) of Johnston. 

With the preoptic recess and the velum transversum fixed, 
the extent of the telencephalic roof plate is determined. Be- 
tween the preoptic recess and the velum transversum should 
appear the lamina terminalis, recessus neuroporicus, lamina 
supraneuroporica, recessus superior, tela chorioidea telencephali 
medii, and paraphysis. 

c. Paraphysis. Just in front of the velum in each embryo 
is a small arch (figs. 18, 20 and 25, p.a.) relatively largest in 
the 19 mm. embryo, and smallest in the 32 mm. embryo. In 
each case it les immediately in front of the velum transversum. 
In each case also the lateral choroid plexuses arise from its 
sides. There can be no doubt that this is the paraphysal arch. 
No indication of the development of a glandular structure 
could be found. The resemblance of the paraphysal arch in 


LOO PERCIVAL BAILEY 


the 19 mm. embryo to those of the 10 mm. cat and 20 hr. chick 
figured by Tilney (715) and of a4 mm. embryo of Platydactylus 
mauritanicus, figured by Tandler and Kantor (’07), is rather 
striking. 

d. Recessus neuroporicus. An identification of this point is 
absolutely essential to a final definition of the boundary be- 
tween lamina terminalis and lamina supraneuroporica. By 
recessus neuroporicus is meant the most caudal point, i.e., last 
point, of closure of the neuropore. In many vertebrates at 
this point a recess appears on the ventricular surface of the roof 
plate (Johnston, ’09). 

It was found impossible to identify with certainty this point 
in the embryos examined. In the 19 mm. embryo, and the 28 
mm. embryo, no such recess is apparent. In the 32 mm. em- 
bryo, there has been noted just in front of the pit in the roof 
plate, a shallow notch on the ventricular surface (fig. 25, r.n.?). 
But there is no evidence that this is the recessus neuroporicus, 
since no such notch appears in either younger embryo. 

e. Lamina terminalis. The upper end of the lamina terminalis, 
as defined by Johnston, has not been determined because the 
recessus neuroporicus is not apparent. Concerning the major 
portion of the lamina, however, there can be no doubt. The 
thick lamina above the recessus preopticus is unmistakable 
(figs. 18, 20 and 22, 1.t.). 

f. Tela chorioidea telencephali medii. Just in front of the 
paraphysal arch in the 28 mm. embryo, and the 32 mm. embryo, 
the roof becomes a single layer of flattened cells (fig. 5, t.c.t.m.). 
This is certainly tela chorioidea telencephali medii. The iden- 
tification of this tela in the 19 mm. embryo is not so easy. How- 
ever, I am inclined to identify the summit of the greater arch 
of the roof plate (fig. 18, 7.s) as the recessus superior. This 
would make the roof between this point (fig. 18, r.s.) and the 
paraphysal arch, tela chorioidea telencephali medii (fig. 18, 
t.c.t.m.). The roof plate here is somewhat thinner than the 
lower limb of the greater arch (fig. 18, l.s.?) and the adjacent 
hemisphere wall is considerably thinner. The angulus terminalis 


DEVELOPMENT OF THE CHOROID PLEXUS 101 


of His probably represents all the membranous parts anterior 
to the velum transversum. 

g. Lamina supraneuroporica. Between what is obviously 
lamina terminalis and what is just as obviously tela chorioidea 
telencephali medii in the 28 mm. embryo and the 32 mm. em- 
bryo, lies the peculiar thickening of the roof plate before noticed 
(fig. 6, t.r-p.) which seems by a process of elimination to be lamina 
supraneuroporica. Whether this interpretation is correct, I 
cannot say, because I have not had material with which to fol- 
low this region in its later development. There is in the 32 
mm. embryo, a short portion of the roof plate between the deep 
pit (fig. 25, d.p.) and the shallow recess on the ventricular sur- 
face (fig. 25, r.n.?) which may be lamina supraneuroporica. 
In the 19 mm. embryo, the lower limb of the greater arch (fig. 
18, l.s.?) seems its most likely location. 

h. Plexus chorioideus ventriculi lateralis. The problem of 
the formation of the lateral choroid plexuses is one of consider- 
able difficulty, because of the complex morphological relations 
involved, but if a few facts of development be remembered, 
the problem becomes relatively simple. 

If the brain of a half-grown frog tadpole be examined, it will 
be found that ‘‘the membranous roof of the forebrain ventricle 
is attached to the massive wall of the hemisphere by the taenia 
fornicis which is directly continuous caudad with the taenia 
thalami’”’ (Herrick, ’10). The taenia thalami is the attachment 
of the roof plate posterior to the velum transversum to the 
lateral wall of the thalamus. It might be added also that the 
taenia fornicis becomes continuous with the taenia thalami 
at the lateral end of the velum transversum. In the middle 
of the membranous roof in front of the velum transversum arises 
the paraphysis. There is no plexus lateralis in the larval or 
adult frog, but in urodele Amphibia it is between the paraphysis 
and the taenia fornicis that the lateral choroid plexus makes its 
appearance, pushing into the ventricle. Warren (’05) in de- 
scribing the development of the lateral plexuses of Necturus 
maculatus says, ‘““The telencephalic plexus develops from the 


102 PERCIVAL BAILEY 
paraphysal arch. ” “he plexuses of the hemispheres 
arise on either side from the origin of the telencephalic plexus 
and pass into the lateral ventricles. ne 

If we turn now to the human embryo and examine, say, the 
19 mm. embryo, H 173, we find that it is easy to trace the tela 
chorioidea telencephali medii over the paraphysal arch and 
velum transversum to the tela chorioidea diencephali. If, 
however, we attempt to follow backward the taenia fornicis, it 
will not be found to be continuous with the taenia thalami at 


Fig. 15 Median view of a model of the forebrain in His’ 6.9 mm. embryo, 
Bio: 


the lateral end of the velum transversum, but is separated from 
the taenia thalami by the fissura chorioidea. In order to ex- 
plain the difference between the condition in the-frog and in 
the human embryo it is necessary to analyze more closely some 
younger human embryos. 

Figure 15 shows a medial view of the forebrain of His’ embryo 
Br 3, 6.9 mm. in length. On it is indicated by a row of crosses 
a line homologous with the taenia fornicis et taenia thalami in 
the tadpole. The anterior limb of the di-telencephalic groove 
is marked by small circles and labeled area chorioidea posterior 
for reasons which will appear later. On the opposite side of 


DEVELOPMENT OF THE CHOROID PLEXUS 103 


the taenia fornicis from the area chorioidea posterior, that is, 
between the taenia fornicis and the mid-dorsal line, is placed 
another area of circles labeled area chorioidea anterior. It is 


Fig. 16 Median view of a model of the forebrain in His’ 13.6 mm. embryo, CR. 

Fig. 17 Hypothetical view of the median wall of the cerebral hemisphere 
from His’ 13.6 mm. embryo, CR. The hemisphere has been excised along the 
line b-a-d in figure 16. 


104 PERCIVAL BAILEY 


in this latter region that the lateral choroid plexus makes its 
appearance in urodele Amphibia as we have just remarked, 
and here also it makes its first appearance in the human em- 
bryo. The hippocampal area is not apparent in this embryo but 
later developments show that its anlage must lie in some such 
position as indicated in figure 15. The part of the telencephalon 
which is evaginated to form the cerebral hemisphere is the part 
anterior to the line d-a-b. The division between telencephalon 
and diencephalon is represented by the line b-a-c. 

Keeping these relations in mind, we may easily understand 
His’ embryo CR, 13.6 mm. in length. Figure 16 shows a me- 
dial view of the forebrain. An invagination of the telencephalic 
roof has taken place between the taenia fornicis and the mid- 
line, forming the fissura chorioidea in the region which was 
marked area chorioidea anterior in figure 15. The hemisphere 
vesicle has enlarged mainly by enlargement and evagination 
of the neo-pallial area (fig. 15, n-p.a.), so that the hippocampal 
area which in figure 15 lay below the roof plate and in front of 
the area chorioidea posterior, now lies above the roof plate and 
behind the area chorioidea posterior, and its ependymal face 
has now come to look lateralward. So also with the area chorioi- 
dea posterior, which lies now in the evaginated cerebral hemi- 
sphere behind the di-telencephalic groove, and its ependymal 
face is now its lateral face. When the hemisphere evaginated, 
the wall bent for the most part along the line of the taenia 
fornicis in figure 15, the di-telencephalic groove as far as the 
point a, and then along the line a-d. 

If now the hemisphere in embryo CR be excised. along a 
line homologous with the line b-a-d in figure 15, the relations 
of the area chorioidea posterior and area chorioidea anterior 
are more clearly seen. The medial surface of such a hypothetical 
hemisphere is represented in figure 17. The area chorioidea 
anterior has invaginated into the lateral ventricle. The result- 
ing plexus has been cut off in figure 17, leaving the fissura chori- 
oidea. Since the invagination takes place between the taenia 
fornicis and the mid-line of the telencephalic roof, the taenia 
fornicis lies along the lateral and upper margin of the fissura 


DEVELOPMENT OF THE CHOROID PLEXUS 105 


chorioidea, or in other words, along the line of attachment of 
the plexus to the medial hemisphere wall. If, therefore, we 
wish to follow the taenia fornicis et thalami in the embryo CR, 
we must follow the line marked with crosses in figures 16 and 17 
along the wall of the diencephalon (figs. 16 and 17, ¢.t.), across 
the velum transversum, and then between the area chorioidea 
posterior and area chorioidea anterior above and lateral to the 
fissura chorioidea (figs. 16 and 17, t,f.). 

It will be seen (fig. 17) that the area chorioidea anterior lies 
entirely anterior to the velum transversum. The lateral choroid 
plexus never approaches the mid-line except along the sides 
of the paraphysal arch, anterior to the velum transversum. 
The area chorioidea posterior has relations entirely analogous with 
its relations in figure 15, except for its change of face. It lies 
between the hippocampus and the di-telencephalic groove, and 
adjoins the posterior end of the taenia fornicis on the upper 
side, which in figure 15, was the lower side of the taenia. 

This stage in the process bears a close resemblance to the 
condition in the brain of Platydactylus mauritanicus, as de- 
scribed by Tandler and Kantor (’07). Concerning the develop- 
ment of the structures in the region of the foramen of Monro, 
they write (italics mine): 


Die zwischen den beiden Foramina Monroi gelegene Decke des 
Ventriculus impar [telencephalon medium] ist rein epithelialer Natur 
and grenzt sich lateralwdrts durch je eine deutliche Furche Sulcus tegmenti 
ab. 


The lateral edge of this sulcus tegmenti is the taenia fornicis. 


Die Fissura chorioidea, welche von hinten her den Raum des Fora- 
men Monroi einengt, entwickelt sich, wie die Durchsich! der Serie lehrt, 
derart, dass sie im vorderen Abschnitt aus dem Sulcus tegmenti selbst, 
hinten aber oberhalb dieser Furche entsteht, und hier die mediale 
Hemisphiren wand einschneidet. 

Der Plexus chorioideus des Ventriculus lateralis stiilpt niamlich 
nur ein ganz kurzes Stiick der Hirnwand ein. Die EHinstiilpwngsstelle 
selbst, liegt wie man um Stadium V zeigen kann, gerade dort, wo das 
hintere Ende des Sulcus tegmenti die Decke des Telencephalon impar von 
der medialen Hemisphdrenwand absetzt. 


106 PERCIVAL BAILEY 


Here again, as in the brain of Necturus maculatus, the half 
grown frog tadpole, and His’ embryo CR, the taenia fornicis 
ean be followed along the lateral margin of the sulcus tegmenti 
past the velum transversum to the taenia thalami. But when 
the sulcus tegmenti invaginates to form the fissura chorioidea, 
the posterior end of the taenia fornicis and a small part of the 
medial hemisphere wall are drawn in also. This latter process 
is carried much farther in the human embryo. 

In the evagination of the hemisphere in the human embryo, 
the point marked x in figure 15, at the junction of the hippo- 
campal area, striatal area, and area chorioidea posterior, re- 
mains always at the posterior pole of the hemisphere. The 
result on the corpus striatum, as the hemisphere extends back- 
ward, is to draw out the tail of the caudate nucleus. The result 
upon the area chorioidea posterior is to draw it out in a thin 
lamina marked in figure 17 by small circles. 

The area chorioidea posterior now buckles into the lateral 
ventricle, and we have the condition found in the 19 mm. em- 
bryo, H 173. The area chorioidea posterior is clearly shown 
by figure 19, being all of the plexus back of the point marked 
z, and has buckled inward only slightly, leaving a very wide 
fissura chorioidea. The point z lies opposite the velum trans- 
versum. 

It thus appears that the lateral choroid plexus is composed 
of two parts, a pars anterior plexus chorioidei ventriculi lateralis 
which is formed by the invagination of the area chorioidea 
anterior, between the paraphysal arch and the taenia fornicis, 
and a pars posterior plexus chorioidei ventriculi lateralis formed 
by the infolding of the area chorioidea posterior in the medial 
wall of the hemisphere. 

If the taenia fornicis be now followed in the 19 mm. embryo, 
it will not be found to become continuous with the taenia thalami 
because the area chorioidea posterior, to which it was attached 
toward its posterior end, has now buckled into the ventricle. 
If, therefore, we follow the attachment of the lateral choroid 
plexus to the medial hemisphere wall, we follow the taenia forni- 


DEVELOPMENT OF THE CHOROID PLEXUS 107 


cis as far as the area chorioidea posterior, then along the upper 
margin of the area chorioidea posterior (anterior margin in fig. 15) 
called also taenia fornicis in adult anatomy, around its posterior 
extremity (lower extremity in fig. 15), and back along its lower 
margin (posterior margin in fig. 15) called also taenia chorioidea, 
to the lateral end of the velum transversum where we finally 
reach the taenia thalami. 

It thus appears that the portion of the taenia fornicis of human 
anatomy which lies adjacent to the pars posterior plexus cho- 
rioidei ventriculi lateralis is not homologous to the taenia fornicis 
in Anura and would better be called taenia suprachorioidea, 
and the taenia chorioidea correspondingly termed taenia infra- 
chorioidea. The illustrations have been labeled in accordance 
with established usage. To place them in accord with the 
foregoing conclusions, in figures 1, 11, 12, 18, 24 and 26 taenia 
fornicis should be changed to taenia suprachorioidea, and taenia 
chorioidea to taenia infrachorioidea. 

In later stages the taeniae infra- et suprachorioidea become 
approximated closely as is found in the 28 mm. embryo and in 
the 32 mm. embryo, and the fissura chorioidea is reduced to a 
narrow slit, its axis in the plexus of the 19 mm. embryo, lying 
probably along the dotted line in figure 19. 

In the 28 mm. embryo and the 32 mm. embryo, it is impos- 
sible to distinguish the dividing line between the pars anterior 
and the pars posterior of the lateral choroid plexus, and the taenia 
fornicis in its restricted sense is relatively of very short length. 

i. Sulcus limitans. The sulcus limitans is lost in a deep re- 
cess between the corpus striatum and hypothalamus. ‘This 
recess has disappeared by fusion of its walls in the 32 mm. em- 
bryo. If the sections of the 19 mm. embryo be followed, the 
beginning of this process can be readily seen (fig. 2). 

j. Corpus striatum. There is nothing extraordinary about the 
corpus striatum in either embryo. The approximation of the 
thalamus and corpus striatum in- the foramen interventriculare 
(fig. 22) is of interest when one remembers Goldstein’s work. 
Of course, the entire connection between the thalamus and 


LOS PERCIVAL BAILEY 


corpus striatum is not formed by fusion, and there is as yet no 
fusion here. The intermediate root of the corpus striatum prob- 
ably extends into the medial hemisphere wall, but the external 
morphology does not suggest it. ‘ 


2. Diencephalon 


a. Tela chorioidea diencephali. It is to be noted that the 
tela chorioidea diencephali shows no indications of folding 
except at its anterior extremity. In the 32 mm. embryo, a 
pouch arises at the anterior end and extends forward over the 
velum transversum and the paraphysal arch (fig. 25, a.p.). 
Streeter, in Keibel and Mall’s textbook, as was mentioned in the 
history, makes the statement that “Orally this choroid roof [of 
the third ventricle] is continued into the telencephalon where 
it forms a pointed pouch overlapping the lamina terminalis and 
the contained commissures. . . . The anterior choroidal 
pouch has been homologized with the paraphysis of the lower 
vertebrates.”’ There is not up to this stage any pointed pouch 
in the telencephalic roof. The pouch noted above lies in the 
roof of the diencephalon, just back of the velum transversum 
and hence cannot be the paraphysis. The true paraphysal 
arch has been pointed out above. In comparing Francotte’s 
figures with this region in embryos of approximately the same 
age in the Chicago collection, I feel sure that it was this anterior 
pouch of the tela chorioidea diencephali which he described as 
the paraphysis. 

I have not followed the thalamic lip fully in later stages, but 
in an embryo of 60.4 mm. greatest length, H 44 of the Chicago 
collection, the thalamic lip is fused with the lateral thalamic wall 
toward the anterior end. 

b. Epiphysis. The epiphysis, in the 19 mm. embryo especially, 
shows marked indications of a differentiation into epiphyseal 
stalk and epiphyseal vesicle. Such a condition is very char- 
acteristic of lower vertebrates, especially reptiles, but has never 
before been noticed in the human embryo. The epiphyses of 
the 28 mm. and 32 mm. embryos show similar indications. 


DEVELOPMENT OF THE CHOROID PLEXUS 109 


In the 32 mm. embryo and in the 28 mm. embryo, the tela 
chorioidea continues on to the epiphyseal evagination. A simi- 
lar condition has been noted in the cat by Thompson (’09). 

c. Habenula. The character of the ridge running from the 
anterior end of the habenula to the subthalamus in the 19 mm. 
embryo (fig. 18, h.s-t.r.) is not apparent. It is probably only 
a temporary fold mechanically produced by inequalities in 
development of the thalamic wall. It has entirely disappeared 
in both older embryos. 

The ridge extending from the posterior end of the habenular 
prominence in the 28 mm. embryo and the 32 mm. embryo 
(figs. 20 and 22, f.r.) indicates the position of the fasciculus 
retroflexus (Meynerti). 

d. Thalamus. In the 19 mm. embryo, only the anterior, 
inferior part of the thalamic wall (fig. 18, th./.) lying between 
the sulcus limitans and the habenulo-subthalamie ridge is thick- 
ened. In the 28 mm. embryo, the ridge has disappeared and 
the anterior portion of the thalamic wall which lay above the 
ridge is thickened (figs. 18 and 20, th. 2). This region probably 
contains the principal nuclei of the thalamus. The posterior 
extremity of the thalamus (figs. 18 and 20,) is still somewhat 
flattened and thin in the 28 mm. embryo, but becomes very thick 
in the 32 mm. embryo (fig. 22). 

e. Sulcus limitans. There is no difficulty in following the sul- 
cus limitans. In the 32 mm. embryo under the posterior pole 
of the thalamus, it is very deep. 

f. Hypothalamus. The hypothalamus in all these embryos 
is of great antero-posterior extent. It will be noticed also that 
the infundibulum, especially in the 19 mm. embryo, lies a con- 
siderable distance back of the optic chiasm. -It seems quite 
probable that, as Johnston (’09) has remarked, in His’ model 
of the embryo CR of 13.6 mm., the recess marked infundibular 
is really postoptic, and the real infundibular recess lies back 
of it and is labelled tuber cinereum. Only in the 19 mm. em- 
bryo is the subthalamus clearly separated externally from the 
hypothalamus. 


L1LO PERCIVAL BAILEY 


SUMMARY 


1. The plexus chorioideus ventriculi lateralis is composed of 
two distinet portions, of which the anterior is developed from 
the roof plate in the angle between the paraphysal arch and the 
medial wall of the hemisphere, and the posterior from that 
part of the medial wall of the hemisphere just anterior to the 
di-telencephalic groove and homologous to the anterior limb of 
the velum transversum. 

2. The taenia fornicis of adult human anatomy, except for 
its extreme anterior end, is not homologous with the taenia 
fornicis of Anura. 

3. The position of the recessus neuroporicus could not with 
certainty be ascertained, and identification of the lamina supra- 
neuroporica was therefore uncertain. 

4. The tela chorioidea telencephali medii is present in the 
embryos of 28 and 32 mm. 

5. The paraphysal arch can be followed to the embryo of 32 
mm., as an arch of the roof plate of the telencephalon. It lies 
just anterior to the velum transversum and from its sides arise 
the lateral choroid plexuses. The anterior pouch of the choroid 
plexus of the third ventricle lies in the diencephalon and is not, 
therefore, homologous to the paraphysis of the lower verte- 
brates. No indication of the development of a glandular strue- 
ture was found. 

6. The velum transversum can be traced up to the 32 mm. 
embryo, joining the anterior extremities of attachment of the 
tela chorioidea diencephali to the lateral thalamic wall, and 
its groove is continuous laterally with the angle between the tae- 
nia chorioidea and the lateral thalamic wall, the di-telencephalic 
groove. 

7. The tela chorioidea diencephali is not folded nor vascular- 
ized except for its extreme anterior end in the embryo of 32 mm. 
It is broadened by the formation of a thalamic lip, very much 
resembling the rhomboidal lip of the rhombencephalon. 

8. The epiphyses of all three embryos show indications of 
the presence of the homologue of the pineal vesicle of the lower 


DEVELOPMENT OF THE CHOROID PLEXUS 111 


vertebrates. In the embryos of 28 and 32 mm., the tela chor- 
ioidea diencephali is continued on to the epiphyseal outgrowth. 

9. The position of the fasciculus retroflexus (Meynerti) is 
indicated in the embryos of 28 and 32 mm. by a pronounced 
ridge. 

10. There is evidence that the connection of the corpus stria- 
tum and thalamus is thickened by fusion of the medial root of 
the corpus striatum with the anterior extremity of the thalamus 
in the foramen interventriculare. 

11. The sulcus limitans is lost in a very deep recess in the 
embryo of 19 mm. between the corpus striatum and hypothala- 
mus, and this recess disappears in later stages by fusion of its 
walls. 

12. The length of the hypothalamus in embryos from 19 to 
32 mm. is relatively very great. In the embryo of 19 mm. the 
subthalamus is separated externally from the hypothalamus. 

13. The infundibular outgrowth is still some distance back of 
the optic chiasm at 19 mm. and becomes shifted nearer only in 
much later stages. 


LITERATURE CITED 


Dexter, F. 1902 The development of the paraphysis in the common fowl. 
Am. Jour. Anat., vol. 2, no. 1, pp. 13-24. 

Fatvre 1854 Les plexus choroides. Revue Medicale de Paris. British and 
Foreign Med. Chir. Review. 

Finpiay, J.W. 1899 The choroid plexuses of the lateral ventricles of the brain. 
Brain, vol. 22, pp. 161-202. 

Francottr, P. 1894 Note sur l’oeil parietal, l’epiphyse, la paraphyse et les 
plexus choroides du troisieme ventricule. Bull. de l’acad. royale, 
des sci. d. Belg., 3 Serie, T. 27, N. 1, pp. 84-113. 

Gaupr, E. 1898 Zirbel, Parietalorgan und Paraphysis. Ergeb. d. Anat. u. 
Entwick., Bd. 7, pp. 208-285. 

GoutpsTteIn, K. 1903 Beitriige zur Entw-gesch. d. Menschl. Gehirns. Arch. 
f. Anat. u. Phys., Anat. Abt. 

Haecket, E. 1860 Zur normalen und pathologischen Anatomie des Plexus 
Choroideus. Schmidt’s Jahrbiicher. Virchow’s Archiv, Bd. 16, 
1859, pp. 253-289. 

Herrick, C. J. 1910 Morphology of the forebrain in Amphibia and Reptilia. 
Jour. Comp. Neur., vol. 20, pp. 413-547. 

Heuser, C. H. 1913 The development of the cerebral ventricles in the pig. 
Am. Jour. Anat., vol. 15, pp. 215-239. 


113 PERCIVAL BAILEY 


His, W. 1880 Die Formentwickelung des menschlichen Vorderhirns vom Ende 
des ersten bis zum Beginne des dritten Monates. Abhdl. d. kénigl. 
sich. Akad. d. Wiss., math. phys. Cl., Bd. 15. 
1893 Uber das frontale Ende des Gehirnrohres. Arch. f. Anat. u. 
Entw. pp. 157-171. 
1904 Die Entwicklung des menschlichen Gehirns. Leipzig. 

Hocusterter, F. 1913 Uber die Entwickelung der Plexus Chorioidei der 
Seitenkammern des menschlichen Gehirns. Anat. Anz., Bd. 45, pp. 
225-238. 

Jounston, J.B. 1909 On the morphology of the forebrain vesicle in vertebrates. 
Jour. Comp. Neur., vol. 19, pp. 457-539. 
1910 The evolution of the cerebral cortex. Anat. Rec., vol. 4, 
no. 4. 
1913 The morphology of the septum, hippocampus, and pallial 
commissures in reptiles and mammals. Jour. Comp. Neur., vol. 23, 
pp. 371-478. 

Lacui, P. 1888 La tela corioidea superiore e i ventricoli cerebrali dell’uomo. 
Atti della Soc. Tosc. d. Sc. nat., vol. 9, Fasc. 1. 

LuscuKa, H. von 1885 Die Adergeflechte des menschlichen Gehirns. Bd. 6, 
174 pp., 4 pl., 4°. Berlin, G. Reimer. Schmidt’s Jahrbiicher, G. 88, 
p. 376. (Review.) 

Meek, W.J. 1907 A study of the choroid plexus. Jour. Comp. Neur., vol. 17, 
pp. 286-306. 

Minor, C. 8S. 1901 On the morphology of the pineal region, based on its de- 
velopment in Acanthias. Am. Jour. Anat., vol. 1. 

SELENKA, E. 1891 Das Stirnorgan der Wirbelthiere. Biolog. Centralbl., Bd. 10, 
pp. 323-326. 

SmiruH, G. Exxtior 1903 On the morphology of the cerebral commissures. 
Trans. of the Linnaen Society, London, vol. 8, part 12, July. 

Sterzi, G. 1907 I] sistema nervoso centrale deivertebrate. Vol. 1, Ciclostomi, 

Padua. 

STREETER, G. L. 1912 The development of the central nervous system. Kei- 
bel and Mall’s Human Embryology. 

TANDLER, J. and Kantor, H. 1907 Beitrige zur Entwickelung des Vertebra- 
tengehirnes. I. Die Entwicklungsgeschichte des Gekohirnes. Anatom. 
Hefte, Bd. 338, Heft 101. 

TuHomeson, P. 1909 Description of the brain of a foetal cat, 20 mm. in length. 
Jour. Anat. Physiol., vol. 48, pp. 134-145. 

TILNEY, FreperiIcK 1915 The morphology of the diencephalic floor. Jour. 
Comp. Neur., vol. 25, no. 3, pp. 213-282. 

WarRREN, JOHN 1905 The development of the paraphysis and pineal region in 
Necturus maculatus. Am. Jour. Anat., vol. 5, pp. 1-29. 


DEVELOPMENT OF THE CHOROID PLEXUS 113 


Fig Fig nese Figs 3+10 
27-6 v f. ta o COM.S. | _&. Y. 
pa, L ; ‘ PS Y -" 


PCVL Pay oo) 


<a 


he 
ae 
? 


acer i ii oor 


" inf 
pohy 


Fig. 18 Median view of a model of the forebrain from the 19 mm. embryo, 
H 173. ‘224. The dotted line follows the sulcus limitans. The limiting 
membrane bounding the epiphyseal vesicle, e.v., ventrally in this figure should 
be represented as incomplete; ef. figure 10. 

In this and some of the following figures the planes of the cross sections in 
the text-figures are indicated. The arrows at the right and left ends of the 
models are marked with the slide and section numbers of the first and last 
sections represented in the models. The reference letters for this and the 


following figures are found on page 88. 


THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL. 26, NO. 1 


114 PERCIVAL BAILEY 


Figs.3rl0 
| Fig. | Fig.2 . 


eex é J 5G aNGe nee | | 27-6 


ihrem. 


Ma eee EN 
pete! ///T/]7-~ ins 


19 


Fig. 19 Lateral view of a model of the forebrain from the 19 mm. embryo, 
H178. X 223. The lateral wall of hemisphere has been cut away, exposing the 


lateral choroid plexus. Median surface of this model is shown in figure 18. 


Fig. 20. Median view of a model of a portion of the forebrain from the 28 mm. 


embryo, H91. X 18. 


Fig. 21 Lateral view of a model of a portion of the forebrain from the 28 mm. 
embryo, H91. X18. The lateral hemisphere wall and corpus striatum have 
been cut away, exposing the lateral choroid plexus. Median surface shown in 


figure 20. 


DEVELOPMENT OF THE CHOROID PLEXUS 115 


20 


32-404 j 


21 


L116 PERCIVAL BAILEY 


Gr 5.1 kr ~ 


Fig. 22. Median view of a model of the forebrain from the 32 mm. embryo, 
H 41. 124. The dotted line follows the sulcus limitans. 

Fig. 23 View of the anterior end of the model of the forebrain from the 19mm. 
embryo, H 173, shown in figure 18. x 223. View taken slightly from the medial 
side and above. A portion of the median hemisphere wall and lamina terminalis 
has been removed exposing the corpus striatum. 

Fig. 24 View of the posterior extremity of the corpus striatum, in the model 
of the forebrain of the 32 mm. embryo, H 41, shown in figure 22. X 123. View 
taken from above, behind and lateralward. 


ro 


rpre: C.O 


23 


24 
117 


118 PERCIVAL BAILEY 


' 
‘ 
’ 
‘ 
4 
i 
y 
1 
' 


: STREEDAIN, 22: 
i | 


ome 


Fig. 25 Median view of a model of the region around the foramen interven- 
triculare from the 32 mm. embryo, H 41. X 50. (The entire forebrain of this 


embryo is shown in figure 22.) The model was made at a magnification of 100 
diameters. Anterior end to the left. 


DEVELOPMENT OF THE CHOROID PLEXUS 119 


STRAABAINS 


. 


Fig. 26 Lateral view of a model of the region around the foramen inter- 
ventriculare from the 32 mm. embryo, H 41. »X 50. Median view of the same 
model is shown in figure 25. For help in orientation, the approximate position of 
the roof plate, which lies on the opposite side of the hemisphere wall, has been 
projected through as a dotted line. The location of the taenia thalami is indi- 
cated by a row of dots and dashes. An arrow lies in the di-telencephalic groove 
and continues across in the groove of the velum transversum. The position of 
its head may be seen by reference to figure 25. Anterior end of model to theright. 


Fig. 27 Photograph of a section through the diencephalon of the 19 mm. em- 
bryo, H173. X 162. Slide 21, Sect. 11. The break in the hemisphere wall to 
the left is an injury to this individual section. Compare figure 1, drawn from the 


same section. 
Fig. 28 Photograph of a section through the lateral choroid plexus of the 


28 mm.embryo,H 91. X 20. Sect. 355. Section taken midway between figures 


7 and il, as shown in figure 20. 
Fig. 29 Photograph of the 19mm. embryo, H 173. X 2. Plane of section 


indicated. 
Fig. 30 Photograph of the 28 mm. embryo, H91. X 1}. Plane of section 


indicated. 
Fig. 31 Photograph of the 32 mm, embryo, H 41.  X 1}. Plane of section 


indicated. 


CONTRIBUTIONS FROM THE ZOOLOGICAL LABORATORY OF THE MUSEUM OF COMPARATIVE ZOOLOGY AT 
HARVARD COLLEGE, NO. 272 


THE MOVEMENTS IN THE VISUAL CELLS AND 
RETINAL PIGMENT OF THE LOWER 
VERTEBRATES 


LESLIE B. AREY 
Northwestern University Medical School 


ONE TEXT FIGURE AND FIVE PLATES 


CONTENTS 
ieanoroduction and DIsGOrical TOVIOW......... «os ds ieaees ewan esmesiresse.a. 121 
Ue RU HALE PUREE ITN son's scones sos < sexes ches ok a silnia wines SOO EMER SoM - » 124 
ee see TAMER TIPU. 5 ahs 2 x Se vce da nev Bee dviclaed ode da RAS ied ws 126 
im. ipetermimation Gf adaption times... .. . .. -ise..0<<ses ch anpaeiees vs» 126 
(e) (ROUND MeI, Foss bx, 0 aces de soeris eiee SAREE S foa. as 127 
Rea MNDISCIELE EDIRNE oe o.-25-5'a.ds Se tie ce G:b« ) we i wreak ee ae 129 
B. Effect of temperature (normal animals)........................0--- 133 
ee Da CaP NINN  BUMOIROIIG oO i dis wns LP We Bs hgh Cale Oe 6 Lk < SORES wechlN's 133 
(oop Noe ewe Ch 2 ee a: ee nore (6 Oe ae 151 
C. Experimentation upon excised eyes........ re es ee ae 163 
TAR eRIGOn OR irnt Nd GAYETOGR. ..o5'5655.o.chas'< 634 oe Ree en . 163 
ayn OeHOL PORDOTHWNNO Ci i. oo in tb icaed «fio vs Rh CER RAE es» 168 
PEO te CMRORGMOUINM Es 5), . bls S's ods bdo 0 4s and edn «eee br 171 
Fi RCE PLGETINESIG, < s\hc'c's Sa Sn Wein a Cice.s s «wv 22,00 iat « a ee oo 172 
MEN IAVASILRLCOLIS: th cles Se Ae Seto ke kas ook Cos ees 175 
Mi MTSCHE MOY ONE ke PE oll yah lwo peck v3) s nip 5 cJorstn cle aloe GRAM a Bede « Sle a 
I. Interrelation of integumentary photoreceptors and retinal elements. 179 
SRE UMN crs) ih CDi AW 50 Ate or k'G Go nye « on Dike aga! « MRM wigeule tanto eca| ck 181 
SR IRNRRRNMNERD 55 ats fre etc a cca Ld SEG hcg Fae ss a ao Ee els ¢ Siena Bee 184 
SRIRAM ee PROTA 2 edly isic bu sce RA ot Rs Melee rok ik tWimeaet 188 


1, INTRODUCTION AND HISTORICAL REVIEW 


The positional changes that occur in the vertebrate retina 
chiefly through the action of light have attracted the attention 
of many workers in continental Europe although, strangely 
enough, neither English nor American investigators have hither- 
to concerned themselves with this particular field of endeavor. 
The results acquired from these researches have proved of such 
interest and significance as to equal, and perhaps even to sur- 

121 


THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL. 26, NO. 2 
APRIL, 1916 


122 LESLIE B. AREY 


pass, those obtained from the two other branches of retinal physiol- 
ogy, namely, the study of the visual purple and of the action 
currents in the optic nerve. 

In order that the reader may understand the present status 
of this subject, it is desirable to summarize briefly the general 
conclusions that have been established relative to the efficiency 
of light in producing movements in the retinal elements. For 
a more extensive review of this literature reference may be made 
to a separate paper by the writer (Arey, 715) or to the excellent 
compilation of the German author, Garten (07). 

Although a variability in the position of the retinal pigment 
had been noted by early workers (H. Miller, 56; Morano, ’72) 
the cause remained unsuspected until Boll (78) and Kihne 
('77) independently discovered that in the light the retinal pig- 
ment of the frog extends nearly to the external limiting mem- 
brane whereas in darkness it retreats, thereby forming a com- 
pact layer next to the choroid! (figs. 1 and 4). Later obser- 
vations have corroborated Kihne’s original view that pigment 
migration is not due to the extension and retraction of cell proc- 
esses but to the movement of pigment granules in the proto- 
plasm of relatively fixed cells. 

The most extensive pigment migration occurs in fishes (Stort 
86) and in anuran amphibians (Boll, ’78, and Kiuhne ’77, on 
the frog; Arcoleo, ’90, on the toad), whereas the positional changes 
in the pigment of urodeles are relatively limited (Angelucci, ’78). 
Well defined movements of the retinal pigment are also found 
among birds, including not only those that are diurnal, but also 
some that are nocturnal in habit. Among reptiles (Angelucci, 
90) and mammals (Angelucci, 778) limited pigment changes 
have probably been detected, notwithstanding the contradictory 
evidence presented by various workers. 


1 Czerny (’67) found that when sunlight was concentrated by a lens on the 
retinas of various animals, the pigment became more highly expanded in the 
affected regions than in other portions which had been exposed to light but 
not so treated. In his experiments no comparisons were made with dark-adapted 
retinas. These somewhat pathological tests were not well substantiated by the 
later work of Deutschmann (’82). 


MOVEMENTS IN THE VISUAL CELLS 123 


Movements of the inner member of the vertebrate cone cell 
were first observed by Stort in 1884, although the earliest an- 
nouncement of this discovery was published by Englemann 
(85) in whose laboratory Stort worked. To the contractile 
portion of the cone’s inner member Englemann applied the sig- 
nificant term ‘myoid’ (figs. 30, 31; my. con.). The contractility 
of the myoid is extraordinary, since in some fishes light produces 
a shortening of this part to 10 per cent of the length which it 
assumes in darkness (figs. 25, 27). If effective at all, light 
always causes a shortening and darkness an elongation of the 
cone cell. 

Stort (’87) extended his first discovery on the frog by experi- 
mentation upon representatives of the various other vertebrate 
classes, thereby showing that in fishes and birds extensive move- 
ments of the cones likewise occur. In the salamander, as a type 
of urodele amphibian, responses of the cones to light have both 
been asserted (Angelucci, 90), and denied (Garten, 07). Among 
a few reptiles (Englemann, ’85) and mammals (Garten, ’07) 
slight changes have apparently been detected. 

The visible response of the rod’s inner member presumably 
is not identical throughout all the vertebrate classes. Angelucci 
(84) was the first to observe a shortening of the frog’s rod after 
exposure to light and later, in 1890, he applied the term ‘myoid’ 
to the contractile portion of the rod’s inner member (figs. 30, 
31; my. bac.) in a sense similar to that in which Englemann has 
previously used it for the corresponding portion of the cone; 
Arcoleo (’90) likewise reported that the rod myoid of the toad 
shortens in the light. Recently, however, Lederer (’08) has 
asserted that the photomechanical change in the rod myoid 
of the frog is not a shortening but an elongation. In all fishes 
that possess cones, the rod myoid lengthens in the light and 
shortens in darkness, as Stort (’87) first believed. The response 
in the rods of day-birds (Stort, ’86) is similar to that in fishes, 
although in night-birds (Garten, ’07) movements of these cells 
may be entirely absent. No experimentation has been per- 
formed upon the rod cells of mammals or of reptiles; in the latter 
group, however, the retina usually lacks these elements. 


124 LESLIE B. AREY 
2. MATERIAL AND TECHNIQUE 


The work of Chiarini (04a, ’06), in which he systematically 
investigated the effect of light on the retinas of representative 
vertebrates, led him to the conclusion that the maximal changes . 
in the position of the pigment and of the cone cells occur in the 
lower vertebrates and particularly in the fishes, whereas in the 
highest vertebrates these changes are extremely limited or may 
even be imperceptible. For experimentation involving varia- 
ble temperature poikilothermous animals are necessary, and no 
vertebrates are more easily available for this purpose than are 
fishes and amphibians. When to availability is added the ad- 
vantages which they offer by the possession of exceptionally 
well developed visual and pigment cells, which are capable of 
undergoing the greatest dimensional changes found in any of 
the vertebrates, it will be seen that these animals are particularly 
favorable for such physiological studies as were conducted in 
the present investigation. ; 

The following fishes were used: the common horned pout 
(Ameiurus nebulosus Lesueur); the common killifish (Fundulus 
heteroclitus Linn.); the shiner (Abramis crysoleucas Mitchill) ; 
and the goldfish (Carassius auratus Linn.). Of these, all but 
Fundulus are fresh water forms, while Fundulus is found in 
brackish water, and especially at the mouths of fresh water 
streams. Abramis and Carassius belong to the same family, 
the others to different families. Goldfish were obtained for 
the most part from dealers although a number of feral animals 
were used. Many of the streams and ponds about Cambridge 
are stocked with feral goldfish, liberated by accident or design. 
The majority of such animals have lost their gold coloration 
and have returned almost completely to the original olivaceous 
type. 

The amphibians used were as follows: the leopard frog (Rana 
pipiens Schreber); the green frog (Rana clamitans Latreille) ; 
the bull frog (Rana catesbiana Shaw); and the mud puppy 
(Necturus maculatus Raf.). Of the three species of frogs, the 
leopard frog was used almost exclusively for experimentation 
on adult animals; the larvae only of the bull frog were employed. 


MOVEMENTS IN THE VISUAL CELLS 125... 


The technique of preparing retinas for microscopical examina 
tion was very simple. Eyes were. removed from their orbits 
in two ways. In Ameiurus, where the eyes are prominent and 
the skin is soft, excision was performed directly. The eyes of 
amphibians and of fishes other than Ameiurus, particularly in 
experiments conducted in the dark where rapidity of operation 
was desirable, were not excised directly but according to the 
following procedure. With heavy scissors the cranium was 
bisected in a median sagittal plane; following this, a trans- 
verse cut just posterior to the orbit freed the two halves of the 
cranium, with the contained eyes, from the rest of the body. 
In either case the operation was performed in a few seconds and 
the eye, without being handled, was allowed to drop into the 
fixing fluid.” 

Both Perenyi’s and Kleinenberg’s fluids gave good fixation 
(Howard, ’08; Palmer, 712), but of the two, Perenyi was pre- 
ferred. The toughness of the sclera and the consequent slow- 
ness in the penetration of fluids demands generous allowances 
of time during the various steps preparatory to imbedding in 
paraffiine. Dehydration and clearing in xylol should, however, 
progress as rapidly as possible since otherwise the sclera becomes 
extremely hard. 

Two methods were used in removing the lens, one of which, 
although longer, gave much more satisfactory results. The first 
somewhat tedious procedure consisted in paring away the front 
face of the eyeball with a razor after the eye had been previously 
imbedded in paraffine. After removing the face of the eyeball 
slightly beyond the ora serrata, the lens was pried from its paraf- 
fine matrix with a dissecting needle; following such manipula- 
tion reimbedding was of course necessary. The second and ° 
simpler method was to remove the face of the eyeball with fine 
curved scissors after the eye had been hardened in absolute alco- 
hol; if, however, the eye was not sufficiently hardened or the 


2A method of simple decapitation used by Pergens has been criticized by 
various workers who contend that slight changes occur in the position of the 
retinal elements after the head is immersed in the killing fluid. In a series of 
controlled experiments, however, I could detect no post-mortem disturbances 
when the head was both bisected and cut from the trunk. 


126 LESLIE B. AREY 


greatest care was not exercised, the retina proper easily separated 
from the pigmented epithelium. On the whole, the first method 
was preferred to the second because of the wrinkling of the retina 
that usually accompanied the use of the latter. 

Sections were cut 7 » to 10 uw thick, and except in a few special 
cases only those passing through the region of the optic nerve 
were retained for examination. Preparations were stained with 
Ehrlich-Biondi’s triple stain or were double stained in Heiden- 
hain’s iron hematoxylin and a plasma counterstain. Ehrlich- 
Biondi in some instances gave excellent differentiation of all 
elements, while at other times it would show the capriciousness 
in producing satisfactory results for which it is notorious; iron 
hematoxylin gave uniformly good preparations. 

It often became necessary to bleach the pigment in order to 
study the visual cells, which would otherwise be masked by the 
partially or completely extended processes.. The method em- 
ployed was essentially that of Mayer (’81), in which nascent 
oxygen’ is the effective agent. 

The aim of the present investigation has been to determine 
the influence of light, temperature, anaesthetics and oxygen 
on the movements of the rods, cones, and retinal pigment in 
the normal and excised eyes of fishes and of amphibians. 

To Prof. G. H. Parker, under whose direction this research 
has been conducted, I wish to acknowledge my indebtedness for 
much inspiration and valuable suggestion. Advantage is taken 
of this opportunity to express appreciation for the facilities of 
the Zoélogical Laboratory placed at my disposal by the Director, 
Prof. E. L. Mark, and for many courtesies extended by him 

during my residence at Cambridge. 


3. EXPERIMENTAL PART 
A. DETERMINATION OF ADAPTION TIMES 


Before extensive experimentation can be undertaken on the 
retinal elements, it is necessary to determine the various lengths 
3 When potassium chlorate and hydrochloric acid interact, it is commonly 


said that nascent chlorine is the agent causing bleaching. As a matter of fact 
the reaction liberates free oxygen. 


MOVEMENTS IN THE VISUAL CELLS 127 


of the time which they require in assuming the positions charac- 
teristic of light and of darkness. In most cases it is not easy 
to state definitely when adaption is completed, for the response 
becomes less vigorous as it nears the end and consequently 
the factor of personal equation is unavoidable. 

Light intensity and temperature ‘(Dittler, ’07) represent 
variables which undoubtedly play a part in the determination 
of adaption time. No attempt was made to discover the exact 
role of either of these factors, although the general experimental 
conditions were kept approximately uniform during successive 
trials. 

The effect upon adaption time of a long or short preliminary 
subjection to light or darkness, has never been taken into ac- 
count, although such influences were asserted by Gaglio (’94). 


a. Retinal pigment 


Pergens (96) found that after 2 minutes’ illumination the 
retinal pigment of Leuciscus began to expand. After 1 minute 
of darkness a noticeable contraction occurred, which was com- 
pleted in 20 minutes. Chiarini (’04b), working on the same 
fish, came to somewhat different conclusions. He observed 
a sensible pigment expansion after direct exposure to sunlight 
for 1 minute, although complete light adaption necessitated 
a period of 1 hour. The reverse process of dark adaption was 
not initiated until the animal had been subjected to darkness 
for from 4 to 5 minutes, and a minimum of 1 hour was required 
to complete the contraction. 

When the retinal pigment of fishes has undergone a maximum 
expansion, it accumulates distally’ near the external limiting 
membrane, whereas the proximal portions of the cells are to a 
greater or less extent devoid of pigment (figs. 9, 10). Such a 


4The term ‘proximal’ as used in this paper will refer to movements toward 
the nuclei of a given cell, either pigment or visual. In like manner ‘distal’ will 
have reference to movements away from the nucleus. Since the distal movement 
in the pigment cells is the reverse of that in the rods and cones, an arbitrary 
nomenclature with reference to the eye-ball becomes confusing, while the ter- 
minology here suggested lends itself readily to descriptions of the moving parts. 


128 LESLIE B. AREY 


condition of distal accumulation requires additional time after 
the ‘front rank’ of pigment granules has arrived at the position 
of maximal extension. In the determinations made by me, 
light adaption was not considered to be completed until the 
pigment was thus maximally expanded. 

The results obtained on the retinal pigment of fishes were 
as follows: 


Ameiurus 

Diffuse daylight 
30 minutes, pigment incompletely extended 
40 minutes, pigment fully extended, but homogeneously distributed 
1 hour, maximal expansion 

Darkness 
45 minutes, pigment about half contracted 

1 hour, maximal contraction 


The movements of Abramis were more rapid, notwithstand- 
ing an apparently heavier pigmentation of the retina. 


Abramis 

Diffuse daylight 

15 minutes, lightly pigmented processes three quarters extended. Most of 
the pigment at the bases of the cells 
30 minutes, processes fully extended. Distal pigment accumulation 

begun 

45 minutes, maximal expansion 

Darkness 
20 minutes, pigment about half contracted 
30 minutes, maximal pigment contraction 


The results of the dark adaption on Fundulus were less satis- 
factory, since the delimitation of the pigment in darkness is 
poorly defined. 


Fundulus 
Diffuse daylight : 
45 minutes, pigment fully extended; but little tendency exhibited towards 
distal accumulation 
1 hour, maximal expansion 
Darkness 
45 minutes to 1 hour, maximal contraction 


As regards the adaption time in the frog, Stort (’87) believed 
that in 1 hour of bright, diffuse daylight or in 4 hours of dark- 


MOVEMENTS IN THE VISUAL CELLS 129 


ness, maximal light- and dark-adaption of the pigment respec- 
tively was produced. Kiihne (’79) stated that complete dark- 
adaption occurred in 1 to 2 hours, this period corresponding 
to the time necessary for the regeneration of the visual purple. 
Chiarini (’04b) maintained that following an exposure of half 
an hour to direct sunlight one and one-half hours were needed 
to complete the adaption in darkness. Although no direct 
experimentation was made on the frog, the general experience 
gained from working with this animal leads me to suspect that 
1 hour for light-adaption and 1 to 2 hours for dark-adaption 
are approximately the proper amounts. 


b. Visual cells 


Relative to the adaption times of the cone cells, Pergens 
(’96) stated that in Leuciscus the first visible shortening occurred 
after-an exposure to light of 1 minute. According to his illus- 
trations, the cones were very much shortened after 2 minutes 
while those that had been subjected to light for 5 minutes were 
practically in the position characteristic for ight. When light- 
adapted animals were introduced into the dark, a lengthening 
of the cones was evident after 1 minute and in 5 minutes the 
elongation was nearly complete, although it did not become 
maximal for 20 minutes. Pergens’ method of allowing decapi- 
tated heads to remain in the light during fixation has been criti- 
cized by Garten (’07), and the later work of Pergens (’99) has 
likewise been questioned by Herzog (’05). These workers assert 
that it is entirely possible that the changes initiated by the action 
of light continue until the actual penetration of the fluid into 
the eye fixes the retina. Certain experiments of Weiss, insti- 
gated by Garten, indicate that the action of light can influence 
the position of cones and retinal pigment after decapitated heads 
have been introduced into the fixitive. These results are in 
opposition to the statement of Chiarini (’04b) that light can- 
not be effective during fixation. 

In an attempt to avoid this source of error, at the compietion 
of an experiment the light-adapted eyes were fixed in exceedingly 


130 LESLIE B. AREY 


dim light. It is probable that this precaution was sufficient, for, 
as will be shown later, the movements of the cones of Abramis, 
the only fish worked upon, are not rapid and moreover no changes 
oecur even when the excised eyes, immersed in water, are sub- 
jected to light or to darkness. 
’ Temperature is an important factor that must be considered 
in the adaption of cones. In anticipation of certain results 
that will be found in another part of this paper, it may be said 
that the cones of fishes are maximally extended at about 25°C. 
in the dark (fig. 27), and in the case of Abramis at least, they are 
also maximally shortened at 5°C. in the dark (fig. 25). Moreover, 
as temperature does not affect the length of the cones when they 
are under the influence of light, the animals may be kept at 
a temperature of 25°C. during an entire experiment and the re- 
sulting movement of the cones will then be solely traceable to 
conditions of light or darkness. 
The results on the cones of Abramis may be summarized 
as follows: 
Diffuse daylight 

15 minutes, cones much shortened—perhaps two thirds 

23 minutes, approximately the same condition as at 15 minutes 

30 minutes, shortening not quite complete in most animals 

45 minutes, maximal light adaption 

Darkness 
13 minutes, cones somewhat extended—one third (?) 


20 minutes, extension practically complete 
30 minutes, maximal dark adaption 


The adaption times of the cones of Abramis are longer than 
those given by Stort for Leuciscus. This, in part, may be due 
to the wider range between the positions of maximal light- and 
dark-adaption which was produced by the aid of elevated 
temperature. 

Englemann (’85), working on the frog, was the first to discover 
that the movements of the cones were not accomplished in- 
stantaneously, but required definite periods of time. He also 
observed that elongation in the dark was a longer process than 
shortening in the light. My results on Abramis do not entirely 
support his latter view. At first the cones of this animal do 


MOVEMENTS IN THE VISUAL CELLS 131 


respond more actively when stimulated by light, but a longer 
time is required to complete the process of light-adaption than 
the reverse changes in the dark. 

No experimentation seems to have been performed upon any 
animal to determine the adaption time of the rod cells. Ameiurus 
was selected for these tests, for the rods differ greatly from those 
which are characteristic of fishes in general. Instead of the 
slender elements 1.5 » to 2.0 « in diameter, which, for example, are 
found in Abramis (fig. 25), the rods in Ameiurus (fig. 31) are 
robust and resemble more closely those of the frog (fig. 35). 
The barrel-shaped ellipsoid measures about 4 u in either dimen- 
sion, and the width of ‘the outer membrane is the same. With 
this can be compared the width of the outer member of the rods 
in the frog, which in my preparations of R. pipiens measured, 
for the most part, 5 u, although Howard (’08) states the width 
as 6 », and H. Miller (56) as from 6 uw to 7 u.* The species 
was not mentioned by either of these writers. 

Unlike the cones, the rods of Ameiurus in the dark form a 
more or less even row close to the external limiting membrane 
(fig. 31), while in the light the myoid elongates carrying the 
ellipsoid and outer member far up‘into the pigment layer (fig. 
30). The extent of these positional changes may be judged from 
measurements of rods in darkness and in light which give ex- 
treme values of 70 u and 7 u respectively for the length of the 
myoid. I know of no fish in which rods of this size have been 
described, although Garten (’07) makes particular mention of 
the pike as possessing ‘grosse Stibchen.’ It is evident that 
the large size and the extensive positional changes which the 
rods of Ameiurus undergo make them especially favorable for 
physiological experimentation. 

The effect of temperature upon the length of the rods is com- 
paratively slight, hence the following determinations on the 
rods of Ameiurus were conducted at room temperature. 


5 Perhaps the fact that my measurements were made on dark-adapted rods 
accounts for this discrepancy, for in rod cells the outer member is said to become 
longer and slenderer in darkness than in light. 


132 LESLIE B. AREY 


Diffused daylight 
30 minutes, rods two-thirds extended 
45 minutes, maximal light adaption. (Cones also light-adapted) 
Darkness 
15 to 20 minutes, rods, in most cases, almost completely shortened 
30 minutes, maximal dark adaption. (Cones still in position of light 


adaption) 


The quicker response of the rods in darkness than in light is 
noteworthy. The rod shortens in the dark, the cone in the light; 
since in both cases the process of shortening is more vigorous 
than the lengthening, it would appear that the contractility of 
either type of cell is the responsible factor and the relative effici- 
ency of light and darkness is not primarily involved. In the last 
analysis, however, the situation may not be reducible to such 
simple terms. If these responses are merely the expression of the 
action of light and darkness on the protoplasmic myoids, why 
should the direction of movement of the two elements be opposed? 
A discussion of this phase of the problem will be found in another 
place. 

I attempted no experimentation upon the cones of the frog. 
Angelucci (’90) stated that after an exposure to candle-light 
for 5 minutes, the cones weré strongly retracted, although other 
experiments of his do not seem to support this conclusion. Her- 
zog (’05) found that at medium light intensity complete light- 
adaption occurred in 24 minutes. 

The most surprising discovery in this series of determina- 
tions, taken as a whole, was the length of time required to com- 
plete the adaption of the rod and cone cells in comparison with 
the retinal pigment. From the results of earlier workers, I 
had expected the positional changes of these cells to be com- 
pleted in a few minutes, hence the actual values obtained were 
wholly unlooked for, and were only accepted after many repe- 
titions of individual experiments. 


MOVEMENTS IN THE VISUAL CELLS 133 


B. EFFECT OF TEMPERATURE (NORMAL ANIMALS) 
a. Retinal pigment 


No investigations have hitherto been made to determine the 
effect of temperature on the retinal pigment of fishes, whereas 
several workers have used the frog for experimentation of this 
kind. The problem considered here was to determine the re- 
sponse of the retinal pigment of normal fishes and amphibians to 
various temperatures, both in light and in darkness. 

1. Fishes. Experiments in the light were performed in the 
following way. Light-adapted® fish were placed in large battery 
jars close to north windows where they received strong diffuse 
daylight; sheets of white paper were always placed under the 
jars to aid reflection.? The highest temperature to which 
it is safe to subject fishes is about 28°C., although by gradual 
elevation a somewhat higher temperature can be withstood 
(Loeb and Wasteneys, ’12). A low temperature that did not 
vary beyond the limits of 3° and 5°C. was produced by intro- 
ducing small pieces of ice directly into jars with the fish. At this 
temperature fish are for the most part inactive, the respiration 
rate decreases and they remain quietly at the bottom of the jars. 

At the end of an experiment, which was never less than three 
hours long, the eyes were excised and immersed in fixing fluid 
in the light.® 

During the earliest trials retinas from the same fish were 
compared, eyes being subjected to the extreme temperatures in 


6 The terms ‘light adaption’ and ‘dark adaption’ as used throughout this paper 
imply that the animals had been previously subjected for a minimum of 4 hours 
either to bright diffuse daylight or to total darkness. 

7 The possibility of a dark background influencing the distribution of retinal 
pigment was considered. Such a visual control, if present, would correspond 
to the known réle of the eye, as determined by Pouchet (’76) and others, in ani- 
mals which adapt their body color to the immediate environment. A series of 
careful comparisons, however, failed to show any recognizable differences be- 
tween the pigment distribution in the retinas of fishes that had been kept over 
dark or light backgrounds. 

§ In all work which involved the use of temperature, the fixing fluid was kept 
at the same temperature as that at which the experiment had been performed. 
Such treatment eliminated a possible source of slight error. 


134 LESLIE B. ARBY 


successive experiments. At first this seemed to be the correct 
procedure, but rigorous controls showed that such precautions 
were unnecessary. When interpreting the results of experiments 
conducted in the light it is not the absolute amount, but rather 
the relative distribution of pigment that serves as a basis for 
decisions. In experimentation in the dark, absolute differences 
in the degrees of pigmentation could give rise to errors in judg- 
ment, for the pigment, gathered into compact masses in the cell 
‘cups’ might mask or apparently reverse the effect of temperature. 
Fortunately, however, the retinas, as judged from the expanded 
light condition seem, on the whole, to be very equally pigmented 
and the width of the contracted pigment zone, therefore, gives 
a fair index of the effect of temperature in the dark. When work- 
ing in the light especially, it was found to be very desirable to 
have the experiments at contrasted temperatures conducted 
simultaneously in order that advantage might be taken of iden- 
tical light conditions, for as will be shown, light intensity is an 
important factor in obtaining the maximum expansion of pigment. 

Experimentation in the dark was conducted as follows. A 
fireless cooker, lined with black paper, was used as a dark cham- 
ber, on account of the minimal loss of heat incurred by it during 
the course of an experiment. If such an apparatus be previously 
brought to the temperature of the introduced jar of water, an 
experiment can be continued for several hours without further 
attention. In all determinations mentioned in this paper which 
were conducted in the dark, precautions were taken against the 
possible influence of light during the few seconds necessary for 
excision and transference of eyes to the fixative. A long series 
of careful comparisons showed identical results whether the 
operation was carried out in total darkness or in light of just 
sufficient strength to permit the operator to see the animal 
and his instruments. Indeed, the results obtained by operating 
in an ordinarily lighted room showed no recognizable differ- 
ences in either pigment, rods, or cones from those secured by 
working in darkness. 

A detailed description follows of the conditions found in each 
of the four fishes: 


MOVEMENTS IN THE VISUAL CELLS 135 


1. Ameiurus. At 25°C. in the light (fig. 2), the characteristic 
position of the expanded pigment is in a broad band about 95 u 
wide, which extends nearly to the external limiting membrane. 
The pigment granules are evenly distributed and show no ten- 
dency to aggregate distally. 

At 15°C. the condition is very similar to that just described. 
The pigment, on the whole, tends to be homogeneously dis- 
tributed, although in many retinas at this temperature there is 
a slight distal accumulation. 

The disposition of pigment at 5°C. is markedly different 
(fig. 1) for it migrates to an extreme d'stal situation and forms 
a dense zone, approximately 30 » wide, close to the external 
limiting membrane, although the pigment of fishes and amphib- 
ians, under normal conditions, never actually touches this 
membrane. Between this heavy pigment-mass and the bases 
of the cells lie scattered granules, but the intervening space, 
nevertheless, appears relatively devoid of pigment. This ex- 
treme condition is best produced on the brightest days, and it is 
impossible to obtain as complete a migration on cloudy days, 
regardless of the temperature. On the other hand, the uniform 
distribution characteristic of incomplete expansion at 25°C., 
is independent of the intensity of diffuse daylight. This would 
suggest that a high temperature is more efficient than light in 
the regulation of pigment distribution, and that cold, that is, 
the absence of heat, merely allowed light to act unrestrained. 
Light and high temperature, then, are antagonistic in their effects. 

The results at 5°, 15°, and 25°C. are, in a way, what might 
have been expected. A temperature of 15° to 25°C. probably 
represents the greatest average warmth to which the animal is 
subjected in nature; this range from 15° to 25°C., then, repre- 
sents the limits of what may be called a warm environment for 
the animal. In the same way from 0° to 10°C. may be called 
a cold environment, and from 10° to 15°C. a neutral environment, 
neither particularly warm nor cold. Hence it is not surprising 
that the results at 15°C. are more similar to those at 25°C. than 
at 5°C. At 10°C. the distribution of pigment approximates 
rather more closely that at 5°C. A curve, therefore, obtained 


136 LESLIE B. AREY 


by plotting temperatures as abscissas and the quantitative 
amount of distal migration as ordinates would show a gradual 
slope from 0° to 10°C., from 10° to 15°C. a rapid drop and from 
15° to 25°C. a nearly horizontal but slightly sloping line. 

The results of experiments performed in the dark, where the 
pigment is highly contracted, are usually not as clear cut as 
those just described. The reason for this is because it is im- 
possible to see the qualitative distribution of the pigment granules 
and hence decisions regarding the effect of temperature must 
depend largely on measurements of the width of the narrow 
pigmented layer. Such a criterion, as has previously been 
pointed out, is open to the criticism that individual eyes may 
vary enough in the absolute amount of contained pigment to 
disconcert judgments concerning the effect of temperature. 
After having studied a great number of preparations, I do not 
believe that such an unequal pigmentation is in truth a factor 
that warrants serious consideration. However this may be, 
an obvious precaution consists in the prolonged repetition of 
each type of experiment. It may be said that in the course of 
my experimentation on the effect of temperature on normal 
fishes alone, over 200 retinas have been examined. 

The evidence obtained from Ameiurus, was more conclusive 
than that from any of the other fishes, with the possible excep- 
tion of Carassius. At 25°C. (fig. 4) the pigment forms a densely 
contracted layer, the mean width of which is about 25 yw. In 
contrast with this is the condition at 5°C. (fig. 3), where the cells 
have short pigmented processes, the total extent of which is 
approximately 38 y». The differences apparent at these two 
extremes of temperature were so slight in comparison with the 
much greater variation in the light, that thorough experimenta- 
tion at the intermediate temperature of. 15°C. was not attempted, 
although it was sufficiently demonstrated that the results at 
this temperature do not vary to any great extent from those at 
the two extremes, and probably more closely approximate the 
highly contracted condition at 25°C. 

The results of some of the temperature determinations were not 
conclusive. Reference has already been made to the fact that 


MOVEMENTS IN THE VISUAL CELLS 137 


on dull days maximal expansion in the light was hard to obtain. 
Among fishes in general, more doubtful cases occurred in ex- 
periments conducted in the dark than in the light, yet in all 
such cases the uncertainty merely lay in deciding between two 
nearly equal conditions, while in practically no instance was there 
evidence of a definite reversal whereby a greater distal migra- 
tion occurred at 25°C. than at 5°C. 

2. Fundulus. The conclusions reached from the study of 
Fundulus, as well as from the other fishes, are similar to those 
given for Ameiurus, but each fish shows individual peculiarities 
in the disposition of the pigment and these will be briefly 
described. 

In the light the pigment of Fundulus tends to migrate to a 
great extent forming a broad zone at the distal ends of the cells, 
much denser and more sharply defined than in Ameiurus. Be- 
tween this zone, which has a width of 30 u, and the bases of 
the cells there is a clearer area, almost devoid of pigment at 
5°C. (fig. 5), while at 25°C. (fig. 6) this region contains a con- 
siderable amount of evenly distributed pigment granules. In 
the latter case, however, the pigment still forms a closely aggre- 
gated zone distally, although it is reduced to a width of 17 ux. 
As in Ameiurus the condition at 15°C. more closely resembles 
that at 25°C. than at 5°C. 

A peculiarity in preparations of the light-adapted retina of 
Fundulus (at least with Perenyi’s fixation) is that at the higher 
temperatures the pigment extends in columns from base to 
periphery of individual cells (fig. 6), while between such columns 
of adjacent cells are elongated areas free of pigment but taking 
the plasma stain. Examination under high magnification does 
not show the presence of excessive shrinkage, although a casual 
observation might suggest that this had occurred. Where the 
pigment is aggregated at the base and periphery of the cells, 
cell boundaries are not distinguishable and the pigment appears 
as homogeneous masses. The separate columns connecting these 
two continuous zones give the whole an appearance not unlike 
a ladder with rungs set very close together. 


THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL. 26, NO. 2 


138 LESLIE B. AREY 


In the dark the contraction is very incomplete and some- 
times tends to give rise to a distal accumulation of pigment 
which faintly resembles the distribution in the light, although 
such a condition is not constantly present as a characteristic 
of dark adaption. It is impossible to state the cause of this 
pigment massing. It may be due to a greater activity at the 
distal ends of the cells in producing a contraction, or what would 
bring about a similar end result, a contraction of pigment en 
masse. 

Measurements were taken of preparations at the extreme 
temperatures in the dark, the degree of variation being greater 
than in any of the other fishes studied. Thirteen retinas at 
5°C. (fig. 7) showed the pigment to be extended a mean dis- 
tance of 20.4 divisions of the ocular micrometer, whereas fif- 
teen retinas at 25°C. (fig. 8) had a corresponding value of 14.4. 

3. Abramis. In the light the condition in the eye of this 
fish is somewhat similar to that in Fundulus. The pigmenta- 
tion is very heavy, forming a broad zone near the external limit- 
ing membrane; between this distal zone and the base of the 
pigment cells pigment granules are also present, the density of 
pigmentation depending on the temperature. 

At 5°C. (fig. 9) the distal zone is wider than at 25°C. (fig. 10) 
in the ratio of 50 uw to 38 u, while the proximal area is much less 
heavily pigmented than at 25°C. Although each area is sharply 
defined, the one appears to grow at the expense of the other. 
It was not possible to get so complete a distal accumulation 
of the pigment as in Fundulus and, as will be presently shown, 
the extent of migration in Fundulus is still less complete than 
that in Carassius. 

The dark phase is one of great contraction, and although 
some retinas show evident differences at the extreme tempera- 
tures, yet throughout the whole set judgment of the eye has 
to be supplemented by actual measurements. Such measure- 
ments show that the mean expansion at 5°C. (fig. 11) exceeds 
that at 25°C. (fig. 12), the values being 30 u and 20 » respectively. 

4. Carassius. The eye of the goldfish is more heavily pig- 
mented than those of the three other fishes. At 5°C. in the 


MOVEMENTS IN THE VISUAL CELLS 139 


light (fig. 13) all the pigment is located distally forming a broad, 
dense zone. Between this band and the base of the cells is a 
narrow clear area in which scattered pigment granules are visible 
only with the aid of high magnifications—for practical purposes 
it may be said to be free of pigment. 

The situation at 25°C. in the light (fig. 14) is variable. In some 
cases it closely simulates that at 5°C., although the clearer 
space is always relatively more heavily pigmented; in other 
instances, the pigment is uniformly distributed from the proxi- 
mal to the distal. extent of the cells. 

The contraction that usually occurs in the dark was less pro- 
nounced in Carassius than in any fish heretofore deseribed; in 
fact in some eyes that had been subjected to a low temperature, 
the actual breadth of the pigment layer nearly equalled that of 
a light-adapted eye. The relative expansion at the extreme 
temperatures in the dark, however; leaves no doubt that at 
5°C. (fig. 15) the contraction is less than at 25°C. (fig. 16). 
Measurements (in terms of the divisions of an ocular micrometer) 
taken from eight eyes at 25°C. and ten eyes at 5°C. gave re- 
spective mean values of 8.0 and 12.1. 

From observations on these four genera certain generaliza- 
tions are suggested. The degree of pigmentation in the eye of 
Ameiurus (figs. 1, 2) is much less extensive than in the other 
fishes, as a comparison with Carassius (figs. 18, 14) shows in 
a striking manner. The three other fishes, however, offer better 
opportunities for comparison since their pigmentation is more 
nearly equal. At 5°C. in the light the pigment of both Carassius 
and Fundulus (figs. 18, 5) migrates distally to such an extent 
that a proximal zone, devoid of pigment, is created. Moreover, 
the dark phase of the pigment, in both these animals (figs. 16, 8), 
is one by no means extreme when compared with the highly 
compact layer in Abramis and Ameiurus (figs. 12, 4). In the 
last named fishes, on the contrary, it is impossible under the most 
favorable conditions of light and temperature to obtain the 
proximal clearer area entirely free from pigment; correlated 
with this absence of complete expansion is the high degree of 
contraction which is evoked in the dark. 


140 LESLIE B. AREY 


It does not seem probable that quantitative differences in the 
degree of pigmentation can be the cause of such relations, for if 
the relative amounts and the distribution of pigment in Fundulus 
and Abramis be compared at 5°C. in the light (figs. 5, 1) and at 
either temperature in the dark (figs. 7, 3) it must be admitted 
that quantitative differences do not adequately explain the 
conditions that exist. 

A clearer insight is gained if these responses are viewed from 
the standpoint of cell organization. We can think of two gen- 
eral types of pigment cells in which the pigment distribution is 
correlated with the behavior outlined above. Thus, in one type, . 
the pigment would tend to remain at the distal end; such a 
cell would show maximal expansion but relatively incomplete 
contraction. In a second type of cell, in which the pigment 
aggregates more proximally, maximal contraction but incom- 
plete expansion would be accomplished. Although other species 
of fishes should be studied before a final conclusion is reached, 
this set of relations may be general. The correlation can be 
stated, at least tentatively, as follows—the highest degrees of 
expansion and contraction in the retinal pigment of fishes are 
mutually exclusive in the same retina. 

2. Frog (adult and larva). Asin many other lines of physiologi- 
cal work, the frog has been used to a large extent by investigators 
of retinal physiology. 

Ewald und Kihne (’78) performed the first experiment in 
which the position of the retinal pigment of the frog was shown 
to be dependent upon temperature. According to their account, 
after 2 hours’ immersion in ice water in the dark, the pigment 
showed a distribution similar to that obtained at 17°C., which 
may be called a state of contraction.’ When, however, frogs 
were subjected to 30°C. for 2 hours, an expansion occurred in 
which lightly pigmented processes were said to extend even to 


9 This, however, does not coincide with an earlier statement (’77, p. 250) of 
the same authors, ‘‘Vor allem ist die Temperatur von ausserordenlichem Ein- 
flusse . . . . Frésche, welche 1-2 Stunden in Eiswasser gehalten wurden, 
liefern schwarze Netzhiute, indem das ganze Epithel mit ausschliipft, und 
nicht viel besser verhalten sich die Priparate von solchen, die bei 5°-10°C. im 
Dunkeln verweilten.”’ 


MOVEMENTS IN THE VISUAL CELLS 141 


the external limiting membrane. These workers were primarily 
interested in discovering under what conditions of temperature 
an accumulation of pigment in the rod area could be avoided, 
for their investigation dealt with the visual purple. Since the 
frogs used for these determinations were treated with curare 
in order to produce pigment relaxation, and, moreover, since 
these animals were in an oedematous condition as a result of this 
poisoning, it is evident, as Herzog (’05) pointed out, that much 
weight can not be given to their conclusions alone. 

To Gradenigro (’85) belongs the credit of having performed 
the first temperature experiment upon the retinal pigment of 
normal animals. He introduced a dark-adapted frog into a 
dry, darkened chamber and removed the whole to a dark room. 
A temperature of 30°C. was maintained until heat rigor set in, 
when, on examination, the pigment was found to be in a condi- 
- tion of maximal light expansion (fig. 19). Gradenigro’s results 
were confirmed by Angelucci (’90) and by Fujita (’11). 

Herzog (’05), without knowledge of Gradenigro’s contribution, 
undertook a detailed study of the relation between temperature 
and pigment distribution. His results not only corroborated 
those of Gradenigro, but also established the additional facts 
that at low temperatures (0°-14°C.) in the dark the distribu- 
tion of pigment is identical with that at high temperatures 
(fig. 17), while only between 14° and 18°C. in the dark (fig. 18) 
is maximal pigment contraction obtained. 

His experiments were performed in the; following manner. 
Dark-adapted frogs (R. temporaria and R. esculenta?) were 
placed in a heating chamber at an initial temperature of 20°C., 
the introduction of the animals cooling the apparatus to 17°-18°C. 
Progressive heating raised the temperature in 15 minutes to 
24°C., in 30 minutes to 32°C., in 45 minutes to 37°C., and in one 
hour to 39-40°C. At each fifteen minute interval frogs were 
removed and their eyes prepared for microscopical examination. 

At 24°C. the position of the pigment was nat essentially dif- 
ferent from the normal state of maximal contraction, although 
delicate fringed processes did extend towards the external lim- 
iting membrane. It is probable that such an experiment did 


142 LESLIE B. AREY 


not fairly test the effect of this temperature on pigment migration. 
One could hardly expect the body temperature of the animals 
to become adjusted to that of the apparatus in such a short time, 
especially since the heating was progressive and the final tem- 
perature was not realized until the end of the experiment. 

At 32°C. the pigment had extended to the maximal distance 
but the distribution was nearly homogeneous. 

A condition of extreme expansion occurred at 37°C. <A dense 
massing of pigment near the external limiting membrane masked 
the rod ellipsoids completely, while the outer members of the 
rods were nearly free from pigment. When the temperature 
was raised to 39° or 40°C. clonic spasms occurred which ended 
in death; the pigment, nevertheless, retained the same position 
as at 37°C, 

A second series of experiments, in which frogs were cooled 
in a refrigerating apparatus, showed that a subjection to 0°C. 
in the dark for 2 hours produced incomplete expansion, while 
after 3 hours the pigment was distributed in a zone of maximum 
breadth, but with only a slight tendency toward distal massing. 
This discovery, which was Herzog’s most interesting contribu- 
tion, is not only in disagreement with the commonly quoted 
result of Ewald und Kiihne’s earlier work, but also has not been 
substantiated by the recent investigation of Fujita (11), who, 
however, states that high temperature does induce pigment 
expansion, as the:other investigators have all maintained. Fujita 
tried the effect of fow temperature on only four animals, the 
duration of his experiments ranging from 30 minutes to 6 hours, 
yet he drew the following positive conclusion (p. 170): “Das 
Resultat war in allen Fallen das gleiche: ich konnte keine Hell- 
stellung konstatieren. Die Zapfen waren nicht kontrahiert, das 
Pigment nicht vorgewandert.”’ 

Since all these results on the frog’s retinal pigment are not 
only fundamentally different from those found by me in fishes, 
but also have no parallel in the movements of vertebrate and 
invertebrate melanophores, and since, moreover, there is no 
general agreement concerning the effect of low temperature, a 
thorough reinvestigation of the problem seemed to be needed. 


MOVEMENTS IN THE VISUAL CELLS 143 


My first effort, therefore, was to repeat Herzog’s work using 
apparatus and methods that essentially agreed with his, in 
order to ascertain whether identical results would be realized. 
If the movements of the frog’s retinal pigment are really ex- 
ceptional among other lower vertebrates, such responses have 
considerable theoretical as well as incidental interest. 

The following apparatus was devised. A cubical cage, made 
of fine wire netting with a cover of the same material, was sup- 
ported by uprights inside a large battery jar which was fitted 
with a glass cover perforated by a small hole to allow slow dif- 
fusion of air. The wire cage, made to receive the frog, did not 
come in contact with any part of the surrounding glass receiver 
which was to serve as a constant temperature chamber. The 
glass receiver sat upon a platform in a large cylindrical metal 
tank. A thermometer passed through the cover of the tank 
and also through the cover of the temperature chamber into 
that chamber itself. A funnel connected with a rubber tube 
to exclude light was fitted into the cover of the tank and served 
for introducing water into the tank. 

Tests were conducted in the following way. A dark-adapted 
frog was wiped dry and placed within the wire cage! inside the 
constant temperature chamber. Previously the chamber had 
been brought to the desired temperature by one of two simple 
methods. If the effect of heating was to be studied, the metal 
tank was partially filled with water at the appropriate temperature 
and the whole system allowed to adjust itself until the ther- 
mometer insde the constant temperature chamber registered 
the required degree of warmth. After the frog had been in- 
troduced, the temperature could be regulated without opening 
the apparatus by admitting warmer water through the funnel 
and drawing off an equivalent amount from the bottom of the 
tank. If, on the other hand, a temperature near the freezing 

10 Herzog lays much stress on the facts that the animal's body was never in 
contact with solids other than the wires of the cage, and that the body was always 
wiped free from secretions or excretions at the commencement of an experiment. 
He seems to fear lest there should be a chemical stimulation due to heated secre- 


tions or excretions that would affect the results in case these precautions were 
not followed. 


144 LESLIE B. AREY 


point was desired, the outer chamber of the tank was filled with 
a mixture of ice and water; if a sufficient amount of ice was 
provided a temperature of from 3° to 5°C. could be maintained, 
without further attention, throughout the whole experiment. 
At the expiration of the time allowed for temperature adaption 
(usually 3 hours), the apparatus was placed in the dark or in 
weak red light and the frog’s eyes were removed and fixed in 
darkness at the same temperature as that at which the experiment 
had been conducted. . 

In the course of these tests nearly 100 eyes were examined, 
yet the general results obtained were identical with those de- 
scribed by Herzog. At 3°C. and 33°C. (figs. 17, 19) the pig- 
ment was expanded:approximating the condition characteristic of 
light; between the temperatures of 14°C. and 19°C. (fig. 18), 
however, the pigment was contracted to a narrow compact layer. 
It should be noted that 14°C. and 19°C. may not represent the 
limiting temperature at which pigment contraction occurs, 
although Herzog states this to be the case; no attempt was 
made by me to determine these intermediate temperature-limits. 
At low temperatures the expansion of pigment was generally 
not as complete as at a high temperature or in the light, and there 
was often considerable variation in different parts of the same 
retina, yet the general result was one of unmistakable expansion. 

In some cases, however, there was little or no evidence of a 
pigment expansion at the lower temperature, yet such instances 
were comparatively rare. Although the cause of discrepant 
results of this kind is not evident, they perhaps furnish additional 
proof for the nervous control of the frog’s retinal pigment, as 
many workers maintain (vide infra). It seems probable that 
Fujita was unfortunate enough, in the few experiments which 
he performed at a low temperature, to obtain nothing but this 
lack of typical results, although in my own work the occurrence 
~ of such anomalous cases was always sporadic. 

A careful comparison was made of the results obtained at 3°C. 
and 33°C. No constant difference in the amount of migration 
could be detected, although Herzog states that at 37°C. the 
distal migration is greater than that obtained at low temperature. 


MOVEMENTS IN THE VISUAL CELLS 145 


Hence we may conclude that the condition found in the frog 
is unlike that found in the fishes. It should be noted, however, 
that between certain limits the two animals show similar ten- 
dencies in their pigment responses. These limits are approxi- 
mately 0° to 19°C. in the dark for the frog, and 0° to 28°C. in 
either darkness or light, for the fishes. 

Since the tendency of pigment migration under the influence 
of temperature agrees between the limits of 0° to 19°C. for the 
dark adapted frog and 0° to 28°C. for the fishes either in light 
or in darkness, the query may be raised—is it not possible that 
if the fishes were subjected to higher temperatures a reversal 
of the temperature effect would occur in which an expansion of 
the pigment would again be found as in the frog? The fact that 
Herzog found but slight differences at 18°C. and 24°C. increases 
this suspicion. I am convinced, however, that a tendency to- 
ward such a response does not exist in the retinal pigment of 
fishes even to the slightest extent. In the first place, as stated 
before, prolonged heating of the frog’s retina at 24°C. would 
be likely to produce more striking changes than Herzog obtained. 
Furthermore, a few experiments in which the temperature of 
fishes was raised to 30°C. and over failed to show, in either 
light or darkness, anything beyond the characteristic response 
of less complete expansion than at the lower temperature. 

The extent to which the retinal pigment of the frog and of 
fishes moves under the influence of temperature differs in a high 
degree. Among the fishes the differences are small and amount 
to little more than a redistribution of pigment in the dark and 
light phases respectively, whereas in the dark-adapted frog 
varying temperature induces the whole range of pigment re- 
sponse usually occasioned by light or darkness. This further 
indicates that the nature of the response in the two kinds of pig- 
ment cells differs fundamentally. In the fishes probably the 
response is through the direct action of temperature on the cell 
protoplasm, while in the frog the pigment migration may be 
produced indirectly through the intervention of the nervous 
system. 


146 LESLIE B. AREY 


All the experiments of previous workers on the effect of tem- 
perature have been performed in the dark and I, therefore, set 
about to discover what results would be obtained in the light. 
The only change in the apparatus from that previously described 
was that a 20 litre jar replaced the metal tank; this, when filled 
with water at the desired temperature, performed the necessary 
heating function while its transparency did not interfere with 
the entrance of light into the inner chamber. Experiments 
were made at the same temperatures as in the dark—3° to 5°C., 
16°C. and 33°C., but the results were, for the most part, condi- 
tions of uniform expansion independent of temperature. 

It was observed that when frogs were subjected to a low 
temperature they became quiescent and tended to keep their 
eyelids closed. Although the lower lid (the only one which 
is movable to any extent) is more or less transparent, the possi- 
bility of its influencing the results led to its removal in a number 
of instances; no difference, however, was obtained by the ob- 
servance of this precaution. From these results, therefore, 
we conclude that in darkness, temperature is the controlling 
factor, while in the light temperature is subordinate to the 
stronger stimulus, light. 

At this juncture a doubt arose as to the exact temperature 
the frog’s retina was experiencing while in the apparatus. It 
is well known that at the surface of the frog’s body rapid evapora- 
tion can take place; hence it is perfectly conceivable that the 
rate of evaporation in the temperature chamber might be such 
as to keep the body temperature for some time considerably 
below that of the surrounding air. This possibility was first 
checked by taking the oesophageal and rectal temperatures of 
animals that had been subjected to various temperatures in the 
apparatus for several hours. The recorded temperature, how- 
ever, was neve found to vary more than a fraction of a degree 
from that of the air in the containing chamber. 

To be absolutely certain on this point, a prolonged set of ex- 
periments was made in which the frogs were immersed in water 
media of appropriate temperatures. It is certain that after a 
short time the animal, as a whole, must assume the temperature 


MOVEMENTS IN THE VISUAL CELLS 147 


of the medium irrespective of activities at the surface of the 
body. ; 

The apparatus for this verification consisted merely of a 
large battery jar, a sheet of coarse wire gauze and suitable weights. 
The jars were filled with water to within a quarter of a centimeter 
of the top, and the gauze, held in place by weights, served as a 
cover. This device worked in the following manner. Animals 
coming to the surface to breathe could only get their nostrils 
above water, the rest of the head and body remaining submerged, 
hence, in a short time the body temperature of the frog neces- 
sarily approximated that of the surrounding medium. 

At 3°C. the body temperature of the animal quickly fell, the 
respiration rate was reduced until it practically ceased and the 
bodily activities diminished until the frogs, for the most part, 
remained quietly at the bottom of the jar, although some animals 
would occasionally swim to the surface to breathe. At 33°C., 
on the contrary, the frogs were very active and had to return at 
short intervals to the surface, where they would sometimes remain 
for several minutes clinging to the netting. 

A series of experiments was performed both in darkness and 
in light. In the dark nothing new was learned beyond the con- 
ditions already described. In the light, the first experiment 
showed a state of extreme pigment expansion at 3°C., which 
was comparable to that of Ameiurus under similar conditions. 
The other trials, at 16°C. and 33°C., on the contrary, showed the 
pigment uniformly distributed. Another experiment, a short 
time after, gave the same result at 3°C. but not so decisively. 
The possibility of discovering a similarity in the pigment re- 
sponses of frogs and fishes in the light, led me to repeat these ex- 
periments many times without, however, again obtaining similar 
results. 

If extreme pigment expansion occurred at 3°C. it would be 
interesting from another standpoint since Herzog reported 
a similar condition, in darkness, at the highest temperatures 
which these animals can withstand. 

Occasionally during experiments both in light and in dark- 
ness, an anomalous condition arose whereby the distribution 


148 LESLIE B. AREY 


of pigment in one part of the retina was markedly different 
from that in the remaining portions." Such conditions may 
have been due to a variety of disturbing factors, Angelucci 
(90) has recorded noises, unilateral pressure on the eyeball and 
‘mechanical or electrical stimulation of the body as causing the 
migration of pigment in dark-adapted animals. Herzog (’05) 
likewise states that a frog tied up for 24 hours in the dark showed 
the pigment in the light position. A whole series of experiments 
and observations are on record to show motor control of some 
sort not well understood. It certainly is evident from a com- 
parative study of pigment in other forms, that is, in the retinas 
of fishes as well as in vertebrate and invertebrate melanophores, 
that the situation in the frog is entirely unlike that in any other 
animal concerning which we have data. 

Herzog explained the temperature responses of the frog’s 
pigment in the following way. If the effect of temperature is 
purely physical, its action presumably consists in accelerating 
or retarding chemical processes in the protoplasm of the pigment 
cell. Since, however, the movements of the dark-adapted 
pigment are not directly correlated with the temperature grad- 
ient, a physical action of temperature is probably not responsi- 
ble for the observed phenomena. If, on the other hand, it is 
assumed that the response of the pigment involves the principle 
of ‘specific energies,’ then any positive stimulus, acting through 
the nervous system, will cause a pigment expansion, and thus 
a satisfactory explanation for the known facts is furnished. 

In connection with the special case offered by the frog an in- 
teresting speculation arises as to the kind of pigment responses 
shown by the frog larva. The larva; in a general way, is com- 
parable to a fish; at least it may be said that the larval stage 
recapitulates certain conditions persistent in the adult fish. Is 
it not possible, therefore, that under the action of temperature 
the pigment of the larval frog will show a distribution similar to 


11 A somewhat similar lack of consistency was also noted by Fick (’90), in 
his attempts to obtain maximal contraction in dark-adapted eyes. He concluded, 
however, that inequality in the pigment distribution was characteristic of dark- 
adapted retinas. 


MOVEMENTS IN THE VISUAL CELLS 149 


that in fishes? The material, on which an answer to this ques- 
tion was sought, was the larva of the bullfrog (R. catesbiana). 
Animals were obtained during the month of April, 1914; these, 
of course, represented larvae hibernated from the previous season, 
since two or even three years are required to complete the meta- 
morphosis of this species. A few experiments were also made 
on animals obtained in November, 1914. 

From the larvae procured, two size limits were selected for 
experimentation. The smallest larvae had a total body length 
of 4.5 em., the hind legs of such animals not being visible; the 
largest larvae had a body length of 7.0 em., and the hind 
legs were developed as two small buds with the digits just 
differentiated. 

In neither the 4.5 em., nor the 7.0 em. animals were the eyes 
as deeply pigmented as in the adult of R. pipiens. Of the two 
sizes, the 7.0 em. larvae had the pigment more highly developed, 
and consequently better differentiation was obtained at the 
various temperatures with these, than with the smaller animals. 

The 4.5 em. larvae at 3°, 26°, and 32°C. in the dark showed 
the pigment in an expanded state (cf. figs. 20, 22) but not in a 
firm zone of uniformly distributed granules. On the contrary, 
the cells displayed pigmented processes that seemed to be more 
or less independent of each other; the appearance of the zone 
being that of a more uniform base with a fringe of pigment 
extending distally from it. The degree of expansion at 26°C. was 
distinctly less than at either 3°C. or 32°C., although at neither 
of the extreme temperatures was the pigment as fully extended 
as in the light. At 16°C., however, a striking difference was 
found (ef. fig. 21), for the pigment lay contracted in a narrow 
compact layer near the choroid. 

The 7.0 cm. animals gave results (figs. 20, 21, 22) quite similar 
to those just described, although the contrasts at the various 
temperatures were considerably sharper due to the heavier pig- 
mentation of the eyes at this stage. 

In these larvae, therefore, the behavior of the retinal pigment 
to temperature is identical with that characteristic of adult frogs. 
Since these animals were always immersed in water, there is no 


150 LESLIE B. AREY 


question concerning the correspondence of their body temperature 
and that of the surrounding medium. 

What would be discovered in a study of earlier stages I can 
not say, but I suspect great difficulty would be encountered in 
interpreting the results due to incomplete pigmentation, for 
the differences at various temperatures in the 4.5 em. tadpole, 
although fairly well marked, showed much less contrast than those 
exhibited by the 7.0 cm. animals. 

5. Necturus. A comparison between the frog and some uro- 
dele suggested another interesting problem. Is the condition 
exhibited in the frog restricted to anurans or is it common to 
the whole group of amphibians? This query becomes all the 
more pertinent when it is recalled that urodeles are not in the 
direct line of ascent to the anurans; that is, they are not amphib- 
ians which have never gone beyond the water inhabiting stage, 
but are more probably a group that were once land animals 
and have again returned to the water as a secondary adaption. 

The common mud puppy, Necturus maculatus, was chosen 
because of the ease with which it is procured and kept in cap- 
tivity. These animals were treated according to the technique 
used for fishes, hence a temperature higher than 28°C. was 
not attempted. 

The first experiments performed were to secure typical light- 
and dark-adapted retinas in order that some basis of compari- 
son might be had. The results were by no means as striking as 
one might wish. In both cases the pigment was extended, not 
in a band with even contours, but generally in large conical proc- 
esses from the individual cells, which, like other cells of Necturus, 
appear to be very large; these conical processes surround the 
distal ends of the outer segments of the rods. 

In the light, the pigment was usually somewhat more ex- 
tended than in darkness. The processes were not of uniform 
length but mean measurements may be expressed by the values 
of about 38 uw in the light and 30 » in the dark,—a condition 
of slight contrast when compared with fishes or the frog. The 
presence of a certain amount of migration in Necturus has 
been previously noted by Howard (’08), who took advantage of 


MOVEMENTS IN THE VISUAL CELLS Tt 


the contraction in the dark in his study of visual cells. This 
situation is comparable to that found in Triton where movements 
of the pigment through the influence of light were discovered 
by Angelucci (’78), Stort (87) and Garten (’07). The extent 
of pigment migration in this animal is described as being very 
limited by Garten (p. 70) who says: “. . . . dieselbe ist 
aber hier unvergleichlich viel schwiicher als bei vielen anderen 
niederen Wirbelthieren.”’ 

In an examination of the effects of various temperatures 
on dark-adapted eyes, however, no constant differences were 
noted. In some instances the processes of the cells seemed to 
be less heavily pigmented and the bases more heavily pigmented 
at 15°C. than at the two extreme temperatures, but this was by 
no means constant. At least, it can be said that there is no 
marked contraction of the pigment in the dark at any temperature. 

The evidence from Necturus and the limited pigment migration 
in Triton conclusively prove that the pigment responses typical 
of the frog are not common to all amphibians. It is probable, 
therefore, that such peculiarities as were described for the frog 
have been developed solely within the anuran group. 


b. Visual cells 


Although the myoids of both the rod and cone cells of fishes 
are capable of a high degree of contractility (a 90 per cent re- 
traction occurs in some instances), the effect of various tempera- 
tures on these cells remains untried up to the present time; in 
fact, the frog is the only animal upon which such work has been 
attempted. Gradenigro (’85), Angelucei (’84b), Herzog (’05), 
and Fujita (11) found that warming produces the same effect 
on the cones of the frog as does light. Herzog also stated that 
cooling to 0°C. likewise caused the cones to shorten, although 
Fujita denies that this occurs. 

There is no record of any attempt to Rercimine the effect of 
temperature upon the rods of vertebrates beyond the statement 
of Gradenigro (’85) that at 30°C. the rod of the frog shortens 
as in light. 


152 LESLIE B. AREY 


The object of the work upon visual cells to be described in 
this section parallels that stated for pigment, that is, to deter- 
mine the effects of various temperatures on the myoid of rod 
cells and of cone cells in normal fishes and amphibians. 

The apparatus and technique employed were similar to those 
which were used in the experimentation upon pigment. Par- 
ticular care was exercised at the termination of experiments 
conducted in the dark to guard against the action of light on the 
highly sensitive visual cells. In the fishes, measurements of 
the cone myoid were made from the external limiting membrane 
to the proximal edge of the ellipsoid, and in the frog, from the 
external limiting membrane to the proximal side of the oil drop 
which is situated at the distal end of the ellipsoid. The lengths 
of the rods were measured from the junction of the inner and 
outer members to the external limiting membrane. Each value 
given in the tables represents the mean measurement, in micra, 
of many (12 to 24) individual elements. 

1. Fishes. (1) Ameiurus. Of the four fishes studied, the 
cones of Ameiurus, in many ways, gave the least satisfactory 
results. These elements are not located at uniform levels and 
the differences between the elongated and shortened condi- 
tion, when stimulated by extremes of temperature, are not strik- 
ing to the eye. The additional fact that the cones, when maxi-. 
mally shortened under the influence of temperature, never 
closely approach the external limiting membrane makes these 
animals rather unsatisfactory for certain kinds of experimental 
work. 

Tables 1 and 2 present data for both rods and cones from typical 
retinas at 5°C. and 25°C. in the dark. 

It will be seen that at the higher temperature the myoids of 
both cones and rods lengthen (figs. 32, 33). Especially in the 
cones is this response unmistakable. The lengths of the rod’s 
inner member, after subjection to the extreme temperatures, 
varies within only a few micra, yet the relative change may be 
25 per cent or more; moreover, since the mean ranges at the 
extreme temperatures do not overlap, these differences are 
presumably significant. If the length of the rod ellipsoid, 4 u, 


MOVEMENTS IN THE VISUAL CELLS 153 


TABLE 1 


Measurements from the retinas of four Ameiurus which had been kept at 5°C. in the 
dark. The values are in micra and represent measurements taken along azes 
coinciding with radii of the eyeball 


NERVE FIBER CHOROID TO 


EXTERNAL LIMITING CONE MYOID | 


NUMBER OF LAYER TO : ROD INNER 
ANIMAL Mt eee ae MEMBRANE MEMBER 
1 35 95 19-32 7-9 
2 35 95 13-20 7-9 
3 50 95 19-31 9 
4 43 100 13-24 9-10 
MMicanc.:.. . 41 96 16-27 8-9 
TABLE 2 


Measurements from the retinas of four Ameiurus which had been kept at 25°C. in 
the dark. The values are in micra and represent measurements taken along azes 
coinciding with radii of the eyeball 


NERVE FIBER 
CHOROID TO j 
NUMBER OF LAYER TO ee Sse a ROD INNER 
ANIMAL EXTERNAL LIMITING sapere saa CONE now | mg MEMBER 
MEMBRANE ok erie 
— oe us _ — 
1 50 105 | 32-38 9-13 
2 43 100 32-38 a 
3 42 87 25-31 
| 
4 50 | 93 31-42 ‘5 iG 
————— — COC | ——_— — —_—_—__ - — 
Mean... 46 96 30-37 10-13 


is subtracted from the mean values, the relation existing be- 
tween the corresponding values at either end of the temper- 
ature range will be expressed by the following proportion: 
25°C: 5°C. = 3:2. The variability in the length of adjacent 
rods in any section is relatively so great that this disparity in the 
length of the myoids is not apparent until actual measurements 
are made. 

In a few observations upon the effect of temperature on the 
extended rods in the light, the range at 25°C. seemed to run higher 
than at 5°C. by 15 to 20 percent. The lengths of the extended 
inner members were approximately 50 wat 5°C. and 60 uw at 25°C. 
Similar tendencies will be noted among certain other fishes. 


THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL. 26, NO. 2 


154 LESLIE B. AREY 


(2) Abramis. The cones, with their large oval ellipsoids 
10 » in length, are very conspicuous in stained preparations. 
The fine rods, on the contrary, are not always. demonstrable 
when treated with the Ehrlich-Biondi stain, which acts in an 
unusually capricious fashion in respect to these elements. 

The elongated cone cells are more or less uniformly extended, 
although the variability in length is greater under these con- 
ditions than when in the retracted state. The rods, however, 
are arranged in the dark at very irregular levels so that retinas 
present a fairly even distribution of them from 8 » to sometimes 
as far as 50 uw from the external limiting membrane. In the light 
the condition is one of more uniform elongation. 

In tables 3, 4, and 5, are presented the data obtained from 
measurements at 5°, 15°, and 25°C. in the dark. 


TABLE 3 


Measurements from the retinas of four Abramis which had been kept at &°C. in the 
dark; the values are in micra and represent measurements taken along axes coin- 
ciding with radii of the eyeball 


NERVE FIBER 


CHOROID TO ROD INNER 
a ee ath a Se EXTERNAL LIMITING CONE MYOID MEMBER _ 
ASR LIRUNS MEMBRANE (MAXIMUM) 
1 90 90 7-12 
2 72 85 . 5-12 
3 85 91 5- 9 40 
4 82 88 6-10 
Meanityen ct. 82 89 6-11 40 
TABLE 4 


Measurements from the retinas of three Abramis which had been kept at 15°C. in the 
dark; the values are in micra and represent measurements taken along axes coin- 
ciding with radii of the eyeball 


NERVE FIBER 
NUMBER OF LAYER TO CHOROID TO ROD INNER 
/GSRRIG EXTERNAL LIMITING|=XTERNAL LIMITING CONE MYOID or 
MEMBRANE WAUSINAE BEND M 

1 95 94 16-29 59 

2 100 105 15-24 60 

3) 88 69 24-34 25 

Means. 22:4 94 89 18-29 48 


MOVEMENTS IN THE VISUAL CELLS 155 


TABLE 5 


Measurements from retinas of four Abramis which had been kept at 25°C.in the dark; 
the values are in micra and represent measurements taken along axes coinciding 
with radii of the eyeball 


NERVE FIBER 


CHOROID TO ROD INNER 
SEL Noy LAYER TO EXTERNAL LIMITING| CONE MYOID MEMBER 
ANIMAL EXTERNAL LIMITING sirsakan! abeavrenn) 
MEMBRANE ‘ 
1 80 89 21-49 
2 72 84 28-40 44 
3 105 84 30-40 
4 76 65 28-40 44 
WWIGATI. oc wictes.- 83 81 27-42 44 


The effect of temperature, therefore, upon the cones of Abramis 
is very marked, the length of the myoid at 5°C. (fig. 25) averag- 
ing only 25 per cent of that at 25°C. (fig. 27), while in extreme 
cases this ratio is as low as 10 per cent. If the mean limits of 
myoid extension are averaged, values of 9, 24, and 35 micra are 
obtained for the temperatures of 5°, 15° and 25°C. respectively. 
These values are in the ratio of 1.0: 2.5:4.0, or in other words, 
the ‘coefficient of expansion’ for the myoid of Abramis is 2+ 
for 10°C. 

As a matter of fact, in the great majority of these temperature 
experiments, 5°C. represents a value too high and similarly 25°C. 
a value too low for the actual temperatures maintained. 3°C. 
and 26°C. are more nearly the actual values. If temperature 
is plotted as abscissas and the myoid length in micra as ordinates, 
the resulting curve (fig. A) is a straight line showing that the 
temperature effect is uniform between these limits. 

The straight line obtained in the plot may indicate that the 
temperature response is the result of two or more opposed chemi- 
cal reactions which operate in a compensatory manner. Since 
the response of the myoid in elongating is directly correlated 
with the temperature gradient, it seems feasible that the effect 
of temperature is physical (in the sense of Herzog), and through 
its action chemical processes in the protoplasm are uniformly 
accelerated. If the length of the myoid is a fair index of the 
chemical activity that causes elongation, and if the effect of 


156 LESLIE B. AREY 


temperature is purely physical, the coefficient of 2+ for 10°C. 
is of interest, on account of its agreement with the value found 
for the temperature coefficient of various vital processes as well 
as of ordinary chemical reactions. 

The incomplete data concerning the maximum lengths of 
the inner members of rods are hardly significant, although 
both of the mean values at 15°C. and 25°C. are slightly above 
the one measurement at 5°C. The Ehrlich-Biondi stain was 
used on most of these preparations and it was only rarely that 
the rod ellipsoid took the stain sufficiently to render its identi- 
fication certain. 


LENGTH OF CONE MYOID IN M’CRA 


TEMPERATURE IN DEGREES CENT. 


Fig. A. Plot showing the relation between temperature and myoid length 
in the cone cells of Abramis. 


Table 6 gives measurements from two retinas which had been 
subjected to extreme temperatures in the light. The measure- 
ments for the cone myoids are identical, and in no one of the 
four fishes was there a demonstrable change ascribable to 
temperature under these conditions. It should be noted that 
the cone measurements at 5°C. in the dark (table 3) are either 
equal to or, as in this case, are actually smaller than those repre- 
senting the highly retracted light condition. This dependence 
upon temperature was taken advantage of in all experimenta- 
tion to be described later where elongation of the cones was 


MOVEMENTS IN THE VISUAL CELLS 157 


desired. Although the disparity between the rod lengths given 
in the table is probably extreme, such values have the same 
purport as the corresponding measurements made on the rods 


of Ameiurus. 
TABLE 6 


Measurements from the retinas of two Abramis, one of which had been kept at 5°C., 
the other at 25°C. in the light; the values are in micra and represent measurements 
taken along axes coinciding with radii of the eyeball 


NERVE >I 
- eau Ti Pee | ROD INNE 
NUMBER OF TEMPERATURE ras . TO EXTERNAL es . Yee hts 
near, on TO EXTERNAL ciMrria CONE MYOID MEMBER 
: LIMITING SSS (MAXIMUM) 
MEMBRANE Sah i 
ees [Sas cae, oe ‘a = 
1 5 54 78 8-14 50 
2 25 88 100 8-14 75 


(8) Fundulus. The retina of this fish is interesting because 
of the presence of prominent ‘double cones.’ Such elements 
are found in representatives of all the vertebrate classes, with 
the exception of mammals (Greeff, 00). They consist of two 
cones with fused inner members, although close examination is 
necessary to demonstrate this union. One component is usually 
larger and is known as the ‘chief cone’ (fig. 29; ell. con.), where- 
as the smaller is the ‘accessory cone’ (fig. 29; con. acc.) The 
chief cone alters its position independently of its accessory 
cone, which remains close to the external limiting membrane and 
is not moved to any great extent by the action of light or other 
stimulating agents. 

The rods, although rather small, are quite in evidence and 
differ from those of Abramis in maintaining fairly uniform de- 
grees of elongation. 

In the tables (7 and 8) which show the characteristic results 
of experimentation in the dark, the mean myoid length of the 
chief cones at 5°C. (fig. 28) is only about one-third that at the 
higher temperature (fig. 29). The differences in the accessory 
cones are not striking, although the lower values are somewhat 
increased at 25°C. Since the movements of the accessory cones 
are very limited, this probably represents a significant elongation. 

The contrast between the extension of the rod at 5° and 25°C. 
is striking and, added to the evidence gained from other fishes, 


158 LESLIE B. AREY 


TABLE 7 


Measurements from the retinas of five Fundulus which had been kept at &°C. in the 
dark; the values are in micra and represent measurements taken along axes coin- 
ciding with radii of the eyeball 


NERVD FIBER 


: CHOROID 
NUMBER OF we role TO EXTERNAL | CHIEF CONE ACCESSORY ROD INNER 
ANIMAL LIMITING . LIMITING MYOID CONE MYOID MEMBER 
MEMBRANE MOM RAND 
1 100 69 5-11 1-5 16-19 
2 106 68 5 1-5 
3 88 69 5 1-5 
4 105 68 3 1-4 11-19 
5 5: 52 5-9 1-5 16-23 
IMGamen cates 95 65 5-7 1-5 14-20 
TABLE 8 


Measurements from the retinas of five Fundulus which had been kept at 25°C. in the 
dark; the values are in micra and represent measurements taken along axes coin- 
ciding with radii of the eyeball 


NERVE FIBRE 


CHOROID 
NUMBER OF LATE TO EXTERNAL | CHIEF CONE ACCESSORY ROD INNER 
ANIMAL pig ear at a LIMITING MYOID CONE MYoID | MEMBER 
MEMBRANE ER ANe 
1 120 87 14-22 2-5 24 
2 120 87 15-22 2-6 31 
3 130 100 17-26 6 7A 
4 135 94 17-27 6 19 
5 135 110 17-26 1-3 38 
Meanie sss: 2. 128 96 16-25 3-5 27 


indicates that in these animals a lengthening of the inner mem- 
bers is favored by a high temperature. 

A series of measurements of chief cones in the light failed to 
show any differences at the extreme temperatures. 

(4) Carassius. The retina of Carassius, as well as that of 
Fundulus, has prominent double cones. 

In the few eyes measured, the chief cone elongated with in 
creased temperature (figs. 23, 24) but the accessory cone did not 
change its position, at least to any extent. 

Table 9 summarizes the results from typical retinas. 


MOVEMENTS IN THE VISUAL CELLS 159 


TABLE 9 


Measurements from the retinas of four Carassius, two of which had been kept at 5°C., 
and two at 25°C. in the dark; the values are in micra and represent measurements 
taken along axes coinciding with radii of the eyeball 


NERVE FIBER | 
LAYER CHOROID 


NUMBER OF TEMPERATURE TO EXTERNAL | CHIEF CONE ACCESSORY 
ANIMAL 5 oH se ates LIMITING MYOID CONE MYOID 
es MEMBRANE 
MEMBRANE 7 
1 25 75 88 21 3-5 
2 25 69 69 21 2-3 
3 5 56 63 1-3 2 
4 5 69 94 3-4 2-3 


2. Frog. (1) Rana pipiens (adult). Gradenigro (’85), Ange- 
lueci (90), Herzog (’05) and Fujita (’11) have all stated that at 
a temperature of 30°C. or more, the cones of the frog shorten 
until they assume the position characteristic of light. Thus, 
Herzog (p. 419) says: “‘Aufenthalt im Briitschrank 1/2 Stunde 
lang, Temperatur von 21° bis 30° C. ansteigend: . . . die- 
selbe zeigt auch, dass die Zapfen maximal contrahiert sind. Die 
Linge der Zapfen betrigt mit ganz vereinzelten Ausnahmen 
0.0091 mm. (v. d. Lim. extern.—QOelkugel exel.).” 

The action of low temperature was first tried by Herzog (05), 
who makes the following statement (p. 424), concerning the re- 
sult of two hours’ cooling to 0°C. in the dark: ‘“‘Dagegen sind 
die Zapfen bereits nahezu héchstgradig verkiirzt. Ihre Linge 
betragt fast durchweg 0.0078 bis 0.0091 mm.”’ 

Fujita experimented upon six frogs, which were kept at 
a low temperature in an ‘Eisgefass’ for periods of 30 minutes 
to 6 hours, after which the decapitated heads were fixed in 
ice-cold fluid. His conclusion (p. 170) is diametrically opposed 
to that of Herzog. ‘Das Resultat war in allen Fallen das gleiche: 
ich konnte keine Hellstellung konstatieren. Die Zapfen waren 
micht kkontrahiert. . . . .” 

Measurements showing the elongation of the cone myoid at 
intermediate temperatures (14° to 19°C.) are not given by Herzog, 
who, however, states (p. 418) concerning retinas that had been 
raised in the course of 15 minutes from 18° to 24°C.: ‘‘Ein 
wesentlicher Einfluss is nicht zu erkennen. . . . . Die 


160 LESLIE B. AREY 


Zapfenlinge (von der Limitans externa bis zur Oelkugel im Ellip- 
soid exclusive) betrigt im Durchschnitt maximal 0.034 mm., im 
Minimum 0.0169 mm.”’ 

With apparatus and methods similar to those described in 
connection with the experimentation upon frog’s retinal pigment, 
an attempt was made to discover the exact responses of the cone 
myoid to high, medium, and especially to low temperatures. 

In table 10, which gives the measurements of visual cells 
from a few typical preparations, it will be observed that the cones 
are greatly shortened at 33°C. (fig. 36), but that at the other 
temperatures they retain the elongated condition typical of 
darkness (figs. 34, 35).1% Measurements of the cones gave two 
modal lengths, which are approximately expressed by the figures 
in the table representing the extreme values. The maintenance 
of elongation at a low temperature is in agreement with Fujita’s 
(711) experiments, but is opposed to Herzog’s conclusion, which 
was apparently based upon exhaustive investigation. It is true 
that the temperature of 0°C. which the latter worker used was 
a few degrees less than the lower limit (3° to 5°C.) employed 


TABLE 10 


Measurements from the retinas of eight Rana pipiens which had been kept at 3°, 14°, 
19°, and 33°C., respectively, in the dark; the values are in micra and represent 
measurements taken along axes coinciding with radii of the eyeball 


NERVE FIBER anencimn 
NUMBER OF TEMPERATURE | ,, ae Ae TO EXTERNAL | CONE INNER ROD INNER 
ANIMAL 2C: LRG: LIMITING MEMBER MEMBER 
MEMBRANE MEMBRANE 

1 3 140 66 12-20 10+ 
Z 3 160 75 13-22 Q=+ 
3 14 103 56 11-20 1+ 
4 14 88 50 11-16 10+ 
5 19 103 61 13-19 13+ 
6 19 94 63 14-23 10+ 
7 33 163 63 8-12 10+ 
8 33 126 63 7-10 11+ 


12 The actual values at medium and high temperatures vary somewhat from 
those given by Herzog for R. temporaria (and R. esculenta?). 

13 This experimentation upon the retinal elements of the frog was practically 
completed before Fujita’s paper was known to me. 


MOVEMENTS IN THE VISUAL CELLS 161 


by me, yet it seems improbable that such a small difference would 
cause the cones to shorten maximally. Any error that could 
be introduced during these determinations would tend to shorten 
the cones, hence Herzog’s results are at a disadvantage in this 
respect. I have continued experiments for 6 hours, yet the 
results were always the same; in no case was there found a gen- 
eral shortening of the cones that in any way resembled the ex- 
treme condition at 33°C. 

From these results, which, in a general way, are the reverse 
of those found in fishes, it is evident that the responses of the 
cones are not comparable to those of the retinal pigment. The 
pigment may indeed be under nervous control, so that stimulat- 
ing agents such as heat, cold and light produce a migration ac- 
cording to the principle of specific energies, yet if the cone cells 
are influenced by the nervous system, these experiments can not 
be said to furnish proof of such a relation. 

In order that there should be no doubt concerning the effect 
of low temperature upon the cone myoid, a further determina- 
tion was made long after the results which are tabulated above 
were obtained. In this later experiment rigorous precautions 
were observed to eliminate possible errors. After frogs (R. 
pipiens of various sizes), kept at a temperature of 18°C., had been 
subjected to a preliminary treatment of darkness for 72 hours, 
they were introduced into a vessel cooled to 1°C., where they 
remained for a period of 4 hours. The eyes were then quickly 
excised in dim, red light, the operation not requiring more than 
15 seconds, after which they were returned into darkness where 
fixation at approximately the freezing temperature ensued. 
The average measurements of the cone myoids in these prepara- 
tions were as follows: 12 to 14 »; 10 to 15.4 uw; 10 to 14 u; 10 to 
11 4313 to 15.4 4;9 4. These values, although somewhat smaller 
than those given in table 10, can not be said to prove that low 
temperature shortens the myoids as does high temperature. 

Gradenigro (’85) found that elevated temperature induced 
a shortening of the rod in the dark. After measuring the myoid 
length in a considerable number of preparations from retinas 
subjected to various temperatures both in light and in darkness, 


162 LESLIE B. AREY 


I was unable to discover constant differences in length that 
could be correlated with definite temperatures. The agreement 
of mean values obtained from various retinas under identical 
temperature conditions was not good, and since the rod myoid 
measures only 6 » to 12 uw in length, even small variations fur- 
nish serious obstacles in determinations of this kind. In any 
one preparation, moreover, variability in the length of adjacent 
myoids tended somewhat to mask a possible temperature effect. 
If anything, my measurements showed the reverse of what 
Gradenigro maintained, the elongation at 33°C. in the dark 
being greater than at 5°C., but, as stated before, these results 
are by no means trustworthy. 

A shortening of the rod through the action of high tempera- 
ture, as claimed by Gradenigro, is of interest because, accord- 
ing to most investigators, ight produces the same result. With 
this can be compared an analogous correlation in fishes, where 
light causes an elongation of the rod myoid and, as I have shown, 
elevated temperature does likewise. It is certain, on the con- 
trary, that although the cones of both frogs and fishes shorten 
in the light, heating produces unlike responses in the dark. 

(2) Rana catesbiana (larvae). Although the cones in both 
the 4.5 cm. and the 7.0 cm. larva of this frog are of large size, 
clearly defined temperature responses were not observed; in- 
deed, the difference between the positions assumed even in light 
and darkness is not striking, the cone myoids in the light re- 
maining well elongated in comparison to those of the adult R. 
pipiens. The variability in length in different preparations is 
considerable, yet if anything, the cones appeared more shortened 
at 3°C. than at higher temperatures. There is no marked shorten- 
ing at 33°C., for the cones under these conditions were as long 
as at lower temperatures, and in some cases longer. It is 
probable that the responses of the cones in adult R. catesbiana 
will be found to agree with those in other species which have been 
studied, although no experimentation was performed to deter- 
mine this point. 

3. Necturus. The rods, and the single and double cones of Nec- 
turus are very large, yet positional changes with varying tempera- 


MOVEMENTS IN THE VISUAL CELLS 163 


ture (3° to 28°C.) were not observed. Individual cells vary more 
or less in the height at which they are situated above the exter- 
nal limiting membrane, yet no constant differences of significant 
amount could be correlated with definite temperatures. With 
these results should be compared Garten’s (’07) denial of a 
change in the position of the cones of the ‘salamander’ through 
the influence of light, such as Angelucci (’90) had previously 
claimed. Stort (’87), however, described movements in both the 
rods and the cones of Triton. 


C. EXPERIMENTATION UPON EXCISED EYES 
a. Effect of light and darkness 


The results of a number of investigators since the first work 
of Englemann (’85) have indicated that the retinal elements of 
some vertebrates, and especially the frog, are subject to a nervous 
control, the action of which is not well understood. 

The réle which the nervous system plays either in producing 
or in assisting the movements of the various retinal elements 
is hard to demonstrate. Experimentation involving the direct 
action of light on excised eyes can not be expected to solve the 
problem decisively, for if no movements result, autoanaestheti- 
zation, or some similar disturbance due to the interrupted blood 
supply, may be the real cause. If, on the other hand, responses 
are called forth by direct stimulation, it by no means follows 
that a similar phenomenon necessarily occurs in the living animal, 
any more than a demonstration of the direct stimulation of 
muscle fibers proves that this rather than a nervous impulse is 
the normal method of muscle stimulation. The limitations which 
restrict a wide interpretation of results, however, do not lessen 
the interest involved in determining the extent to which the 
retinal elements can be directly stimulated. 

Hamburger (’89) maintained that the cones and retinal pig- 
ment in excised eyes of the frog assumed the positions characteris- 
tic of light or darkness according to the conditions of the experi- 
ment. Dittler (’07) working on isolated frog’s retinas obtained 


164 LESLIE B. AREY 


a shortening of the cones in localized areas through the action 
of light but found darkness to be ineffectual. The spread of 
the response to portions of the retina unstimulated by light, led 
Dittler to investigate further the cause of cone retraction. He 
was able to furnish experimental proof that weak acids, resulting 
from catabolic processes in the retina, caused the cone myoid 
to shorten; hence he concluded that the cone myoid was not of 
itself ‘lichtempfindlich,’ as Englemann (’85) had believed, but 
was stimulated to movement through chemical agents result- 
ing from the action of light on the retina. Fujita (11), as a 
result of very limited experimentation, stated that the pigment 
of the excised eye of a frog expanded in the light but did not con- 
tract in the dark. 

Ringer’s solution, normal saline solution, and tap water were 
used by me for the immersion of excised eyes. When the first 
two media were employed the movements of the rods and cones 
of Ameiurus, through the action of light, were never clearly 
demonstrated; possibly such results are to be interpreted as 
evidence of a chemical control somewhat comparable to that de- 
seribed by Spaeth (’13) for the melanophores of Fundulus. Tap 
water did not inhibit the movements of any of the retinal ele- 
ments of Ameliurus and consequently it was used in all subse- 
quent experimentation. 

The pigment of Ameiurus, Abramis and Fundulus did not 
contract when excised eyes from light-adapted fishes were sub- 
jected to darkness for periods of 4 hours or less. At most there 
was only evidence of a retraction of the distal accumulation of 
pigment, which is characteristic of light-adapted eyes, to form 
a more homogeneously pigmented zone (figs. 2, 10, 6). When 
the reverse experiment (subjection to light) was performed, 
the pigment of Ameiurus became maximally expanded in 2 hours 
(figs. 1, 3). Only the slightest tendency toward expansion, 
however, could be found after similar experimentation on the 
two other fishes. It thus appears that light acts directly on 
the pigment of Ameiurus only, while darkness is totally ineffec- 
tive on all three animals. 


MOVEMENTS IN THE VISUAL CELLS 165 


When the rods and cones of Ameiurus and the cones only of 
Abramis and Fundulus were tested, the following results were 
obtained. The rods of Ameiurus moved both in light and in 
darkness, whereas the cones were stimulated only by light, no 
elongation occurring in the dark even when the experiment 
continued for 4 hours. The cones of Abramis and Fundulus 
did not change their positions to any extent either in light or 
in darkness. In the light the cones of Abramis, which were more 
carefully investigated than those of Fundulus, at most showed 
only slight retraction and never closely approached the exter- 
nal limiting membrane. If light did exert a direct influence 
on the cones or retinal pigment of this fish, the changes would 
be extremely easy to distinguish due to the wide difference 
between the light- and dark-adapted phases. 

Since neither the cone cells nor retinal pigment of Abramis 
underwent movements under these conditions, it is possible 
that the accumulation of catabolic products, occasioned by the 
interruption of the blood supply, was responsible. Experimenta- 
tion of the following kind shows the importance of maintaining 
the vascular circulation. If the optic nerve only of Abramis is 
cut, the retinal elements undergo their normal movements in 
darkness and in light. If, however, all the blood vessels and 
muscles are cut and the eye ball is attached to the body by the 
optic nerve only, no movements result. The objection may be 
made that some nervous mechanism is deranged by cutting 
these muscles and blood vessels, but this is hardly probable, 
as further experimentation, to be presented in a subsequent 
paper, on this and other fishes has shown. 

In eases like the movements of the pigment or cones in excised 
eyes of Ameiurus through the action of light only, it is probable 
that an inhibitory tendency is also present, but the response to 
the stimulus furnished by light is sufficiently vigorous to over- 
come it. Either a less vigorous stimulus or response may eX- 
plain why no movements of the cones and pigment occur in 
darkness. 

An inhibition due to the presence of unremoved catabolic 
products, as postulated here, would be merely a form of auto- 


166 LESLIE B. AREY 


anaesthesia. As will be shown, carbon dioxide and other anaes- 
theties do, in fact, arrest the movements of all the retinal elements 
of fishes. 

Dittler (07) accounted differently for the absences of elonga- 
tion in the cones of isolated frog’s retinas which were introduced 
from light into darkness. In order to appreciate his way of 
viewing this situation, it is necessary to understand the general 
theory advanced by him to explain the movements of the cones. 
In darkness an equilibrium was supposed to exist in the metab- 
olism of the retina, the elongated cone myoid representing an 
unstimulated condition. Through the action of light, however, 
catabolic processes preponderate, and the accumulated acid 
wastes chemically stimulate the myoid to shorten. These con- 
clusions were based upon experimental evidence by which it 
was shown that weak, free acids could be detected if isolated 
retinas were subjected to light in limited amounts of Ringer’s 
solution, and further that such an acid solution was capable of 
causing other dark-adapted cones to shorten while still in the 
dark. To return to the case under consideration, Dittler believed 
that the accumulation of the catabolic products formed in the 
light merely continued its contractile influence after the isolated 
retina was removed into the dark, and since these products were 
not removed, the metabolic equilibrium could never be restored 
and consequently elongation failed to occur. 

This theory of chemical stimulation is not supported by the 
condition in Fundulus and especially in Abramis, where the 
cones of excised eyes do not shorten even when exposed to light, 
for under these favorable conditions the tendency toward the 
production of a catabolic excess should be maximum. In still 
another way Dittler’s theory does not explain a typical response 
of the cones of fishes. The cone myoid in isolated retinas of 
the frog shortens when the temperature is raised to 30°C. or 
more in the dark (fig. 36), and this fact Dittler used to sup- 
port his view in the following logical manner. It is well known 
that most chemical reactions are accelerated by raising the 
temperature; hence in this case the autonomic equilibrium 
normally existing in the dark would become disturbed, the re- 


MOVEMENTS IN THE VISUAL CELLS 167 


sulting increase of catabolic products causing the cone myoid 
to shorten.4 Although Dittler’s statement is not altogether 
clear, it seems evident that low temperature was supposed not 
only to reduce the metabolism of the retina to a low level, 
but also to render the cone myoid less ‘empfindlich’ to chemical 
stimulation. 

The conditions in the dark-adapted cones of fishes, however, 
are entirely different, for here not only do the cone myoids 
elongate when the temperature is increased, but also elongated 
cones can be made to shorten by the use of low temperatures. 
If the shortening of the cones of fishes in the light were due to 
a chemical stimulation, how can the elongation of these elements 
in the dark through the action of heat be explained, since mani- 
festly in this case the metabolic equilibrium tends to become 
destroyed, the result being the formation of an excess of catabolic 
wastes, which by analogy with the conditions in the frog, should 
cause the cone myoids to shorten? Moreover, the efficiency 
of low temperature in retracting elongated cones, and the cor- 
relation between the uniform degree of myoid elongation and 
the temperature gradient (p. 155) finds no explanation through 
Dittler’s hypothesis. 

It should be said, however, that Dittler strictly limited his 
conclusions, the experimental evidence for which appears to 
be well established, to the material upon which he worked, and 
was even reserved in suggesting the occurrence of a similar 
method of stimulation in living animals. 

The reason why the rods of Ameiurus move in darkness, while 
the cones do not, may be as follows. The rod myoid normally 
shortens in the dark, whereas the cone myoid elongates. It is 
probable that the contractile function of the myoid is more 
vigorous than the reverse process of elongation. This is not 
only substantiated by the fact that dark adaption of the rod 
takes less time than light adaption, but also by experiments 

14 Dittler did not formulate this conception in extenso as I have expressed it, 
yet several statements (pp. 317-318) show that this was his belief. . A concluding 
quotation reads: . . . . ‘der Einfluss der Temperatur iiberhaupt ganz nach 


physikalischen Modus zu Wirken scheint, und berechtigt uns, seine Wirkung rein 
in diesem Sinne zu fassen.”’ 


168 LESLIE B. AREY 


in which the optic nerve was cut. In such cases the rod never 
elongated in the light although it often showed a tendency to 
shorten when introduced into the dark, whereas the behavior 
of the cone cells in light and darkness was the exact opposite 
to that of the rod, since they tended to shorten in the light but 
remained unchanged in the dark. On these grounds, therefore, 
an explanation is offered to show why it is that through the strong 
stimulation produced by the direct action of light, both types 
of cells show characteristic responses, while in the dark only 
the more vigorous contractility of the rod myoid becomes effec- 
tive. It must be remembered, however, that although the direc- 
tion of the movements of the rod and cone cells are opposed, 
the real response of the protoplasmic myoid may be similar in 
both cases. If this were true, the apparent inconsistency in 
the movements of these elements would be due to a difference 
in the axis of contractility in the two kinds of myoids, and the 
explanation just advanced would not stand. Reference will 
be made to these possibilities in another place. 


b. Effect of temperature 


Previous attempts to determine the direct influence of tem- 
perature upon the retinal elements have been confined to the frog. 
Gradenigro (’85) found that if excised eyes of dark-adapted 
animals were subjected to a temperature of 30° to 386°C. the 
rods and cones shortened and the pigment expanded, both 
end-results béing characteristic of light-adaption. Dittler (07) 
was able to confirm Gradenigro’s discovery concerning the cone 
cells. When the isolated retina was heated to 35° to 37°C. in 
the dark for 50 to 60 minutes, the cone myoids shortened. After 
retinas had been subjected to a temperature of 1° to 2°C. for 
many hours, on the contrary, no shortening of the cone myoid 
was observed. 

The apparatus and methods used by me were similar to those 
described in connection with the experiments upon living fishes. 
The excised eyes were contained in test tubes which were sus- 
pended in jars of water kept at appropriate temperatures. Eyes 


MOVEMENTS IN THE VISUAL CELLS 169 


from the same animal were used simultaneously, one at each 
temperature extreme. 

1. Effect of temperature upon retinal pigment. After many 
trials it was found that sharp differentiation of the retinal pig- 
ment of fishes was hard to secure at the extreme temperatures 
when the initial temperature had been intermediate. Accord- 
ingly, the expedient was employed of subjecting the living ani- 
mals to a preliminary treatment either at 3°C. or at 25°C., and 
as a result uniformly satisfactory differentiation was obtained. 

The retinal pigment of Ameiurus behaved precisely as in 
living animals. At a low temperature, both in darkness and 
in light (figs. 3, 1), the degree of distal migration was greater 
than that at a high temperature (figs. 4, 2). Particularly in 
experiments conducted in the light was this strikingly appar- 
ent, since in many preparations at 3°C. the pigment migrated 
so far distally that the more proximal portions of the cells were 
free of granules, while a sharp line of demarcation existed be- 
tween the pigmented and non-pigmented zones. 

A few experiments were made upon the dark-adapted eyes 
of Abramis. In this case also, a greater pigment expansion 
was found at 3°C. than at 25°C. (figs. 11, 12). 

Reference has been made to the contention of many investi- 
gators that there is an apparent nervous control over the move- 
ments of the frog’s retinal elements. It has been shown that at 
a medium temperature in the dark the pigment is maximally 
contracted (fig. 18), whereas at higher and lower temperatures 
(figs. 17, 19) a considerable degree of expansion is effected. Not 
only is the amount of migration occasioned by temperature 
much more extensive than that in fishes, but also the similarity 
between the effects of the two temperature extremes as con- 
trasted with an intermediate temperature, has no parallel among 
other pigment cells or even in melanophores. 

Since it is at least agreed that the pigment of excised eyes of 
the frog expands in the light, it ought to be possible to observe 
the effect of temperature, if this agent acts directly upon the 
pigment cells. Excised eyes from animals that previously had 
been at an intermediate temperature in the dark, were subjected 


THE JOURNAL OF COMPARATIVE NBUROLOGY, VOL. 26, NO. 2 


170 LESLIE B. AREY 


to temperatures of 3°, 16°, and 33°C., but no changes were ob- 
served in the position of the retinal pigment. These results, 
which do not agree with Gradenigro’s statement, by no means 
furnish conclusive proof that in the living frog temperature 
operates through the nervous system, yet when supported by a 
comparative study of the retinal pigment and melanophores of 
other animals, the conclusion reached by Herzog (’05), that the 
expansion of the frog’s retinal pigment under these circumstances 
is of nervous origin, involving the principle of specific energies, 
becomes highly probable. 

According to Fujita (11), when excised eyes from dark-adapted 
frogs are retained in the dark for 20 minutes, the pigment assumes 
a partial light position. My observations do not confirm this 
result, for no migration of any consequence occurred. 

2. Effect of temperature upon visual cells. When a study of 
the visual cells of Ameiurus was made, the effect of temperature 
was found to be essentially similar to that upon normal animals. 

When eyes from dark-adapted individuals of Abramis that 
had previously been kept at a temperature of 25°C. were like- 
wise excised and subjected to 5°C. and 25°C. in the dark, 
those at 25°C. (fig. 27) retained the elongated position while 
those at 5°C. (fig. 25) shortened to a considerable extent 
although somewhat less than in living animals. Mention has 
already been made of the significance of these results in con- 
nection with the applicability of Dittler’s theory of chemical 
stimulation to the cones of fishes. Although the rods of Abramis 
at both temperatures showed a distribution extending over wide 
limits, yet the shortest measured 12 u at 5°C. as compared with 
20 wat 25°C., and the modal elongation at the same temperatures, 
as judged by the eye, was 18 » and 25 u respectively. It is 
interesting to note that in the excised eyes of dark-adapted 
Abramis, temperature is able to produce changes in both the 
retinal pigment and visual cells, notwithstanding the fact that 
light and darkness are wholly ineffectual in this respect. 

Table 11 summarizes these results from typical retinas. 

A few experiments were performed upon the cone cells of 
the frog. The results obtained were identical with those stated 


MOVEMENTS IN THE VISUAL CELLS 171 


TABLE 11 
Measurements of the visual cells from the retinas of two Ameiurus and two Abramis, 
of which one of each had been kept at 5°C. and the other at 25°C. in the dark; the 
values are in micra and represent measurements taken along axes coinciding with 
radii of the eyeball 


FISH TEMPERATURE PG: CONE MYOID ROD INNER MEMBER 
Ameiurus No. 1....... 5 4-16 5-6 
Ameiurus No. 2:...... 25 19-32 8-12 
IApramis No. b:.;.:3:..| 5 10-30 18 
Abramis No. 2 7 25 35-50 25 


by Dittler (07), who used isolated retinas. In dark-adapted eyes 
which were placed in water at a temperature of 33°C. the cone 
myoids shortened (fig. 36), while at 3°C. or 16°C. (fgs. 34, 35) 
the myoids remained for the most part unchanged. These re- 
sults are identical with those found by Fujita ('11) and myself 
on the cones of living animals, and indicate that, unlike the pig- 
ment cells, the movements of the cones are not dependent upon 
nervous control. If an influence of the nervous system over 
these elements exists in the normal animal, it is at least not mani- 
fested as is the control over the retinal pigment, in which changes 
at both high and low temperatures can be interpreted according 
to the principle of specific energies. 


D. EFFECT OF ANAESTHETICS 


Various instances have been noted throughout this paper in 
which the behavior of the retinal pigment and the visual cells, 
when deprived of their blood supply, cast suspicion upon auto- 
anaesthetization as being the factor causing suspension of move- 
ment. Certain conditions discovered in the responses of melano- 
phores in the web of the frog’s foot had previously suggested 
such a possibility; indeed, it was this difficulty which led to 
the abandonment of the frog’s melanophore as material for an 
investigation somewhat similar to the present one. In this way 
my interest was aroused to determine the effect of anaesthetics 
on the movements of the retinal elements, both in normal animals 
and through the more direct action upon excised eyes. 


172 LESLIE B. AREY 


The effect of certain drugs, as quinine and strychnine, upon 
the retinal pigment (‘protoplasmagifte’) is in dispute. It is 
clear, however, from the work of Ovio (95) and of Lodato (’95) 
that cocaine can arrest pigment migration. 

As a precaution against a possible source of error, animals 
were never introduced from one condition of light or darkness 
to the other without having been previously subjected to a 
brief preliminary treatment of the anaesthetic which was to be 
tested. 


a. Retinal pigment 


1. Carbon dioxide. The carbon dioxide used in these experi- 
ments was a commercial soda-water product sold under the trade 
name of ‘Pureoxia.’ Quantitative determinations of the con- 
centrations used were made by titration with #4) sodium 
carbonate, using phenolphthalein as an indicator. 

In the first experiments made on Ameiurus the movement of 
the pigment was arrested by a strong solution of carbon dioxide, 
but since none of the animals survived such treatment the- 
obvious objection exists that the pigment cells also may have 
been killed. 

A slight refinement in method consisted of revivifying the 
fishes at intervals, by temporary removal to running water, 
until opercular movements were restored. By this method 
fishes were kept alive for 2 hours, during which time four or 
five revivifying treatments were necessary. The migration of 
retinal pigment was shown to be checked both in light and in 
darkness, yet controls proved that the cells were not permanently 
injured. 

A method which gave more satisfactory results was devised 
after repeated trials had given a mixture of tap water and car- 
bonated water of sufficient strength to anaesthetize an Ameiurus 
but not to prohibit opercular movements of greatly reduced 
amplitude. The record of an experiment will well illustrate 
both the method and the results. 


Experiment 4.1.6. A dark-adapted Ameiurus was placed in a 
mixture of 1 part of ‘Pureoxia’ to 4 parts of tap water, and after re- 


MOVEMENTS IN THE VISUAL CELLS 173 


maining 10 minutes in the dark the jar was removed into strong dif- 
fuse daylight for 1{ hours. During this time, the fish was practically 
motionless except for a very weak but rhythmical pulsation of the oper- 
cular rims. At the end of the experiment one eye was removed and 
fixed. The Ameiurus was allowed to recover until the next day when 
the other eye was removed. The pigment in the eye which had been 
subjected to carbon dioxide was in the typical dark position (cf. fig. 4) 
while the pigment of the control eye was maximally expanded (ef. 
fig. 2). Titration of the anaesthetizing solution showed that the 
concentration of carbon dioxide had been in the ratio of 60.14 ec. per 
litre of water. 


In the converse experiment from light to darkness an Ameiurus 
lived 3 hours in a similar solution (53.07 ee. of carbon dioxide 
per litre) during which time the pigment retained its light dis- 
tribution, whereas the control eye removed on the next day, 
showed maximal contraction. 

These results prove conclusively that in the presence of certain 
concentrations of carbon dioxide the pigment cells are not injured 
but are in a condition of anaesthetization whereby there is a 
failure to respond to the normal stimulus causing contraction 
and expansion. Such experimentation, however, does not show 
whether this failure is due to a direct effect upon the pigment 
cells or to an inhibition through the central nervous system. 

To demonstrate which alternative is true, the effect of carbon 
dioxide was tested on excised eyes of Ameiurus. If, under these 
conditions, a migration occurs a direct influence of the anaesthetic 
on the cell itself will be disproven, while on the other hand, 
‘ if no migration ensues one can only infer that a similar direct 
action on the pigment cell is responsible for the whole course of 
events in the living fish, whereas an inhibition through the cen- 
tral nervous system may be involved as well. 

For such an experiment the excised eye of Ameiurus is well 
adapted, since its pigment has been shown to migrate from the 
dark to the light position, although the reverse process does 
not occur. Excised eyes of dark-adapted fish were exposed to 
light in a solution of carbon dioxide having a strength of about 
60 ce. per litre. The pigment in each case was arrested in the 
contracted position. 


174 LESLIE B. AREY 


The work described for living Ameiurus has been repeated 
on both Abramis and Fundulus with identical results. In every 
case the pigment maintained the position it occupied previous 
to the application of the anaesthetic. 

The results obtained in this study, as a whole, are very differ- 
ent from those of Fick (’90), who concluded that the retinal pig- 
ment of the frog expanded when the animals were subjected 
to an atmosphere of carbon dioxide gas. Fick attributed this 
result to asphyxiation and it is certain that the experimental 
conditions in his work differed greatly from those in my tests. 
In order to make the experiments more comparable, frogs should 
be treated with a mixture of oxygen and carbon dioxide gases 
in which they could live. 

2. Ether. The anaesthetic effect of ether on the retinal pig- 
ment was demonstrated by a series of tests that duplicate those 
described with carbon dioxide. Care must be observed against 
using an excess of ether since otherwise a partial or complete 
disintegration of the pigment cells results. 

Both dark and light trials were made on Ameiurus, Abramis, 
and Fundulus. In each animal the pigment was found to be 
completely arrested in whatever position it occupied at the be- 
ginning of the experiment. Controls proved that ether, if 
used in small amounts, does not permanently injure the pigment 
cells. 

Ether also checked the migration of pigment in excised eyes 
of dark-adapted Ameiurus when such eyes were subjected to light. 

3. Chloretone and urethane. ‘These substances are such satis- 
factory narcotizing agents that their effect was tested upon the 
retinal pigment of Ameiurus. Individuals lived in 0.1 per cent 
chloretone or in 1.0 per cent urethane, but the pigment was not 
arrested in its movement from the dark to the light phase. In 
concentrations of 0.5 per cent chloretone and 2.5 per cent ure- 
thane, the pigment migrated when fish were brought from dark- 
ness to light although the animals died in both cases. 

The results from all the foregoing experimentation are of 
interest in showing the difference in the effect upon pigment 
cells of four powerful anaesthetics, of which only two were 


MOVEMENTS IN THE VISUAL CELLS 175 


efficient. The experiments with chloretone and urethane also 
prove that even though the animal as an organism dies, the 
pigment, nevertheless, can expand independently. 


b. Visual cells 


Experiments similar to those just described were repeated 
in order to determine the action of anaesthetics on both rod and 
cone cells. Since the cone myoid is maximally elongated at 
about 25°C. in the dark (figs. 25, 27), this condition was taken 
advantage of in producing sharp contrasts between dark and 
light phases. The cones of Abramis and Fundulus, on account 
of their great contractility, were particularly favorable for ob- 
servation, as were the rods of Ameiurus because of their large size. 

The results of these experiments are shown in table 12. 


TABLE 12 


A tabulation of the effects of carbon dioxide and ether upon the movements of the 
visual cells of Ameiurus, Abramis, and Fundulus; X indicates that the movements 
of the elements were completely arrested; conditions corresponding to the blank 
spaces were nol investigated 


CONE ROD 
FISH eS ee eee od ms “ 
Dark to Light | Light to Dark Dark to Light Light to Dark 
Ameiurus....... 4 xX 
Abramis......<.. 4 x xX . 4 
MUNOUMISes <2. X x 


The conclusion is, therefore, that both ether and carbon diox- 
ide anaesthetize the visual cells of normal fishes to such an extent 
that neither light nor temperature is effective in causing positional 
changes. 

A few experiments upon the excised eyes of Ameiurus showed 
that both carbon dioxide and ether have the same anaesthetic 
effect on the rods and cones as that described for normal animals. 

Whether or not autoanaesthetization prevented movements 
of the retinal elements, as was suspected in previously described 
experiments when the normal blood supply was interrupted, it 
is at least demonstrable that certain anaesthetics do act in a 


176 LESLIE B. AREY 


similar way. The effect of carbon dioxide is especially interesting 
for, as it is the commonest catabolic product, it may have been 
the agent that prevented movements in those cases. This con- 
ception is opposed to Dittler’s (07) view, which assumes the 
existence of a balance in the metabolism of the unstimulated 
cone cells which is disturbed by the increased catabolism through 
the action of light. The movement of the cone cells in the 
isolated retina of the frog was stated by Dittler to be due to 
the action of a weak free acid, the product of increased catabolism. 
This conclusion, which was supported by experimental evidence, 
is opposed to that postulated by me; nevertheless, it must be 
pointed out that the responses of the retinal elements differ 
considerably in fishes and in the frog, and while evidence for 
autoanaesthetization is indirect yet the results obtained from 
experimentation upon fishes can be consistently interpreted 
in this way, whereas Dittler’s hypothesis does not meet all the 
known facts. <A discussion of these points was given in another 
section of this paper. 


EK. EFFECT OF OXYGEN 


Spaeth (’13) showed that the isolated melanophores of Fundu- 
lus contract in the absence of oxygen, but contracted melano- 
phores do not expand when oxygen is the only stimulating 
agent present. Fick (90) deprived dark-adapted frogs of 
oxygen by submergence in water or by introducing them into 
an atmosphere of hydrogen or carbon dioxide. As a result of 
this treatment he asserts that the retinal pigment underwent 
expansion. Dittler (07) states that in frogs which are about 
to hibernate the cones are never as fully elongated as in active 
animals; but after subjection to an atmosphere of pure oxygen 
the cones can again be obtained in the maximal dark position. 
It seems probable that in this case the effect of oxygen was 
indirect, and the increased activity of the cone cells accompanied 
pari passu the return of other body activities. 

In order to test whether or not the amount of oxygen avail- 
able to a fish in any way controls the distribution of its retinal 


MOVEMENTS IN THE VISUAL CELLS 177 


pigment, a series of experiments, chiefly upon Ameiurus, were 
performed. 

In experiments involving a reduced oxygen supply, the ap- 
paratus was simple. A 33 litre flask was filled with boiling 
water. The flask was then closed with a three-hole rubber 
stopper through which passed, (1) a glass tube extending to 
the bottom of the flask, which served to introduce gas from a 
hydrogen generator, (2) a glass overflow tube extending about 
three-quarters of the way down the flask, which served chiefly 
as an outlet for the hydrogen gas, (3) a mercury pressure regu- 
lator. As soon as the flask of boiling water was stoppered, the 
hydrogen supply was turned on and as a result, water was 
forced to escape through the overflow tube, its place being taken 
by hydrogen gas. When the water level reached the bottom 
of the overflow tube no more escaped, but the gas after bubbling 
through the water did do so and was conducted through a water 
trap to the outside air. As the water cooled down to room 
temperature it could not take up oxygen since none was present, 
and furthermore, the bubbling hydrogen gas tended to expel 
mechanically any residual oxygen present in the boiled water. 

Water containing an excess of oxygen was prepared by bub- 
bling oxygen gas through water in a flask similar to that de- 
scribed in the former experiment, whence it escaped by means 
of an overflow tube leading into a water trap. The water used 
had previously been boiled and reoxygenated by an aquarium 
aerating device. 

Quantitative determinations of the oxygen content were made 
at the expiration of all experiments by the method of Winkler 
(Treadwell and Hall, ’05). 

Ameiurus was used in most of the experimentation, although 
Abramis served in a few cases. The description which follows 
applies particularly to Ameiurus. 

It was possible to reduce the oxygen supply to an amount 
in which the fish could not live.* This, for example, happened 


16 The normal oxygen saturation of water at 20°C. is 6.356 cc. per liter (Tread- 
well and Hall, ’05). Boiled water which had been cooled rapidly was found 
to contain about 0.93 ec. per liter. 


L178 LESLIE B. AREY 


when only 0.8 cc. of oxygen per litre was present. On the 
other hand, in water containing an excess of oxygen (7.5 ce. 
per litre) respiratory movements of the operculum ceased, 
the fins appeared reddish in color and respiration may have 
been largely cutaneous. 

In parallel experiments conducted both in the dark and at 
various light intensities, no difference could be detected in 
the positions of the pigment or visual cells under the extreme 
conditions of oxygen supply. This is not surprising, for pre- 
sumably but little oxygen is needed to permit the cells to func- 
tion, and since for the success of the experiment, the animal 
must have enough oxygen with which to keep itself alive, a 
crucial test involving a complete elimination of oxygen is not 
possible. Since the pigment can not be made to contract in 
excised eyes, but only to expand, a decisive experiment in which 
all oxygen might in this way be eliminated (similar to Spaeth’s 
work on isolated chromatophores) was impossible. 

Pigment, rods, and cones respond in a normal fashion when 
brought from darkness to light or vice versa in water contain- 
ing the minimum oxygen content, about 0.9-1.0 ec. per litre, 
in which the Ameiurus can live. 

Since no indication was observed of a tendency toward ex- 
pansion in the retinal pigment cells of fishes which were deprived 
of oxygen, it is evident that the expansion described by Fick 
(90) in dark-adapted frogs whose respiration rate had been re- 
duced by covering the head with a velvet hood, is exceptional. 
In view of the well known respiratory function of the frog’s skin 
it is possible that Fick’s results are open to other interpretations, 
especially since his experiment, a repetition of the earlier work 
of Englemann (’85), is not in agreement with the latter’s con- 
clusion relative to the absence of movement in the retinal ele- 
ments when frogs provided with velvet hoods were retained in 
the dark as controls to other experiments. 

The chief value, therefore, of the work done by me is to show 
that within normal experimental limits the retinal pigment 
and visual cells of fishes are not affected by an increased or 
diminished oxygen supply. 


MOVEMENTS IN THE VISUAL CELLS 179 


F. INTERRELATION OF INTEGUMENTARY PHOTO-RECEPTORS AND 
RETINAL ELEMENTS 


The skin of several lower vertebrates has been shown to be 
sensitive to light. Among the fishes, Eigenmann (’00) stated 
that certain blind forms living in caves gave motor responses 
when stimulated by light, the photo-receptors presumably being 
located in the skin. Parker (’05) followed up some negative 
results obtained by one of his students on Fundulus by an investi- 
gation on ammocoetes, and proved that the integumentary 
nerves were sensitive to light, causing movements of the animal 
that were both ‘phototropic and photodynamic.’ A_photo- 
‘receptivity of the skin of certain other vertebrates was first 
demonstrated by the following workers: Graber (’84) on Triton; 
Dubois (90) on Proteus; Kordnyi (’92) on the frog; Carleton 
(03) on Anolis; and Eycleshymer (’08) on Necturus. 

Englemann (’85) covered the heads of dark-adapted frogs 
with a velvet cap and exposed the bodies to sunlight. Under 
these conditions, he asserted that in 15 minutes the pigment 
and cone cells assumed the maximal light position, whereas 
the same elements in control experiments conducted in the dark 
remained unchanged. Illumination of the skin for longer periods 
was said to result in a falling off (herabsteigen’) in the expansion 
of the retinal pigment and to a weakened response on the part 
of the cones. From these results he concluded (p. 507): “‘. . dass 
Zapfen und Pigment des Auges von entfernten Kérpergegenden 
aus reflectorisch in Bewegung gebracht werden kénnen.”’ 

Fick (’90), in repeating this experiment of Englemann, 
found the pigment in an expanded condition while the frogs, 
ready for the test, were still in the dark, and after supplementary 
experimentation of various kinds he decided that pigment ex- 
pansion accompanied disturbed respiration. In the case under 
consideration the velvet hood was supposed to have caused partial 
asphyxiation. 

Koranyi (’92) refers to the similarity in the responses of the 
retinal pigment resulting either from the illumination of the 
retina or of the skin only, yet he does not state that he actually 
observed this condition himself. 


180 LESLIE B. AREY 


More recently Fujita (’11) asserted that after the head 
and forward extremities were bound with wet black cloth and 
the rest of the body was exposed to sunlight for 15 to 20 minutes, 
the ‘eyes’ remained in the dark-adapted condition. 

I, myself, before learning of Fujita’s work, had performed 
several experiments of the same kind. The head and fore body 
of dark-adapted frogs were bandaged with many thicknesses 
of black velvet and the remainder of the body and the hind legs 
were exposed to direct or diffuse sunlight for periods of 15 minutes 
to 1 hour. Care was taken to keep the skin moist and to 
guard against the heating tendency of direct sunlight by using 
a heat filter. Although the animals were in an active condition 
at the end of the experiment, there was in no case a distinct 
change in the position of the cone cells or retinal pigment as 
Englemann maintained. 

In an unpublished seen by Mr. 8. G. Wright, no 
direct responses were observed when Ameiurus, from which 
the eyes had been removed, were illuminated with the light 
of an electric are. The normal fish, as is well known, is a night 
feeder, yet it also frequented equally the light and dark halves 
of an aquarium Jar. 

It is, conceivable, however, that the soft skin of this animal 
contains a photoreceptive mechanism, even though the motor 
responses to light fail to indicate its presence. One of the pos- 
sible ways in which an integumentary photosensitivity could 
be manifested is through an influence on the position of the 
retinal elements, similar to the relation which Englemann be- 
lieved to exist in the frog. In view of such a possibility several 
experiments were performed in which dark-adapted fish with 
bandaged heads were exposed to daylight and to the light of 
an electric arc for periods of 45 minutes to 1 hour. In no 
case, however, was there the slightest tendency toward movement 
on the part of the rods, cones, or retinal pigment. 

These results indicate that neither in the frog nor in Ameiurus 
are movements of the retinal pigment or visual cells evoked 
by a codperation of dermal photosensitivity and ‘retino-motor’ 
nerve fibers. Consequently, Englemann’s experiment does not 


MOVEMENTS IN THE VISUAL CELLS 181 


have the significance that he believed it possessed, when he 
attempted to furnish physiological proof that in the frog: “‘Jeden- 
falls aber laufen . . . . auch retinomotorische Fasern von 
den grossen Nervencentren aus durch den Sehnerv zum Auge” 
(85, p. 506). 


4. DISCUSSION 


We have so accustomed ourselves to view the phenomena 
exhibited by living organisms from the evolutionary standpoint 
that an ‘explanation’ which will reveal the adaptiveness of an 
organism to its environment is demanded whenever a system 
of relations involving constant responses to definite stimulating 
agents is discovered. 

To make dogmatic assertions regarding the presence or ab- 
sence of adaptation in a set of responses is, obviously, a matter 
of exceeding danger, yet if the phenomena exhibited by a series 
of representative animals to definite stimulating agents are 
shown to be variable, it is at least evident that a single inclusive 
explanation will not be forthcoming. 

The writer has attempted to show elsewhere (Arey 15) that 
the discontinuous occurrence of photomechanical responses in 
the visual cells and retinal pigment, both in the various verte- 
brate classes and among different representatives of certain 
individual classes, renders it extremely difficult from the adapta- 
tional standpoint to devise a satisfactory explanation for the 
meaning of these movements. The majority of theories which 
attempt to lnk the known responses of the retinal ele- 
ments with the mechanism of light perception are, without 
doubt, highly speculative, and since for the most part they lack 
an experimental basis of any kind, they must remain of inter- 
est only as ingenious and interesting possibilities. In the case 
of the retinal pigment at least, it is possible to compare the 
responses to light and darkness with those exhibited by melano- 
phores in general (Parker ’06, p. 413), and from our present 
knowledge we are unable to see either in the reactions of the 
retinal pigment or in those of the rods and cones anything more 


182 LESLIE B. AREY 


than the presence of constant protoplasmic responses to a definite 
stimulating agent. 

When the effect of temperature is considered, a lack of uni- 
formity at once becomes apparent. 

Among the fishes the effect of temperature upon the retinal 
pigment is in agreement with Parker’s (’06) general conclusion 
that in all melanophores, low temperature has the same effect as 
light and high temperature the same effect as darkness.!° In 
the retina of the living frog where this statement holds true only 
between the temperatures of 0° and 19°C. in the dark, notwith- 
standing the identical photic responses of the retinal pigment 
in fishes and amphibians, it is probable that a nervous control 
is responsible for the expanded condition both at low and at 
high temperatures; hence the responses in this animal are not 
comparable to those in fishes. Since only a limited temperature 
reaction occurs among fishes, and this in darkness as well as 
in light, it probably has no immediate adaptive significance, 
but merely represents the survival of a tendency shown in 
the responses of melanophores in general.'?7 In homothermous 
animals temperature, of course, can play no part in the normal 
activity of the retinal pigment, and even if the temperature 
responses in poikilothermous animals have an adaptive signifi- 
cance, it is evident that this particular advantage must be un- 
available to warm blooded vertebrates, in which keenness of 
sight is best developed. 

Moreover, the action of temperature upon the cone cells in 
the dark is variable. In the fishes, the cone myoids greatly 
elongate when warmed and shorten when cooled, while in the 
frog the cones are maximally shortened at high temperatures 
only and at all other temperatures remain unchanged. 

The inconsistent action of temperature in producing move- 
ments of the visual cells apparently has no common adaptive 

16 Tt should be pointed out, however, that in the case of the melanophores of 
the frog’s skin, the reverse of this statement is true, since high temperature pro- 
duces similar effects to light, and low temperature to darkness. 

17 Below 38°C. temperature has practically no effect on the rate of decom- 


position of the visual purple, as Kiihne (’79) showed, hence the changes of the 
pigment through the action of temperature are not related to this phenomenon. 


MOVEMENTS IN THE VISUAL CELLS 183 


value, and, moreover, as in the case of retinal pigment, variable 
temperature can play no part in the normal movements of these 
cells in homothermous animals. 

It may be asked how it happens that high temperature in 
the dark produces diametrically opposed results upon the cone 
myoids of fishes and of the frog, when the responses to light 
are identical. The nervous system, presumably, is not involved 
in these reactions since temperature has a similar effect upon 
excised eyes. Between the minimal and optimal limits the 
movement of undifferentiated protoplasm is uniformly acceler- 
ated with increased temperature, and as has been shown in 
another place (p. 135) the responses in length of the cones of 
Abramis are markedly similar in this respect. Whatever the 
details of the process may be, it seems evident that among the 
fishes temperature acts purely as a physical agent in controlling 
the velocity of the reactions leading to positional changes of the 
cones. In frogs temperature apparently acts in an entirely 
different manner. The fact that only high temperature is 
effective in producing a change, a shortening of the myoid, is 
best explained on the basis of Dittler’s theory of chemical stimu- 
lation, whereby increased temperature can be conceived of as 
favoring the formation of catabolic wastes which chemically 
stimulate the cones to shorten, while low temperature probably 
acts both in retarding catabolism, and by reducing the sensitivity 
of the myoid toward such products as are formed even under 
these conditions. 

A comparative study of the responses of the visual cells, 
throughout the various classes, to light and temperature re- 
veals difficulties in explaining the mechanism by which their 
positional changes are accomplished. Why do the rod cells 
of some animals shorten in the light whereas others lengthen? 
Why do the rods of fishes and of birds lengthen through the 
action of light, whereas the cones shorten? 

From the physiology of simple protoplasm, two alternatives 
are open. Either the myoids of the visual cells have become 
specialized to respond to certain stimulating agents in different 
ways, or, in various retinas the morphological structure of these 


184 LESLIE B. AREY 


contractile regions is such, that although the protoplasm re- 
sponds similarly in all cases, the visible result upon the position 
of the entire cell is variable. The latter alternative would be 
realized if the active protoplasm were differentiated into ana- 
logues of myofibrillae which were arranged in some cases 
axially in the myoid and in other cases transversely or spirally. 
To become effective in elongating the myoid through contraction, 
a simple spiral ‘myofibril’ would have to make an angle greater 
than 45 degrees with the long axis of the myoid.'!8 

In attempting to interpret the adaptiveness of the movements 
of the vertebrate retinal elements, it is evident from the fore- 
going discussion that neither with respect to the action of tem- 
perature nor of light has a satisfactory and constructive conclu- 
sion been reached. From the present state of our knowledge, 
therefore, the situation may be summarized in the following 
way. Although the movements of the visual cells and retinal 
pigment, when present, may have a certain unknown significance 
in connection with the mechanism of light perception, such 
movements can be interpreted at present only in terms of proto- 
plasmic responses to definite stimulating agents. 


5. SUMMARY 


1. The retinal pigment of the fishes studied requires 45 min- 
utes to 1 hour for light adaption and from 30 minutes to 1 hour 
for dark adaption. The cones of Abramis assume the light 
(shortened) position in 45 minutes, the dark (elongated) posi- 
tion in 30 minutes. Maximal elongation of the rods of Ameiurus 
in the light occurs in 45 minutes, maximal shortening in dark- 
ness in 30 minutes. 

2. Both in light and in darkness, the retinal pigment of fishes 
shows greater expansion at a low than at a high temperature. 
High temperature is apparently more efficient in causing this 


18 Tn connection with these possibilities mention will be made of only the 
fibrils found by Hesse (’04) in both the inner and outer members of the rods and 
cones of several vertebrates, and of the longitudinal fibrils which Howard (’08) 
was able to trace throughout both the rodsand cones of Necturus. Both workers, 
however, considered such structures as neuroid in character. 


MOVEMENTS IN THE VISUAL CELLS 185 


redistribution of pigment than is low temperature. The results 
obtained from the four fishes studied indicate that extreme 
conditions both of expansion and of contraction are not to be 
found in the retinal pigment of any one fish. 

3. In darkness, the retinal pigment of the frog undergoes 
striking expansion between the temperatures of 0° to 14°C. 
and 19° to 33°C., whereas at the intermediate temperatures of 
14° to 19°C. it is highly contracted. Temperature is without 
effect upon light-adapted retinas. Since the retinal pigment 
of the larvae of Rana catesbiana shows temperature responses 
identical with those characteristic of adults, there is no tem- 
porary larval recapitulation of the responses characteristic of 
fishes. 

4. Light and darkness produce but limited changes in the 
distribution of the retinal pigment of Necturus. No definite 
effect of temperature could be detected. It is probable that the 
peculiar responses found in the frog have been developed within 
the anuran group. 

5. The cone myoids of fishes shorten at low temperatures in 
the dark; at high temperatures they lengthen. Not only is 
elongation progressive, extreme conditions being found at 0° 
and 30°+ C., but also the rate of change is directly proportional 
to the temperature gradient. Temperature is imeffectual in 
the light. 

6. The myoid of the rods of fishes also elongates at high tem- 
peratures and shortens at low temperatures, but the extent of 
change is much less than that of cone cells. 

7. The cone myoid of adult frogs in the dark shortens when 
subjected to a temperature of 19° to 338°C., but remains 
elongated at all lower temperatures. No definite temperature 
responses of the cone myoids were found in the larvae of 
Rana catesbiana. 

8. A correlation between the length of the rod myoid and 
temperature was not detected in the frog either in darkness 
or in light. 

9. The positions of the visual cells of Necturus are not affected 
by temperatures between the limits of 0° and 28°C. 


THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL. 26, NO. 2 


186 LESLIE B. AREY 


10. In excised eyes of the four fishes studied, light causes 
a migration of the retinal pigment in Ameiurus only, whereas 
the pigment of none of the fishes moves in darkness. The rods 
of excised eyes of Ameiurus undergo movements both in the 
light and in the dark, the cones move in the ight only. Neither 
exposure to darkness nor to light produces positional changes 
in the cone cells of excised eyes of Abramis or Fundulus.. Where 
tested, temperature was found to cause movements in the retinal 
pigment and cone cells in the excised eyes of fishes. 

11. It is probable that the absence of responses in the excised 
eyes of fishes is due to an autoanaesthetization caused by the 
accumulation of catabolic products. 

12. Dittler’s theory of the chemical stimulation of the cone 
myoid, propounded to explain the movements of the cone cells 
in isolated frog’s retinas, does not satisfactorily meet many 
conditions found in the responses of the cones of fishes. 

13. The effects of temperature upon the rods, cones, and retinal 
pigment of the excised eyes of fishes are identical with those 
found in living animals, hence it is probable that temperature 
has a direct action upon these elements, its effect being physical 
in the sense that the chemical activity of the protoplasm is 
thereby accelerated to varying degrees. 

14. Temperature has no effect upon the retinal pigment of 
the excised eye of the frog, therefore it 1s plausible that the 
action of temperature in living animals is physiological, whereby 
any adequate stimulus acting through the central nervous sys- 
tem can produce a striking pigment expansion according to the 
principle of specific energies. As in living animals the cone cell 
of the excised frog’s eye responds by a shortening at an elevated 
temperature only; it is probable that temperature acts directly 
upon the cone myoid, for this response, unlike that of the pig- 
ment, can not be interpreted by the principle of specific energies. 

15. Neither in the frog nor m Ameiurus are movements of 
the retinal elements evoked by exposure of the skin only to 
light. Hence the existence of an interrelation between dermal 
photosensitivity and the responses of the retinal elements by 


MOVEMENTS IN THE VISUAL CELLS 187 


means of ‘retino-motor’ nerve fibers, as maintained by Engle- 
mann, is not substantiated. 

16. Within the experimental limits at which fishes can be 
kept alive, the retinal pigment and visual cells are not affected 
by an increased or diminished oxygen supply. 

17. Both in darkness and in light, and in excised as well as 
in normal eyes, carbon dioxide and ether completely check the 
movements of all the retinal elements of fishes. Chloretone and 
urethane, on the contrary, are inefficient in this respect. The 
action of carbon dioxide suggests that this may be the catabolic 
product that in many cases restrains the movements of the 
retinal elements when the circulation of the blood is interrupted. 

18. Although the movements of the visual cells and retinal 
pigment, when present, may have a certain unknown significance 
in connection with the mechanism of light perception, such 
movements can be interpreted at present only in terms of proto- 
plasmic responses to definite stimulating agents. 


Cambridge, Mass., April 10, 1915. 
POSTSCRIPT 


A study of the influence of light on the movements of the 
frog’s rod has just been completed by the writer. Careful meas- 
urements prove that these elements are extended in the light and 
are retracted in darkness. Hence the results of the older workers 
(p. 123), who believed that the photic responses of the frog’s rod- 
myoid are the reverse of those occurring in fishes and birds, are 
not substantiated. 


LSS LESLIE B. AREY 


BIBLIOGRAPHY 
Papers marked with an asterisk have not been accessible in the original. 


AnGceLucci, A. 1878 Histologische Untersuchungen itiber das retinale Pig- 
mentepithel der Wirbelthiere. Arch. f. Anat. u. Physiol., Physiol. 
Abth., Jahrg., 1878, pp. 353-386, 2 Taf. 

*1884a Una nuova teoria sulla visione. Communic. preventiva 
presentata all’ Accad. med. di Roma, 14 Jugho. 

1884b Una nuova teoria sulla visione. Gazetta Medica di Roma, 
1884, pp. 205-210: 217-223. 

1890 Untersuchungen iiber die Sehthatigkeit der Netzhaut und des 
Gehirns. Untersuch. zur Naturlehre d. Menschen u. d. Thiere (Mole- 
schott), Bd. 14, Heft 3, pp. 231-357, 2 Taf. 

*1905 Physiologie générale de oeil. Enecyclopedie frangaise d’oph- 
thalmologie, Tome 2, pp. 1-141. 

ArcoLeo, EK. 1890 Osservazioni sperimentali sugli elementi contratili della 
retina negli animalia sangue freddo. Annali d’Ottalmologia, Anno 19, 
Fasc. 3 e 4, pp. 253-262. 

Arry, L. B. 1915 The occurrence and the significance of photomechanical 
changes in the vertebrate retina—An historical survey. Jour. Comp. 
Neur., vol. 25, no. 6, pp. 535-554. 

Bout, F. 1878 Zusatz zur Mitteil. vom 11. Januar, mitgeteilt in der Sitzung 
vom 19. Februar. Monatsber, d. kénigl. preussischen Akad. d. Wis- 
sensch. zu Berlin aus dem Jahre 1877, pp. 72-74. 

Car.eton, F. C. 1903 The color changes in the skin of the so-called Florida 
chameleon, Anolis carolinensis Cuv. Proce. Amer. Acad. Arts and Sc1., 
vol. 39, no. 10, pp. 259-276, 1 pl. 

CuIARINI, P. 1904a Changements morphologiques, que l’on observe dans la 
rétine des vertebrés par l’action de la lumiére et de ’obscurité. Prem- 
iére Partie. La rétine des poissons et des amphibies. Arch. ital. de 
Biol., Tom. 42, Fasc. 2, pp. 303-322, 4 figs. 

*1904 b Cambiamenti morphologici che si verificano nella retina 
dei vertebrati per azione della luce e dell’ oscurité. Parte 1. La 
retina dei pesci e degli anfibi. Boll. della R. Acead. Med. di Roma, 
Anno 30, pp. 75-110. 

1906 Changements morphologiques qui se produisent dans la rétine 
des vertebrés par l’action de la lumiére et de Vobseurité. Deuxiéme 
Partie. La rétine des reptiles, des oiseaux et des mammiféres. Arch. 
ital. de Biol., Tom. 45, Fasc. 3, pp. 337-382, 8 figs. 

Czerny, V. 1867 Uber Blendung der Netzhaut durch Sonnenlicht. Sit- 
zungsber. d. Wiener Akad. d. Wissensch., Math-naturw. KI., Bd. 56, 
Abth. 2, pp. 409-428. 1 Taf. 

DerutscumMann, R. 1882 Uber die Blendung der Netzhaut durch directes 
Sonnenlicht. Arch. f. Ophthal., Bd. 28, Abth. 3, pp. 241-254. 

Dirrier, R. 1907 Uber die Zapfenkontraktion an der isolierten Froschneiz- 
haut. Arch. f.d. gesam. Physiol., Bd. 117, Heft 5 u. 6, pp. 295-328, 
1 Taf. 


MOVEMENTS IN THE VISUAL CELLS 189 


Dusotis, R. 1890 Sur la perception des radiations lumineuses par la peau, 
chez les Protées aveugles des grottes de la Carniole. Compt. rend. 
Acad. Sci. Paris, Tom. 110, pp. 358-361. 

EIGENMANN, C. H. 1900 The blind-fishes. Biological lectures; Marine Bio- 
logical Laboratory at Woods Holl, 1899, Ginn & Co., Boston, pp. 113- 
126. 

ENGLEMANN, T. W. 1885 Uber Bewegungen der Zapfen und Pigmentzellen 
der Netzhaut unter dem Einfluss des Lichtes und des Nervensystems. 
(Nach 1884 gehaltenen Vortrag). Arch. f. d. gesam. Physiol. Bd. 
35, Heft 10, 11, u. 12, pp. 498-508, 1 Taf. 

Ewa.p, A., UNp Kiune, W. 1877 Untersuchungen iiberden Sehpurpur. Teil 2. 
Entstehung der Retinafarbe. Untersuch. aus. d. Physiol. Inst. d. 
Univ. Heidelberg, Bd. 1, Heft 3, pp. 248-290. 

1878 Untersuchungen iiber den Sehpurpur. Teil 3. Verinderungen 
des Sehpurpurs und der Retina im Leben. Untersuchung. aus. d. 
Physiol. Inst. d. Univ. Heidelberg, Bd. 1, Heft 4. pp. 370-422. 

EycLtesHyMer, A. C. 1908 The reaction to light of the decapitated young 
Necturus. Jour. Comp. Neur., vol. 18, no. 3, pp. 303-308. 

Fick, A. E. 1890 Uber die Ursachen der Pigmentwanderung in der Netzhaut. 
Vierteljahrsschrift d. naturf. Ges. zu Ziirich, Jahrg. 35, pp. 53-86. 

Fusira, H. 1911 Pigmentbewegung und Zapfenkontraction im Dunkelauge 
des Frosches bei Einwirkung verschiedener Reize. Arch. f. vergleich. 
Ophthal., Jahrg. 2, Heft 2, No. 6, pp. 164-179, 7 Taf. 

*Gacuiio, E. 1894 Le modificazioni del pigmento all’oscurita ed alla luce nella 
retina della rana. Arch. di Ottalmologia, vol. 1, pp. 225-233. 

GARTEN, 8S. 1907 Die Veriinderungen der Netzhaut durch Licht. Graefe- 
Saemisch Handbuch der gesam. Augenheilkunde. Leipzig, Aufl. 
2, Bd. 3, Kap. 12, Anhang, 30 pp., 5 Taf., 49 Textfig. 

GrRaBER, V. 1884 Grundlinien zur Erforschung des Helligkeits- und Farbensin- 
nes der Tiere. Prag u. Leipzig, 8 vo, vili + 322 pp. 

Grapenticro, G. 1885 Uber den Einfluss des Lichtes und der Wirme auf die 
Retina des Frosches. Allg. Wiener med. Zeitung, Bd. 30, No. 29 u. 
30, pp. 343-344, u. 353. 

Greerr, R. 1900 Die mikroskopische Anatomie des Sehnerven und der Netz- 
haut. Graefe-Saemisch Handbuch der gesam. Augenheilkunde. Leip- 
zig, Aufl. 2, Bd. 1, Kap. 5, 212 pp., 2 Taf., 53 Textfig. 

*HamBurGER, D. J. 1889 Dorsnijding van der rervus opticus bij Kikvorschen, 
in Verband met de Beweging van Pigment en Kegels in het Netvlies, 
onder den Invloed van Licht en Duister. Onderzoekingen d. Utrech- 
tsche Hoogeschol, Reeks 3, vol. 11, pp. 58-67. 

Herzoc, H. 1905 Experimentelle Untersuchung zur Physiologie der Bewegungs- 
vorgiinge in der Netzhaut. Arch. f. Anat. u. Physiol., Physiol. Abth., 
Jahrg. 1905, Heft 5 u. 6, pp. 418-464, 1 Taf. 

Hesse, R. 1904 Uber den feinern Bau der Staibchen und Zapfen einiger Wir- 
beltiere. Zool. Jahrb., Suppl., Bd. 7, pp. 471-518, 1 Taf. 

Howarp, A. D. 1908 The visual cells in vertebrates, chiefly in Necturus 
maculosus. Jour. Morph., vol. 19, no. 3, pp. 561-631, 5 pls. 


190 LESLIE B. AREY 


KorAnyr, A. vy. 1892 Uber die Reizbarkeit der Froschhaut gegen Licht und 
Wirme. Centralbl. f. Physiol., Bd. 6, pp. 6-8. 

*Ktiunn, W. 1877. Uber den Sehpurpur. Untersuch. a. d. Phyiol. Inst. d. 
Univ. Heidelberg, Bd. 1, pp. 15-104, 1 Taf. 

1879 Chemische Vorgiinge in der Netzhaut. Hermann’s Handbuch 
der Physiologie, Leipzig, Bd. 3, Teil 1, pp. 235-337. 

Leprerer, R. 1908 Verinderungen an den Stibchen der Froschnetzhaut unter 
Einwirkung von Licht und Dunkelheit. Centralbl. f. Physiol., Bd. 
22, No. 24, pp. 762-764. 

*Lopato, G. 1895 Ricerche sulla fisiologia dello strato neuroepitheliale della 
retina. Arch. di Ottalmologia, vol. 3, pp. 141-148. 

Lors, J., AND WasTENEYS, H. 1912 On the adaption of fish (Fundulus) to 
higher temperatures. Jour. Exp. Zool., vol. 12, no. 4, pp. 543-557. 

Mayer, P. 1881 Uber die in der zoologischen Station zu Neapel gebrauch- 
lichen Methoden zur mikroskopischen Untersuchung. Mitth. Zool. 
Stat. Neapel. Bd. 2, pp. 1-27. 

Morano, F. 1872 Die Pigmentschicht der Netzhaut. Arch. f. mikrosk. 
Anat., Bd. 8, pp. 81-91, 1 Taf. 

Miuuer, H. 1856 Anatomisch-physiologische Untersuchungen iiber die Retina 
bei Menschen und Wirbelthieren. Zeitschr. f. wissensch. Zool., Bd. 8, 
Heft 1, pp. 1-122, 2 Taf. 

Ovio, G. 1895 Diun speciale axione della cocain sulla funzione visiva. Annali 
di Ottalmologia, Anno 24, Suppl. al Fase. 4, p. 23. 

Pautmer, 8. C. 1912 The numerical relations of the histological elements in the 
retina of Necturus maculosus (Raf.). Jour. Comp. Neur., vol. 22, 
no. 5, pp. 405-445, 12 fig. 

Parker, G. H. 1905 The stimulation of the integumentary nerves of fishes by 
light. Am. Jour. Physiol., vol. 14, no. 5, pp. 413-420. 

1906 The influence of light and heat on the movement of the melan- 
ophore pigment, especially in lizards. Jour. Exp. Zodél., vol. 3, no. 3, 
pp. 401-414, 3 figs. 

*PrerRGENS, E. 1896 Action de la lumiére sur la rétine. Travaux de Lab. de 
l Inst. Solvay, Tome 2, pp. 1-38. 

1899 Vorgiinge in der Netzhaut bei farbiger Belichtung gleicher 
Intensitit. Zeitschr. f. Augenheilk., Bd. 2, pp. 125-141. 

Poucuet, G. 1876 Des changements de coloration sous l’influence des nerfs. 
Jour. de l Anat. et de la Physiol., pp. 1-90 et 113-165. 

Spagetu, R. A. 1913 The physiology of the chromatophores of fishes. Jour. 
Exp. Zodél., vol. 15, no. 4, pp. 527-585, 4 pl. 

Srort, A. G. H. vaN GenpeREN 1886 Uber Form und Ortsverinderungen 
der Elemente in der Sehzellenschicht nach Beleuchtung. Bericht 
iiber d. 18. Versamm. d. Ophthal. Gesell. zu Heidelberg, 1886, pp. 43-49. 
1887 Mouvements des éléments de la rétine sous l’influence de ia 
lumiére. Arch. néerlandaises des Sciences exact et naturelles, publ. 
p. l. Soe. holland. des Sciences, Tom. 21, Livr. 4, pp. 316-386. 

TREADWELL, F. P. anp Hauu, W. T. 1905 Analytical chemistry, vol. 2. 
Wiley and Sons, New York, 8vo, x1i+654 pp., 96 fig. 


EXPLANATION OF PLATES 


The figures of Plates 1 to 4 are photomicrographs; the figures of Plate 5 were 
drawn with the aid of a camera lucida. 


ABBREVIATIONS 


bac., rod my.con., cone myoid 

con., cone pd.cl.pig., base of pigment cell 
con.acc., accessory cone prs.dst.bac., rod outer member 
ell.bac., rod ellipsoid prs.dst.con., cone outer member 
ell.con., cone ellipsoid rin., retina 

gtt.ol., oil globule scl., sclera 
mb.lim.ex., external limiting membrane — st.nl.ex., external nuclear layer 
my.bac., rod myoid st.pig., pigment layer 


191 


PLATE 1 
EXPLANATION OF FIGURES 


The photographs in this plate, which show the influence of temperature on 
the distribution of the retinal pigment of fishes, are all at a magnification of 170 
diameters. 

1. Ameiurus, 5°C. in the light. 

2. Ameiurus, 25°C. in the light. 


. Ameiurus, 5°C. in the dark. 
. Ameiurus, 25°C. in the dark. 


. Fundulus, 5°C. in the light. 
. Fundulus, 25°C. in the light. 
. Fundulus, 5°C. in the dark. 
. Fundulus, 25°C. in the dark. 


CcOonT mS Ot Re 


MOVEMENTS IN THE VISUAL CELLS PLATE 1 
LESLIE B. AREY 


__-st. bac. con: 


_-mb. lim. ex. 


pd. cl. pig:-- 


a ae _-sbthe 


nen ABE. LEC COM eae a 


~~~ mb. lim. ex---~ 


~~rins--~ 


193 


PILATE: 2 
EXPLANATION OF FIGURES 


The photographs in this plate, which show the influence of temperature on 
the distribution of the retinal pigment of fishes, are all at a magnification of 
170 diameters. 

9 Abramis, 5°C. in the light. 

10 Abramis, 25°C. in the light. 

11 Abramis, 5°C. in the dark. 

12 Abramis, 25°C. in the dark. 

13. Carassius, 5°C. in the light. 

14 Carassius, 25°C. in the light. 

15 Carassius, 5°C. in the dark. 

16 Carassius, 25°C. in the dark. 


194 


MOVEMENTS IN THE VISUAL CELLS PLATE 2 
LESLIE B. AREY 


— ~ 


- mb. lim CT >~__ 


EIS Wb | Sp Be So 


st. bac. con; 


tI [2 


, 
scl, 


pd, el mg: 


sl. pigs 


rin; 


scl 


pd. cl. ptg: 


st. ng: 


mb. lim. ex.-- 


rin.- 


195 


PLATE 3 
EXPLANATION OF FIGURES 


Figures 17 to 22, which are photographs showing the influence of temperature 
on the distribution of the retinal pigment of the frog (adults and larvae), are all 
at a magnification of 170 diameters. The larval Rana catesbiana, from which 
figures 20 to 22 were made, had a total body length of 7.0 em. A Leitz ;': homo- 
geneous immersion objective was used in making figures 23 and 24, which are 
magnified 715 diameters. 

17 R. pipiens (adult), 3°C. in the dark. 

18 R. pipiens (adult), 18°C. in the dark. 

19 R. pipiens (adult), 33°C. in the dark. 

20 R. catesbiana (larva), 3°C. in the dark. 

21 R. catesbiana (larva), 16°C. in the dark. 

22 R. catesbiana (larva), 32°C. in the dark. 

23 Carassius; 27°C. in the dark. 

24 Carassius, 3°C. in the dark. 


196 


MOVEMENTS IN THE VISUAL CELLS PLATE 3 
LESLIE B. AREY 


- a > 
---- TENs--__ : vgs AS: See 
——_ 


scl: 
st. [tyz-- 


Cp aa oF st.bac. con. 


oe Deere wp 
<2 eee 


ene SER rin; 
be ue dk ~, a 

md, ee ‘ 

<% x 


scl; 


ral aaah ae BAL, 


sl. hac. con: 


. ? . r) cra 
19 292 
. zi 


mb. lim. ex. 


Sab Tes OX 


23 24 


197 


PLATE 4 


EXPLANATION OF FIGURES 


These photographs, taken with a Leitz y's homogeneous immersion objective, 
are at a magnification of 715 diameters and show the responses of the myoids of 
the cone cells of fishes when under the influence of temperature. 

25 Abramis, 3°C. in the dark. 

26 Abramis, 15°C. in the dark. 

27 Abramis, 27°C. in the dark. 

28 Fundulus, 3°C. in the dark. 

299 Fundulus, 27°C. in the dark. 


198 


MOVEMENTS IN THE VISUAL CELLS PLATE 4 
LESLIE B. AREY 


---sl, ply. 


_-ell. bac. 


mb. lim.exz> 


ell. con 
st. nl. ex, 


con, ace, 


~~mb. lim. ex 


~~-sl. nl, ex 


{ : i 
ee 


mb. lum. ex 


rem MN gare a 
j , M jé ig. y, 
i . iy’ ; 


cll. con, 


ell. con.- 


--con, acc, 


ell. bacs.. a ’ u 
7. . . ; : 


mb. lim. ex. ci " . 
CON. ACCz -- --=-===-=--- 


mb. lim. ez. 


---st. nl, ex. bc Page od or 
st. nl. ex~-- = 


af 29 


199 


PLATE 5 
EXPLANATION OF FIGURES 


All drawings in this plate are at a magnification of 930 diameters, a Leitz 
js homogeneous immersion objective being used. 

30 Showing the positions of the rods and cones in a typical light-adapted 
retina of Ameiurus. 

31 Showing the positions of the rods and cones in a typical dark-adapted 
retina of Ameiurus. 

32 Showing the effect of low temperature (8°C.) on the position of the dark- 
adapted rod- and cone-cells of Ameiurus. 

33. Showing the effect of high temperature (27°C.) on the position of the dark- 
adapted rod- and cone-cells of Ameiurus. 

34 From the retina of a Rana pipiens that had been previously kept at a 
temperature of 3°C. in the dark. The cone myoids remain elongated as at an 
intermediate temperature. 

35 From the retina of a Rana pipiens that had previously been kept at 18°C. 
in the dark. The cone myoids are elongated. 

36 From the retina of a Rana pipiens that had been kept at 33°C. in the 
dark, showing the resulting shortening of the cone myoids. 


200 


MOVEMENTS IN THE VISUAL CELLS PLATE 5 
LESLIE B. AREY 
_-prs.dst.bac. 
we prs.dst.bac. 


prs.dst.con:------ 
prs.dst.con 
--.ell.bac ell.con------ 
a gtt.ol 
Penn prs.dst.bac.. 
cee ell. bac. 
my .con------ 
oo my .bac <3 
----prs.dst.con, = es fee aaery 
inl ~~mb.lim.ex. 
st.nl.ex.----- 
y---H----ell.con. 34 
ell.bac:---- 
---my.bac. prs.dst.bac.. my.bac:----- 
prs.dst.con.. 
---MY.con, 
ell.bac 


my.bac. _ Px 


= “nb.lim.ex. 
~-st.nlex, 


gtt.ol---- a 
ellcon.—¥, 


---.my.con, 


prs.dst.bac. +---ell.bac. 
my .con:--- 

SS _ell.bac. . 7 
Sa mb.lim.ex: 


---my.bac. 


ell. bac. --4 : 


SES my bac. __- 
\ -----my.bac. 


~--mb.lim.ex, ; 
~~~st.nl.ex. 3 5 mb.lim.exs 


sstnlets 


201 


THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL. 26, No. 2 


REGENERATION IN THE BRAIN OF AMBLYSTOMA 


I. THE REGENERATION OF THE FOREBRAIN 


H. SAXTON BURR 


The Anatomical Laboratory of The School of Medicine, Yale University 
FOUR FIGURES 


Up to the present time the data dealing with regeneration in 
the central nervous system have been exceedingly conflicting. 
The early workers reported the regeneration of various definitive 
parts after they were extirpated. As far back as 1890 Danielew- 
sky found that the removal of the cerebral hemispheres of the 
frog resulted in the formation of a ‘cerebral mass’ which he 
believed contained embryonic nerve cells, though this mass 
was in no way a new hemisphere. More recently Bell, in 1907, 
removed the ‘lateral half’ of the brain of the frog and found 
that the brain nearly always is reformed though never does it 
reach the normal size. 

On the other hand, Schaper (’98) found that the removal of 
portions or of the entire brain of Rana esculenta was never 
followed by regeneration. Rubin in 1903 working with Rana 
fusca larvae found that no regeneration occurred after the re- 
moval of the brain or of parts of it. 

The results of experiments published elsewhere (Burr ‘16) 
show conclusively that the nasal placode of Amblystoma and 
also of the frog does not regenerate when it is completely re- 
moved. In addition, the large amount of work by Lewis and 
others has shown that the eye will not regenerate when all of 
the anlage is removed. It was deemed probable, therefore, 
that complete extirpation of the cerebral hemisphere would not 
be followed by even a partial regeneration of the part removed. 
The discrepancy in the results previously reported might be 
due to incomplete operations. The regeneration of parts of 

203 


204 H. SAXTON BURR 


the brain would then be due to the fact that the entire substance 
of the hemisphere was not removed, enough being left to carry 
the regeneration through to a considerable degree of complete- 
ness. A condition similar to this has been shown by Harrison 
to exist in the extirpation of limb buds in Amblystoma (Harri- 
son 715). 

So far as it has been possible to ascertain, no attempt has 
previously been made to control the stimulus afforded by the 
functional activity of the end organ normally connected with 
the part of the central nervous system removed. Extirpation 
of a portion of the brain has invariably carried with it the re- 
moval of the end organ. It is a little difficult then to see why 
a part should regenerate when its activity has ceased owing to 
the removal of the end organ. 

In order to obtain some answer to the above question the 
following experiments were performed on Amblystoma larvae. 
Two series of operations were undertaken on embryos possess- 
ing neither peripheral nerves nor a circulatory system. In 
the first of these the right cerebral hemisphere and the right 
nasal plaéode were extirpated, the cut that severed the hemi- 
sphere from its connections passing directly in front of the optic 
stalk. In the second series the right telencephalon was removed 
but the right nasal placode was left in position. This was ac- 
complished by turning back the flap of skin containing the pla- 
code and removing the underlying forebrain. The flap was 
then returned to its former position and held in place until the 
wound had healed. Great care was taken to remove all of the 
cells of the cerebral hemisphere in all the operations. 

The above experiments subject the brain tissue left by the 
extirpation of the hemisphere to two conditions. In the first 
series of operations the nervous tissue is left to regenerate with- 
out the possibility of any stimulus from the end organ that is 
normally connected with it. In the second series the end organ, 
the nasal placode, is left in its normal position and may there- 
fore act as a stimulus to the nervous tissue (Burr 716). 

In the first series of experiments in which the right hemisphere 
and the right nasal placode were removed the wound usually 


BRAIN REGENERATION IN AMBLYSTOMA 205 


healed within the first twenty-four hours. Five days after the 
operation a new wall had formed connecting the wall of the 
right diencephalon with the wall of the left telencephalon. At 
first this wall consists of a narrow band of new cells bridging 
the interventricular foramen. As growth proceeds the narrow 


Fig. 1 Transverse section of embryo five days after operation, by which 
the cerebral hemisphere was removed, leaving the nasal placode in place, show- 
ing curtain of cells across the interventricular foramen. X 50. 


band of cells is drawn out into a thin plate, never more than two 
or three cells thick, stretching across the foramen (figs. 1 and 3). 

A eareful inspection of a series of operated larvae makes it 
quite clear that the new tissue thus formed is derived from the 
primary ependymal cells that line the neural tube. Figure 2 
drawn from a section of an operated larva of the second series 
shows the origin of this new tissue from the margin of the dien- 


206 H. SAXTON BURR 


cephalon. It is composed entirely of the typical columnar 
cells that make up the ependyma. As growth proceeds the 
cells of the new membrane lose their columnar shape and _ be- 
come metamorphosed into flattened quadrilateral cells. In the 
three months old larva, a section of whose brain is shown in 
figure 3, it is evident that this metamorphosis is accompanied 
by a thinning of the tissue so that at this time the membrane is 


Fig. 2. Portion of the regenerated curtain of an embryo seven days after 
the same operation as in figure 1, showing regeneration from the margin of the 
diencephalon. 50. 


composed of only a single layer of cells whose somewhat elon- 
gated nuclei are parallel to the surface. 

It is evident from the above that the regenerated portion is 
not made up of nerve cells but rather of the primitive germinal 
epithelium which lines the neural tube, the only regeneration 
that takes place being that necessary to close the wound made 
by the removal of the hemisphere. We have then, forming as 
a result of the operation, a curtain of primary ependyma across 
the interventricular foramen. This is not nervous and hence 
constitutes no true regeneration of the telencephalon. 


BRAIN REGENERATION IN AMBLYSTOMA 207 


Hardesty (04) has shown that the dividing elements of the 
primary ependyma, the germinal cells, give rise to a nuclear 
layer just outside of the ependyma from which later develop 
the neuroblasts and the spongioblasts. It is conceivable that 
the differentiation of the nuclear layer into nerve cells results 
from the stimulus derived from the ingrowth of nerve fibers 


eX ry ay 
O09". e | 


LR 
are) 


Fig. 3. Transverse section of embryo without the right hemisphere or nasal 
placode three months after the operation, showing the thin curtain across the 
interventricular foramen. X 50. 


from other definitive parts of the nervous system. Evidence 
of this is seen in the fact that the fiber tracts of the telencephalon 
do not develop until the olfactory nerve fibers have grown into 
the peripheral part of the hemisphere from the nasal placode. 
The possibility of such a stimulus has been used by Kappers 
(14) in his neuro-biotaxis theory of the phylogenetic migration 
of nuclear centers in the central nervous system. 


208 H. SAXTON BURR 


On the other hand Harrison (10) has shown that embryonic 
nerve cells taken from the medullary cord of frog larvae will 
produce protoplasmic nerve processes when cultivated in vitro, 
but it is manifestly impossible to determine by such methods 
the extent to which development will proceed without the inter- 
vention of functional activity. As has been shown elsewhere 
(Burr 716), development and differentiation of the central ner- 
vous system will progress to a certain point without the presence 
of a functioning end organ. Beyond this poimt functional 
activity is essential for continued growth. In the experiments 
just referred to, the telencephalon was quite completely organ- 
ized without the ingrowth of the olfactory fibers, but this nerve 
is not the only one whose fibers penetrate the hemisphere, for 
Herrick (10) has shown that in the posterior third of the telen- 
cephalon of Amblystoma there are at least two centers that are 
directly connected with the hypothalamus and the pars dorsalis 
thalami by ascending projection fibers. Hence it seems probable 
that the apparent self-differentiation of the telencephalon may 
be due in some degree to the presence of the forward growing 
nerve fibers which establish this connection, the functional stim- 
ulus imparted by them being sufficient to carry differentiation 
some distance. Experimental evidence that is in favor of this 
view is seen in the fact that in the posterior margin of the cur- 
tain which develops across the foramen of Monro on the re- 
moval of the telencephalon and nasal placode, there appears in 
the older larvae a rounded mass of nerve cells and fibers. This 
mass is in direct communication with the nucleus habenulae 
and the hypothalamus, a tract of fibers reaching it from each 
of these regions. Evidently the neuroblasts at the posterior 
margin of the interventricular foramen have been stimulated 
to further development by the ingrowth of axones from lower 
centers. This mass is, then, in all probability the rudiment 
of the primitive pallium. 

The second series of experiments involved the removal of the 
right hemisphere without the removal of the nasal placode. 
The results here are definite and conclusive. <A curtain of epen- 
dymal cells is formed across the foramen of Monro as in the 


BRAIN REGENERATION IN AMBLYSTOMA 209 


previous experiment. The healing of the wound brings the 
nasal placode back to approximately its normal position though 
a slight shifting anteriorly has been noted. Five days after 
the operation the curtain is composed of a single layer of cells 
(fig. 1). Two days later this single layer of cells has become 
converted into a thickened mass of columnar cells that are evi- 
dently of ependymal origin (fg. 2). By the time the twelfth 
day is reached the olfactory nerve has established a connection 
with the curtain, which has in the mean time increased con- 
siderably in size. The establishment of the nervous connection 
between the nasal placode and the telencephalon takes place 
in the normal larva about the tenth day, only a slight delay 
in this union occurring as a result of the operation. In the 
unoperated larvae the nasal placode lies in close apposition to 
the thick wall of the telencephalon, the olfactory fibers growing 
into the hemisphere at the point of contact. In the operated 
forms the nasal anlage lies close to the newly formed curtain, 
which as indicated above shows a slight thickening not seen in 
the operated larvae possessing no nasal placode. This thicken- 
ing anticipates the ingrowth of the olfactory nerve by several 
days. It is not possible at this time to give any explanation 
of this thickening, though it is conceivable that the presence of 
the olfactory anlage stimulates the ependymal cells to grow and 
divide through some mechanical or chemical factor. 

Two days after the ingrowth of the olfactory nerve the new 
telencephalon, as it must now be called, has reached quite an 
advanced stage of development. During these two days the 
thickened curtain has been differentiated into a small but never- 
theless typical telencephalon. Neuroblasts are present and 
a ventricle has appeared. Outwardly the form has changed. 
A forward growth has converted the thick flat plate of cells 
into an ovoid mass of neuroblasts and spongioblasts. The 
ingrowth of the olfactory nerve on the lateral aspect of the mass 
is accompanied by the appearance of the large lateral forebrain 
tract (fig. 4). Almost simultaneously there appears from the 
dorsal aspect of the olfactory bulb the tractus olfactorius dorso- 
lateralis and from the ventral aspect the tractus olfactorius 


210 H. SAXTON BURR 


medialis. At this time the new telencephalon is similar in or- 
ganization to its fellow but it is considerably smaller in. size. 
From this time on growth and differentiation continue in both 
hemispheres in an entirely normal fashion, though apparently 
there is to some extent a compensating increase in the rate of 
erowth of the new telencephalon for in the oldest larva of the 


POO S EOS: 


Fig. 4 Transverse section of an embryo one and one-half months after the 
operation showing the regenerating telencephalon as a result of the presence of 
the nasal placode. X 44. 


series, a larva some three and a half months old, the new hemi- 
sphere had completely regenerated and could not be in any way 
differentiated from its fellow. 


SUMMARY 


It is evident from the above experiments that the forebrain 
of Amblystoma will not regenerate when it and its functional 
end organ are completely extirpated. The healing of the wound 


BRAIN REGENERATION IN AMBLYSTOMA ot 


results in the formation of a curtain across the interventricular 
foramen derived from the ependyma lining the neural tube. 
If, on the other hand, the forebrain is removed without remov- 
ing the end organ, the nasal placode, the presence of the latter 
acts as a stimulus to the regeneration of a new telencephalon 
through the ingrowth of the olfactory nerve. The pallial re- 
gion of the telencephalon is regenerated in all cases owing to 
the stimulus afforded by the forward growth of axones from 
lower centers in the brain and cord. 


LITERATURE CITED 


Beut, E. T. 1907 Some experiments on the development and regeneration of 
the eye and nasal organ in frog embryos. Arch. f. Entw.-Mech., 
Bd. 23, S. 457. 

Burr, H. S. 1916 The effect of the removal of the nasal placodes on Ambly- 
stoma embryos. Jour. Exp. Zoél., vol. 20, p. 27. 

DanieELEwsky, B. 1890 Uber die Regeneration der Grosshirn Hemisphiren 
beim Frosch. Verh. X. inter. med. Kongr., Bd. 2. 

Harpesty, I. 1904 On the development and nature of neuroglia. Am. Jour. 
Anat., vol. 18, p. 229. 

Harrison, R. G. 1910 The outgrowth of the nerve fiber as a mode of proto- 
plasmic movement. Jour. Exp. Zodél., vol. 9, p. 787. 

Herrick, C. J. 1910 The morphology of the forebrain in Amphibia and Rep- 
tilia. Jour. Comp. Neur., vol. 20, p. 413. 

Kaprrers, C. U. Arriins. 1914 Weitere Mitteilungen iiber Neurobiotaxis. 
VIII. Folia-Neurobio., Bd. 8, S. 383 

Lewis, W. H. 1904 Experimental studies on the development of the eye in 
Amphibia. Am. Jour. Anat., vol. 3, p. 505. 

Rusrn, R. 1903 Beziehung des Nervensystems zur Regeneration bei Amphibien. 
Inaug. Dissert. Rostock. 

Scuarer, A. 1898 Experimentelle Studien an Amphibienlarven. Arch. f. 
Entw.-Mech., Bd. 6, 8. 151. 


} 


CONTRIBUTIONS FROM THE ZOOLOGICAL LABORATORY OF THE MUSEUM OF COMPARATIVE ZOOLOGY AT 
HARVARD COLLEGE. No. 276. 


THE FUNCTION OF THE EFFERENT FIBERS OF THE 
OPTIC NERVE OF FISHES 


LESLIE B. AREY 
Northwestern University Medical School 


TWELVE FIGURES (TWO PLATES 


CONTENTS 

eR a> Ji. / CEM ele Bia Apia viv a ou we Se sb Se Sage eye RR RIEN As to 213 
Rae UCOL VUNG! DEODUAMING coh v ocap a oh wie sae «a v's 6s oie ody emia ot EEE Weta eS 215 
PECUSEICL RG CCORHICHL TUBUMOUEs ac .n5 osc c css cecsaccc tenes Se aaMEME Es 84 sie 217 
TU CeRPEATIOILURLET TITLE Us cy bc eC K MEE ae ovis doe o's ob ves 6 vivid tla We 0d En an Oe 218 
SEeE SHELIMIETILALLON: WOM AINOLUINUS) oes isi6s cai 5 0k Sib wwe intels Clem ReRINIS  s ere 218 

DE SES DM TASU RDN BFA OTE Sheree gcse veo Nliesa salip iel-s'y 4.» vik, av Kate Adin eee Tor 219 

OSSR CLT A ae ee eee re re 8 PR 230 

b. Experimentation upon Abramis and Fundulus........................ 233 
PERONOTAUEL GOUBICGPAUIONE. oi loys Wc oie va Ves ou oss ove stds cn ee nee eR 234 
a Sy a eee ee tee 239 
Oh a a ie. i re se 240 

PRELIMINARY 


A number of observations have been recorded, chiefly upon 
the frog, which indicate that the retinal pigment and visual 
cells are at least partially under the control of the central nervous 
system, although the manner and extent of this influence are 
not well understood. An interrelation between the retinal ele- 
ments of the two eyes has also been maintained by which stimu- 
lating agents, such as light or salt crystals, applied to one retina 
induce changes! in the other. It has not only been declared by 


1 Extensive experimentation has demonstrated the existence of photomechani- 
cal responses in the retinal pigment of most of the lower vertebrates. In all 
eases in which movements of the retinal pigment are demonstrable, light causes 
an expansion (i.e. a migration toward the external limiting membrane), and dark- 
ness a contraction of the pigment (figs. 1 and 4). A portion of the cone’s inner 
member, to which the appropriate term ‘myoid’ has been applied, is also capable 


. of actively shortening when the retina is stimulated by light, a compensatory 


213 


THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL. 26, NO. 3, 
JUNE, 1916 


214 LESLIE B. AREY 


some workers that this interrelation is independent of the brain, 
but also that it continues after the optic nerve is cut distal to 
the chiasma. The latter assertion, however, has been the sub- 
ject of much controversy. 

Observations on the frog relative to these conditions have 
- been presented by the following workers: Englemann (’85), Grijns 
(91); Nahmmacher (’93); Angelucci (90; ’05); Lodato e Pirrone 
(01); Chiarini (04) and Herzog (05). Similar statements were 
made by Pergens (’96) for fishes, and by Birch-Hirschfeld (’06) 
for the pigeon. The diversity of stimuli which have been reported 
as being effective in producing positional changes in the cones 
and pigment has caused Fick (’89; ’90; 791) to doubt the reflex 
nature of the process, although some of his own experiments by 
no means disprove many of the contentions of the other investi- 
gators. More recently Fujita (11) has shown wherein the older 
experimentation on the interdependence in the responses of the 
elements of the two eyes through the action of light, is not 
trustworthy. 

Englemann (’85), who was the pioneer in asserting the pres- 
ence of ‘retino-motor’ nerve fibers, further supported his view 
with results obtained by illuminating the skin only of dark- 
adapted frogs, whence changes in the cone cells and retinal pig- 
ment were said to occur. This observation, however, is not in 
agreement with those of Fujita (11) and myself (Arey, 716). 

A number of other statements are on record which ascribe a 
control over the movements of these elements to the central 
nervous system. An enumeration of stimulating agents supposed 
to act in this way would include such as the following: noises, 
unilateral pressure on the eyeball, mechanical irritation (Ange- 
lucci, ’90), and trussing up a frog for 24 hours (Herzog, ’05). 

The first direct experimentation in which the relation of the 
optic nerve to movements of the retinal elements was tested 


elongation taking place when the retina is again subjected to darkness (figs. 10 
and 11). In fishes and birds, at least, the rod myoid is likewise contractile (figs. 
10 and 11), although the direction of movement in light and in darkness is the 
exact reverse of that executed by the cones. For a review of the literature on 
this question, reference may be made to an earlier paper by the writer (Arey, 
715). 


EFFERENT FIBERS OF THE OPTIC NERVE 215 


was performed by Hamburger (’89), whose results were corrobo- 
rated by Fick (91). After the optic chiasma of the frog was 
severed, light- and dark-adaption occurred as in normal animals, 
hence the pigment and cone cells were viewed as independent 
structures, responding to direct stimulation, the movements of 
which are not dependent upon the integrity of the optic nerve. 
Arcoleo (’90) likewise found that the retinal elements of the 
pithed toad exhibited photomechanical changes. On the other 
hand, Nahmmacher (’93) showed that stimulation of the frog’s 
optic chiasma with salt crystals induced changes in the cones, 
only if the optic nerve was intact. 

Hence it seems probable that, in the frog, the retinal elements 
are capable of more or less independent movement, but over 
this is superimposed a nervous (efferent) control, the nature of 
which is not altogether evident. 

Apart from the work of Pergens (’96), who believed that the 
illumination of one eye of a fish induced cone contraction in the 
other, practically no attempt has been made to determine the 
conditions under which movements of the retinal elements of 
fishes are accomplished. The negative results of Gertz (11) 
concerning the effect of electrical stimulation of the eyes of 
Abramis brama differ only slightly from those of Fujita (11), 
who, however, maintained that induction shocks caused an insig- 
nificant pigment contraction in the light-adapted eye of the 
‘Weissfisch.’ 


STATEMENT OF THE PROBLEM 


It has been shown that the retinal pigment of most fishes 
expands to a marked degree when the normal animal is brought 
from a situation of darkness to one of light, and that the pig- 
ment again contracts when the procedure is reversed (figs. 1 and 
4). 

The writer (Arey, ’16) has also found that the pigment in the 
excised eyes of Ameiurus, at least, is capable of expanding when 
subjected to light, which, in this case, presumably acts as a direct 
stimulus on the pigment cells; a contraction in darkness, how- 


216 LESLIE B. AREY 


ever, does not occur. In several other fishes (Abramis; Fundulus; 
Carassius) a general pigment migration could not be demon- 
strated under these conditions even in the light, and such a lack 
of response was accounted for by postulating the occurrence of 
an autoanaesthetization of the pigment cells, presumably through 
the accumulation of catabolic wastes. Furthermore, the rods of 
the excised eyes of Ameiurus undergo movements both in the 
light and in the dark, whereas the cones move in the light only; 
yet neither darkness nor light induces positional changes of the 
cones in the excised eyes of certain other fishes (Abramis and 
Fundulus). As will be seen farther on, the total absence of 
response in the excised eyes of these fishes throws important 
light upon the question of a probable autoanaesthetization of the 
retinal elements. 

If the optic nerve transmits afferent impulses exclusively, as 
hitherto has been believed, no disturbance in the movement of 
the retinal pigment should be introduced when the optic nerve 
only is severed, and the retina is thereby freed from this source 
of cranial innervation. The present paper will be devoted to 
the description of experiments devised to test the validity of 
this hypothesis. 

In order that the reader may better interpret the various 
experiments about to be described, as well as the significance of 
their progressive sequence, it is advisable to state in advance 
the thesis which this paper aims to establish. Experimental 
evidence will be presented which offers physiological proof for 
the existence in Ameiurus of two distinct components of a 
mechanism, through the balanced action of which are move- 
ments of the visual cells and retinal pigment alone possible. One 
component involves the efferent nerve fibers of the optic nerve, 
whereas the second component (possibly the ciliary, autonomic, 
nerves) is closely associated with the eye muscles. This latter 
set of nerve fibers exerts a constant. inhibition upon the move- 
ments of the retinal elements, while the impulses in the efferent 
optic nerve fibers, on the other hand, serve only as a block to 
this tonic inhibition, thus allowing photomechanical responses 
to occur. 


EFFERENT FIBERS OF THE OPTIC NERVE 217 


The following paper presents one phase of an investigation 
upon the visual cells and retinal pigment, pursued at Harvard 
University under the supervision of Prof. G. H. Parker. To Pro- 
fessor Parker I am greatly indebted for the continued interest 
and kindly criticism that characterizes his instruction. 


MATERIAL AND TECHNICAL METHODS 


The fishes used in this investigation were as follows: the com- 
mon horned pout, Ameiurus nebulosus Lesueur; the shiner, 
Abramis erysoleucas Mitchill; and the common killifish, Fundu- 
lus heteroclitus Linn. A greater part of the experimentation 
was done upon Ameiurus, since in this animal the activities of 
the retinal pigment and visual cells proved to be especially 
favorable under the experimental conditions imposed upon them. 
Furthermore, the occurrence of responses in the excised eyes of 
Ameiurus was a happy concomitant circumstance, since the 
results were thereby controlled at several important points. 

~The technical methods used in preparing retinas for micro- 
scopical examination were simple. The eyes of Ameiurus were 
excised directly, for the skin of this animal is soft and the eyes 
are prominent. In the two other fishes, especially when rapidity 
of operation was desirable, the following procedure was observed. 
With heavy scissors the cranium was bisected in the sagittal 
plane; following this, a transverse cut just posterior to the orbit 
freed the two halves of the cranium, with the contained eyes, 
from the rest of the body. In either case the operation was per- 
formed in a few seconds, and the eye, without being handled, was 
allowed to drop into the fixing fluid. 

Perenyi’s fluid gave good fixation and was used exclusively. 
Fixatives containing nitric acid have long been recognized as 
favorable in the obtainment of faithful preservation of the retina. 
The preparatory steps prior to embedding in paraffine demand 
generous allowances of time, yet the processes of dehydration 
and clearing should progress as rapidly as possible, since other- - 
wise the sclera becomes extremely hard. 

Two methods were used in removing the lens, one of which, 
although longer, gave much more satisfactory results. The first, 


218 LESLIE B. AREY 


somewhat tedious, procedure consisted in paring away the front 
face of the eyeball with a razor after the eye had previously been 
imbedded in paraffine. After removing the face of the eyeball 
slightly beyond the ora serrata, the lens was pried from its 
paraffine matrix with a dissecting needle; following such manipu- 
lation retmbedding was of course necessary. The second and 
simpler method was to remove the face of the eyeball with fine 
curved scissors after the eye had been hardened in absolute 
alcohol; unless, however, the eye was sufficiently hardened and 
the greatest care exercised, the retina proper easily separated 
from the pigmented epithelium. On the whole, the first method 
was preferred to the second because of the wrinkling of the retina 
that usually accompanied the use of the latter. 

Sections were cut 7p to 10 thick, and only those passing 
through the region of the optic nerve were retained for examina- 
tion. Preparations were stained with Ehrlich-Biondi’s triple 
stain or were double stained in Heidenhain’s iron haematoxylin 
and a plasma counterstain. Ehrlich-Biondi in some instances 
gave excellent differentiation of all elements, while at other 
times it would show the capriciousness for which it is notorious; 
iron haematoxylin gave uniformly good preparations. 

When it became necessary to bleach the pigment, in order to 
study the visual cells, which would otherwise be masked by the 
partially or completely extended processes, the method employed 
was essentially that of Mayer, in which nascent oxygen? is the 
effective agent. 


EXPERIMENTAL PART 
a. Experimentation upon Ameiurus 


Ameiurus is well adapted for operations involving the optic 
nerve. The soft skin and the relatively small eye allow easy 
access to the orbit without causing the serious disturbance and 
shock that is almost unavoidable when operating on fishes which 
have eyes set in prominent bony sockets. 

2 When potassium chlorate and hydrochloric acid interact, it is commonly 


said that nascent chlorine is the agent causing bleaching. As a matter of fact 
the reaction liberates free oxygen. 


EFFERENT FIBERS OF THE OPTIC NERVE 219 


The method of operation was as follows. With curved scissors 
an incision was made in the skin ventral to the eye and the cut 
carried upward around both sides of the ball until a semicircular 
(or greater) incision resulted. The eyeball could now be rolled 
back, and when thus displaced, no excessive strain was exerted 
on the conjoined tissues. By dissecting through the aperture 
thus exposed, the optic nerve was separated from the overlying 
muscles and severed. If a delicate razor-edge scalpel, such as 
ophthalmologists use in cataract operations, is employed, it is 
comparatively easy to cut the nerve without injury to the sur- 
rounding parts. Ameiurus has no large central artery and vein 
in the optic nerve, as is the case in mammals. A blood-vessel, 
however, runs beside the optic nerve, but if care is taken it need 
not be injured by the operation. 

1. Retinal pigment. My first effort was to discover the effects 
upon pigment migration of the severance of the optic nerve only, 
A typical experiment will illustrate the response of the pigment 
when a previously dark-adapted fish was operated on and exposed 
to light. 


Experiment 8.1.12. The optic nerve of a dark-adapted fish was 
severed and the animal allowed to recover from any shock effect until 
the next day, when it was exposed to diffuse daylight for 2} hours. 
At the expiration of this time both the operated and the normal eye 
were excised. Subsequent examination showed that the retinal pig- 
ment in the operated eye had retained the position typical of darkness 
(fig. 4), while the pigment in the control eye had migrated normally 
(fig. 1). 


Thus it is seen that the integrity of the optic nerve must be 
maintained in order that the retinal pigment of an otherwise 
intact animal may undergo a positional change when stimulated 
by light. 

In successful experiments of this type the length of exposure 
varied between 45 minutes and 3 hours, yet uniform results 
were obtained.? A possible influence of operative shock, on ani- 
mals that were subjected to experimentation directly after the 


3’ Under normal conditions the pigment is fully extended by light in about 45 
minutes (Arey, 716). 


220 LESLIE B. AREY 


optic nerve was severed, was shown to be non-existent. It was 
also shown that roughly performed operations, in which various 
mutilations of the eye muscles and the adjacent blood vessels 
were intentionally made, did not interfere with the obtainment 
of essentially typical results. 

This type of experiment has been repeated very many times. 
Among animals used during both the spring and the fall of 1913, 
no deviation was found from the conditions recorded above. 
When work was begun again on a new supply of Ameiurus 
in the spring of 1914, the results secured from operated animals 
brought from darkness into the light, were not always as decisive 
as those of the preceding year. Some experiments showed a 
complete retention of the pigment, while others resulted in a 
partial migration, which, in some cases, was quite extensive 
although never as complete as in normal light adaption. Further 
experimentation on other animals procured in the fall of 1914 
also gave inconsistent results, varying from complete to very 
incomplete control. 

I am unable to state the cause of these discrepancies. That 
the results of the work done in the year 1913, which embraced 
the larger part of this experimentation, are beyond question, I 
feel confident; the remarkable consistency of a long series of 
experiments about to be described substantiates this conviction. 
On the other hand, the persistent occurrence of more or less 
incomplete, mingled with complete control in the work of the 
year 1914, was also undeniable. 

It is possible that individual fish vary considerably in their, 
ability to inhibit the migration of pigment after sectioning of 
the optic nerve, or what amounts to the same thing, the activity 
of the pigment in producing expansion may be variable. I shall 
attempt to present evidence, based on experimentation to be 
described later, that demonstrates an inhibiting mechanism which 
prevents the pigment from migrating when the optic nerve is 
severed. Since it has been proven (Arey, 716) that the action 
of light on completely excised eyes can directly stimulate the 
retinal pigment to movement, it is evident that there must be 
a competition between light stimulation and this inhibiting mech- 


EFFERENT FIBERS OF THE OPTIC NERVE 22a 


anism for the control of pigment migration. If the inhibiting 
mechanism is weak or becomes exhausted, or if the response of 
the pigment cells to the direct action of light is especially strong, 
the common result will be the production of a more or less exten- 
sive movement of the pigment.‘ 

This, nevertheless, does not explain how different lots of ani- 
mals, obtained in the spring and fall of one year, differed as a 
whole from other lots, procured in corresponding seasons of the 
following year. All the fish used were supplied by one collector, 
who secured them from several small ponds, according to their 
varying abundance from year to year. It is possible that the 
stocks of different ponds may show more or less consistency, due 
to common inheritance, in the behavior of their retinal pigment 
under the previously mentioned experimental conditions. The 
writer, however, wishes to present this suggestion merely as a 
possibility that may or may not be true. 

A second type of experiment, in which light-adapted animals 
were subjected to darkness, is illustrated by the following record. 


Experiment 8.1.49. One optic nerve only of a light-adapted Ameiurus 
was severed. The animal was subjected to darkness for 18 hours, 
after which both eyes were excised. No pigment contraction was 
found to have occurred in the operated eye (fig. 2), whereas the pig- 
ment of the control eye showed the extreme contraction of dark- 
adapted retinas (fig. 4). 


The greatest amount of contraction ever observed was a with- 
drawal of the distal accumulation of pigment, which even at 
room temperature is often characteristic of light adaption, to 
form a uniform zone (fig. 2). Sometimes a slight thinning out 
of pigment was seen, whereby the expanded mass appeared less 
dense than normally. In no case, however, was there observed 
a withdrawal of pigment, as a whole, from the zone usually 


4 A few determinations of the rapidity of pigment adaption seemed to indicate 
that the migration occurred more slowly in the fishes used during the first season, 
which gave constant results, than in the animals used during the following year. 
This fact not only tends to support the hypothesis that a variability in the 
activity of pigment migration was responsible for the lack of consistency in the 
later results, but also indicates that the résponse of the pigment in the animals 
of the first year, taken as a whole, may have been less vigorous. 


222 LESLIE B. AREY 


occupied during light adaption. In this type of experiment, 
therefore, it is also evident that severance of the optic nerve 
restrains the normal pigment migration. The fact that the pig- 
ment of excised eyes does not contract in darkness, and that 
pigment contraction in general seems to involve a less vigorous 
response than does expansion, explains why consistent results 
were obtained in all these experiments, for the competition with 
the inhibiting mechanism presumably was less keen. 

To remove any doubt that the integrity of the optic nerve is 
necessary for the migration of pigment in light and in darkness, 
the nerve was cut inside the cranium near the chiasma. In order 
to do this, a small aperture was made in the cranium over the 
region of the optic nerve and the severance of the nerve was 
accomplished by introducing a cataract scalpel through the open- 
ing thus made. The removal of bone from the thick cranivm 
had to be done with eare or the fish failed to recover from the 
resulting shock effect. In successful experiments the results 
obtained were similar to those previously recorded. Since the 
eye and the surrounding parts were left intact during the whole 
procedure, it is safe to conclude that in Ameiurus the optic nerve 
is intimately related to the phenomenon of pigment migration. 

A microscopical study of preparations demonstrating the rela- 
tion of the optic nerve to the nerve-fiber layer of the retina shows 
that the arrangement of nerve fibers is approximately of a radial 
nature, and that the fibers from any sector of the retina are 
extended into the adjacent side of the optic nerve. These rela- 
tions probably become disturbed a short distance from the 
emergence of the optic nerve from the eyeball. The optic nerve of 
Ameiurus joins the retina by several roots (fig. 7, rdz. n. opt.), 
hence this condition is not shown as diagrammatically as in 
preparations of the eye of Abramis which cut the optic nerve 
in aradial plane. Here the nerve fibers form a distinct V-shaped 
‘parting’ as they pass to the two sides of the retina. On the 
periphery of the nerve this is especially evident although at the 
center there is undoubtedly more decussation, as was shown by 
Kohl (’92) for several of the lower vertebrates. 


EFFERENT FIBERS OF THE OPTIC NERVE 223 


This relation should stand the physiological test imposed by 
partial section of the optic nerve. In view of the results already 
established, one would expect to find the pigment situated in 
a sector adjacent to the cut to be unaffected by varying condi- 
tions of light and darkness, while the pigment adjacent to the 
intact portion of the nerve should expand or contract in a normal 
manner. The description of an actual experiment will make this 
clear. 

Experiment 8.1.38. The optic nerve of a light-adapted Ameiurus 
was two-thirds severed close to the eyeball, and the fish was trans- 
ferred to darkness. After 2 hours the eye was excised, a pointed flap 
of skin being left attached to the eyeball to insure proper orientation. 
In a preparation, sectioned to include portions of the retina adjacent to 
both the cut and the uncut fibers, a striking contrast was evident 
(fig. 7). On the uncut side the pigment was contracted maximally, 
while on the cut side it still remained in the expanded position charac- 
teristic of light adaption. 


I can think of no experiment that could be devised to support 
better the view concerning the rdle which the optic nerve plays 
in the movement of retinal pigment, than the one just described. 

The similar treatment of dark-adapted fish which were sub- 
jected to light did not give as decisive results as those of the 
reciprocal set. The pigment on the ‘cut’ side of the eye was 
not retained in a contracted state, but migrated to a considerable 
extent (fig.6). In no case, however, was the expansion maximal, 
that is, involving an accumulation near the external limiting 
membrane, but at most only a broad evenly pigmented zone was 
formed, which was markedly in contrast with the maximal expan- 
sion of pigment in that half of the retina adjacent to the intact 
portion of the nerve. 

By the evidence of previous experimentation, a reason for this 
difference in behavior is suggested. It has already been shown 
that light is an efficient’ stimulus in producing expansion of the 
pigment in excised eyes, whereas darkness does not contract 
expanded pigment. Furthermore, after complete section of the 
optic nerve, the pigment in certain cases still tended to migrate 
when the animal was exposed to light. In the experiment last 
described, it is probable that there was a more or less extensive 


224 LESLIE B. AREY 


decussation of fibers distal to the cut, and that these fibers, 
although few in number, since they come from the intact portion 
of the nerve and distribute themselves through the retina of the 
opposite side are capable of allowing the pigment to migrate when 
the eye is subjected to.the action of light. In producing this re- 
sult, these fibers are doubtless servient to the direct action of light 
on the pigment cells themselves. In darkness, without the aid of 
some independent and efficient factor, such as the direct stimulus 
of light, it may well be that these stray fibers are not sufficiently 
potent to overcome the inertia of the pigment. 

If the cause of the apparently inconsistent behavior described 
in these two types of experiment is something of the nature of 
that just outlined, then the evidence for the control of pigment 
migration through the optic nerve thereby receives additional 
support. 

It was thought that if a small cut were made through the 
eyeball and retina, near the entrance of the optic nerve, the por- 
tion of the retina peripheral to the cut would be freed from all 
connections with the optic nerve, and consequently some decision 
could be reached concerning the réle of the decussating fibers 
just considered. This method, if successful, would also corrobo- 
rate the general conclusions which were drawn as to the control 
of pigment migration by the optic nerve fibers. 

On the whole, the results showed quite conclusively that in 
the earlier experiments the presence of decussating fibers had 
caused the pigment expansion on the side of the retina adjacent 
to the cut optic nerve, when the eye was stimulated with light. 
Cuts about 1.0 mm..in length were made with a cataract scalpel. 
Although the blade of the scalpel was very thin, tapering to a 
needle-like point and the edge was of great keenness, nevertheless, 
in many cases the retina was torn away from the contracted 
pigment layer. The ease with which a separation of the retina 
and the pigment layer is effected, especially when the pigment 
is not expanded to strengthen the union, is well known. 

In successful operations (fig. 5), the pigment lying peripheral 
to the cut showed but little expansion, although it was often the 
case that toward the extreme periphery of the retina, pigment 


EFFERENT FIBERS OF THE OPTIC NERVE 225 


expansion was again found. Since it is probable that the fibers 
of the nerve-fiber layer do not have a precise radial distribution, 
the latter condition is readily explained by assuming the presence 
in this peripheral expanded region of intact nerve fibers, which, 
near the fundus, bordered on the incision. 

The decrease in intraocular pressure necessitated by cutting 
the eyeball has no effect on the activity of the pigment, for nor- 
mal animals whose corneas have been punctured show typical 
responses. ; 

One is driven to the conclusion, by all the experiments hereto- 
fore described, that there must be some mechanism, either in the 
eye muscles or in the blood vessels of the eye, that exerts an inhi- 
bition on the movements of the retinal pigment, for proper con- 
trol experiments have eliminated the skin as a possible factor. 

If this supposition is true, the retinal pigment should undergo 
expansion when all the eye muscles and blood vessels of a dark- 
adapted fish are cut, and the eye, connected to the body by the 
optic nerve only, is subjected to light. This condition was indeed 
realized, the pigment distribution being essentially like that in 
totally excised eyes. The same experiment under reversed light 
conditions did not result in a contraction of the pigment. As in 
the similar failure of excised eyes to show contracted pigment 
when subjected to darkness, I believe there is a strong proba- 
bility of an anaesthetic action on the pigment cells due to the 
accumulation of catabolic waste. If the vascular circulation 
could be preserved, it is probable that the pigment cells would 
contract in an experiment of this kind. Possibly with artificial 
circulation the pigment of an excised eye would also contract in 
the dark. ' 

The next step was to discover whether the presence of certain 
eye muscles could be correlated with the inhibition of the pigment 
response. When dark-adapted Ameiurus, having the optic nerves 
cut, were brought into the light and the dorsal oblique and pos- 
terior rectus muscles (those innervated by trochlear and abducens 
nerves respectively) were severed, no pigment migration occurred. 
It is evident, therefore, that the inhibiting mechanism does not 
involve these muscles alone. Reciprocally, the dorsal, ventral 


226 LESLIE B. AREY 


and anterior rectus and the ventral oblique muscles (those inner- 
vated by the oculomotor nerve) were cut, leaving the eyeball 
attached to the body by the two remaining muscles. Under 
these conditions, the pigment migrated much as in a normal 
animal. 

Hence it appears that the inhibitory mechanism has been 
located as existing in association with the muscles (or possibly 
the blood vessels of the muscles) which are innervated by the 
oculomotor nerve. The objection may be raised, however, that 
there is a possibility of the dorsal oblique and posterior rectus 
muscles possessing an inhibitory function also, but that the inhibi- 
tion produced when only two muscles are left in connection with 
the eye is not sufficient to prevent the pigment response. The 
evident check experiment which answers this criticism consists 
in severing all the muscles except two which are innervated by 
the oculomotor nerve. When this was done (the dorsal rectus~ 
and the inferior oblique being left intact) no movement of the 
pigment was observed. This result indicates that the inhibi- 
tory mechanism is found associated with those muscles inner- 
vated by the oculomotor nerve and is not demonstrably asso- 
ciated with the other eye muscles. 

Since the dorsal oblique and posterior rectus muscles are not 
potent in restraining pigment migration, it was possible to make 
the following experiment. All the muscles excepting these two 
were cut and light-adapted fish with the optic nerves either intact 
or severed were placed in a dark situation. If the retinal pig- 
ment contracted under these conditions, it could be reasonably 
assumed that the presence of a partial vascular circulation had 
been responsible for the change. However, no movement of the 
pigment was detected, yet as the blood vessels in these muscles 
are small in number and size, the results neither support nor 
detract from the general view that absence of blood supply pre- 
vents pigment contraction in the dark. 

Although the relation of the autonomic fibers to the oculo- 
motor nerve of teleosts has not been worked out as completely 
as in some other forms, it is at least evident that from the ciliary 
ganglion the so-called ciliary nerves pass to the eyeball, probably 


EFFERENT FIBERS OF THE OPTIC NERVE 227 


followivg the same general courses as do the branches of the 
oculomotor nerve. In man, where the ciliary nerves have been 
traced rather carefully, autonomic fibers supply the sclerotic and 
choroid coats, ciliary muscle, iris, and cornea of the eyeball 
(Carpenter, ’06). The ciliary ganglion of fishes according to 
Onodi (’01) also receives fibers from the trigemival nerve, the 
relation being more intimate than Schwalbe (’79) believed. 

The action of atropine upon sympathetic nerves is to paralyze 
the endings of the postganglionic fibers. If the mechanism that 
inhibits pigment migration involves sympathetic nerves, it was 
thought that it might be possible to eliminate its action by the 
use of this drug. 

By supplying fishes with water through a tube, they could be 
retained indefinitely in the air. Dark-adapted fish with severed 
optic nerves were brought into the light, and oecasionaly during 
the course of an hour small amounts of a 0.5 per cent solution of 
atropine sulphate were introduced into the orbit. Subsequent 
examination of the retina showed that a migration of pigment to 
the light phase had occurred. 

The possibility of a direct stimulation of the pigment cells 
presented a difficulty, however, that had to be tested, for Spaeth 
(13) found that a 1.0 per cent solution of atropine caused a rapid 
expansion of the melanophores of Fundulus. When dark-adapted 
eyes were placed in a 0.5 or 1.0 per cent solution of the drug and 
were returned to darkness for an hour or more, the pigment did 
expand to an extent which nearly equalled that caused by in- 
complete ight adaption. 

It is certain, therefore, that any evidence gained from experi- 
mentation of the kind described is valueless, and it is doubtful 
if much more dependence can be placed on the significance of a 
pigment migration when atropine was painted on to the eye 
muscles in such limited amounts that a direct stimulation of the 
pigment cells seemed improbable, or when small quantities of 
atropine were injected into the cranial cavity.® 

°> The use of nicotine, which paralyzes the synapse of the preganglionic fiber 
with the sympathetic nerve cell, might furnish more interesting results, provided 


it did not have a toxie effect on the pigment cells. Unfortunately, no experi- 
ments of this kind were performed. 


228 LESLIE B. AREY 


An attempt was made to cut the oculomotor nerve and thus to 
separate the sympathetic fibers which arise with it from the brain. 
To make a large opening through the cranium leads to operative 
shock from which Ameiurus does not recover. Accordingly, a 
relatively small aperture was made, through which a cataract 
scalpel was introduced, all the nerves presumably being cut on 
one side of the brain between the optic and the trigeminal. A 


dark-adapted Ameiurus with severed optic nerve, when exposed — 


to the light after this treatment, showed pigment expansion. 
This experiment, although properly controlled, was not of a 
‘ refined type. It is possible that disturbances other than the 
mere section of the oculomotor verve may have led to the ob- 
served results. 

It is not intended that any of the experimentation in which 
an attempt was made to locate the fibers of the inhibiting mecha- 
nism, should be received as conclusive. The results obtained 
from cutting muscles and from the severance of the oculomotor 
nerve suggest that autonomic fibers are involved, and that these 
enter with the oculomotor nerve. <A strong suspicion is there- 
fore cast upon the sympathetic fibers in connection with the 
ciliary ganglion as the causal agents in preventing movement of 
the retinal pigment when the optic nerve is cut. 

No statement has previously been made concerning the nature 
of the mechanism in the optic nerve, the integrity of which is 
necessary for allowing positional changes of the pigment to occur. 
If the inhibitory fibers associated with the eye muscles are con- 
ceived as acting after the manner of a brake, it follows that the 
optic nerve must contain active components, which in some way, 
directly or indirectly, permit expansion and contraction of the 
pigment. 

If the optic nerve contains fibers of an efferent nature, it is of 
interest to discover whether these components can be made to 
function by electrical stimulation. The description of a typical 
experiment will best illustrate this point. 


Experiment 8.1.23. A previously dark-adapted Ameiurus, in which 
both optic nerves had been cut, was retained in the air, respiratory 
water being supplied through a tube. The peripheral end of the cut 


EFFERENT FIBERS OF THE OPTIC NERVE 229 


optic nerve of one eye was exposed and stimulated for 1} hours, in the 
light, with a weak current from an inductorium. Inspection of the 
two retinas showed that the pigment in the eye that had been elec- 
trically stimulated was uniformly expanded (fig. 2), while in the control 
eye the pigment still remained contracted as is characteristic of dark 
adaption (fig. 3). 


Since repeated trials confirmed these results, the conclusion 
follows that there must be nervous elements of an efferent nature 
in the optic nerve of this fish which normally receive impulses 
from the central nervous system, and can be made to function 
experimentally by electrical stimulation. The result in either 
case is a release from an inhibition exerted upon the pigment cells 
by a mechanism possibly involving the autonomic fibers from the 
ciliary ganglion. 

Previous experimentation (Arey, 716) has shown that both in 
light and in darkness the retinal pigment of fishes is more highly 
expanded at low temperature (0°C. +) than at high temperatures 
(25°C. +), hence a series of experiments was next made to discover 
whether after the optic nerve was cut temperature would still 
be efficient in producing the characteristic temperature responses. 
In determinations made both in light and in darkness, the results 
were identical with those found in normal animals, for at 5°C., 
the expansion of pigment was greater than at 25°C. (figs. 1 and 
2;3 and 4). 

In order to produce sharp contrasts (in the light at least), 
before the optic nerve was cut, a preliminary treatment at a 
temperature of 5°C. was necessary. If the preliminary treat- 
ment was at 25°C. the pigment remained in the position charac- 
teristic of that temperature, regardless of the temperature that 
followed. From this result additional evidence is obtained, over 
that already advanced (Arey, 16), to show that a high tempera- 
ture is more efficient in causing positional changes of the retinal 
pigment than is a low temperature, and that temperature is 
more efficient than either light or darkness.°® 


5 It should be remembered, however, that temperature merely produces a 
quantitative redistribution of the already expanded or contracted pigment, the 
extent of its influence always being limited in this way. 


THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL. 26, NO. 3 


230 LESLIE B. AREY 


2. Visual cells. ‘The more important experiments, which show 
the relation of the optic nerve to the pigment migration of 
Ameiurus, were repeated in order to discover whether a similar 
set of relations exists between the optic nerve and the visual 
cells. 

It will be remembered that the myoid of the cone cells of fishes 
elongates in the dark and shortens in the light (figs. 10 and 11), 
whereas the behavior of the rod myoid is the converse of this. 
Moreover, in excised eyes both types of cells are stimulated to 
movement through the direct action of light, whereas only the 
rods show a response in darkness. 

In general it may be said that these experiments, which ex- 
tended discontinuously throughout a period of two years, showed 
that whenever the migration of pigment was inhibited the move- 
ments of the rods and cones likewise failed to occur. In those 
cases, discussed previously (p. 220), where cutting of the optic 
nerve only partially inhibited the pigment response, the move- 
ments of the visual cells usually were not prevented. Practically 
all of the experiments to be described were performed in the 
spring and autumn of 1913 before cases of incomplete control 
were discovered. 

When the optic nerve of a dark-adapted fish was severed, and 
the animal was subjected to light (fig. 12), the rods remained as 
in the condition of dark adaption. The cones, on the contrary, 
usually became more or less shortened, although they did not 
approach the external limiting membrane as closely as when 
exposed to light under normal conditions. 

In the converse experiment, in which operated fish were trans- 
ferred from light to darkness, the cones showed no tendency to 
elongate, but in some cases remained maximally shortened for 
several hours. The rods, however, did not always retain the 
highly elongated position characteristic of darkness, for they 
commonly exhibited partial retraction, although the final position 
was unmistakably that of semi-elongation. 

From these two types of experiment one striking correlation 
is evident. The cones have a greater tendency to undergo posi- 
tional changes in the light, while the rods behave similarly in 


EFFERENT FIBERS OF THE OPTIC NERVE 231 


darkness. The end result of this tendency towards movement, 
nevertheless, is identical in both kinds of visual cells, for a shorten- 
ing of the myoid results in either case. This would suggest that 
the process of retraction, as in simple contractile tissue in general, 
is More vigorous than is elongation. This behavior of the cones 
is comparable to that observed in excised eyes of Ameiurus, 
where these cells shortened in the light but remained unchanged 
in darkness; the rods of excised eyes, on the contrary, moved both 
in light and in darkness. The posulation of a more vigorous 
response in retraction offers an explanation for the observed facts, 
for under such conditions the inhibitory mechanism would be 
less likely to check completely the positional changes of these 
elements. 

Reference has been made (p. 223) to experiments in which the 
optic nerves of light-adapted fish were partly severed. When 
these animals were transferred into the dark, the retinal pig- 
ment showed distinct areas of expansion and contraction, corre- 
sponding respectively to the cut and uncut sides of the optic 
nerve. In one such preparation which had been stained, the 
position of the cone cells was also observed to vary with respect 
to the same areas. Where the pigment had undergone move- 
ment, the cones were elongated, whereas in the half of the retina 
adjacent to the cut fibers the cones remained maximally shortened. 

A control experiment, in which the light-adapted eye, con- 
nected to the body by the optic nerve only, was subjected to 
darkness, showed that the rod cells had shortened although not 
always completely. The cones, as in excised eyes, did not change 
their positions in the least. There is no evidence that the optic 
nerve of itself prohibits the movements of the rod and cone 
cells. 

The results which have thus far been outlined indicate that, 
just as with the retinal pigment, there is a mechanism asso- 
ciated with the muscles or blood vessels of the eye which tends 
to inhibit movements of the visual cells. When the dorsal 
oblique and posterior rectus muscles are cut, and after section 
of the optic nerve the dark-adapted Ameiurus is exposed to 
light, no changes in the position of the rods and only incomplete 


232 LESLIE B. AREY 


changes of the cones occur. The muscles innervated by the 
trochlear and abducens nerves evidently are not essential in 
ausing an inhibition of movement. 

A few reciprocal experiments were performed by cutting all 
the eye muscles except the dorsal oblique and the posterior 
rectus. Although the extent of positional change was not as 
striking as in excised eyes, it seems probable that the same 
mechanism that controls the migration of pigment, and is pre- 
sumably associated with the muscles innervated by the oculo- 
motor nerve, also acts on the visual cells. 

The evidence, therefore, suggests that normally there are im- 
pulses which travel in the optic nerve and render ineffective the 
inhibition produced by a second set of fibers. Hence it becomes 
a matter of interest to discover whether it is possible, by arti- 
ficial stimulation of the severed optic nerve, to cause the visual 
cells to move more freely than they would otherwise do. The 
details of these experiments were similar to those previously 
described when the migration of pigment was tested.. The cut 
optic nerves of dark-adapted fish were stimulated in the light by 
a weak faradic current. Both rod and cone cells assumed their 
characteristic light positions, while in control eyes of the same 
animals, the optic nerves of which had been cut but not stimu- 
lated, the visual cells remained for the most part as in darkness. 

These results on the visual cells of Ameiurus, taken as a whole, 
closely agree with those obtained from the study of retinal pig- 
ment, and the general conclusion concerning their significance 
is so identical with one previously stated (p. 229) that it hardly 
needs to be repeated here. 

lt was shown in a former paper (Arey, 716) that in fishes an 
elevated temperature (25°C. +) causes an elongation of the my- 
olds of dark-adapted cone-visual cells, whereas a low tempera- 
ture (0°C. +) induces a shortening of the myoids. The rod 
myoid exhibits a similar behavior but in a less marked degree. 

The effect of temperature was tested on the visual cells of dark- 
adapted fish the optic nerves of which had been cut. At 5°C. 
in the dark the cone myoid measured 18 yu, the rod myoid 8 ph + 
(fig. 8). At 25°C. in the dark the cone myoid elongated to 25 yu, 
the rod to 10 » + (fig. 9). These results are in agreement with 


EFFERENT FIBERS OF THE OPTIC NERVE 200 


those obtained with normal animals. The effective action of 
high temperature in causing an elongation of the cone cells is 
noteworthy, since without the aid of temperature no change in 
the position of the cones occurred under these conditions. 


b. Experimentation upon Abramis and. Fundulus 


Having presented evidence which indicates that the optic nerve 
fibers of Ameiurus are not all afferent in function, the question 
arises as to the occurrence of this condition among other fishes 
as well as among other classes of vertebrates. A series of experi- 
ments was carried out upon Abramis and Fundulus in which the 
effect produced on the retinal elements by the severance of the 
optic nerve was tried. 

Both in light and in darkness the retinal pigment and the cone 
cells of Abramis underwent movements which were essentially 
normal, and hence independent of the cut nerve. Both in the 
cone myoid, which is capable of a 90 per cent retraction, and in 
the pigment, which exhibits extreme conditions of contraction 
and expansion, was an independent movement strikingly ex- 
hibited. The stain used in’ making these preparations did not 
demonstrate the rods to advantage, except in one instance where 
a light-adapted fish had been subjected to darkness. In this 
case the rods were shortened, occupying the characteristic dark 
position. 

When all the eye muscles and blood vessels of dark-adapted 
Abramis were cut, the eye remaining attached to the body by 
the optic nerve only, no marked change in the position of the 
pigment or cones accompanied a removal into the light. Since 
the pigment and cones show no movements when the eyes are 
completely excised and exposed to light or darkness, these experi- 
ments favor the view that the absence of movement in such cases 
is not due to an inhibition through the optic nerve but to an 
interruption in the vascular supply.’ 

7 In several instances a large blood vessel that lies near the optic nerve was 
accidentally cut in experiments where the optic nerve only was being severed, 
yet the changes in the retinal elements occurred as before. This further sup- 
ports the view that the blood supply to the retina by means of vessels in con- 
nection with the eye muscles is important in allowing the positional changes 
to occur. 


23 LESLIE B. AREY 


The effect of severing the optic nerve of Fundulus was investi- 
gated in a similar manner. When brought from darkness to 
light the pigment migrated normally after the usual operation. 
The converse experiment also showed an independent movement 
of the pigment, although the poorly defined contraction which 
characterizes dark adaption rendered this type of experiment 
less decisive than the clear cut results on Abramis. 

It is evident, then, that unlike Ameiurus the movements of 
the retinal elements of Abramis and Fundulus are not dependent 
upon the integrity of the optic nerve. From the results of Ham- 
burger (89) and Fick (91) upon the frog, who showed that the 
retinal pigment underwent movements after the optic nerve was 
cut at the chiasma, it is safe to conclude that in this animal also 
the optic nerve does not control pigment migration. The experi- 
ments of these latter workers are all the more interesting since 
in many observations which have been recorded showing the 
retinal elements of the frog to be under nervous control, it has 
generally been directly stated, or at least implied, that the optic 
nerve was involved. It would be unprofitable to speculate con- 
cerning the further occurrence among the vertebrates, or even 
among the fishes, of individuals possessing efferent optic nerve 
fibers which act similarly to those in Ameiurus. 


THEORETICAL CONSIDERATIONS 


Although serious doubt has been cast on many of the earlier 
results upon the frog which were supposed to demonstrate the 
presence of efferent fibers, both between the brain and the 
retina and between the two retinas by way of the optic chiasma, 
yet it is probable that there is a residue of truth in the general 
proposition of a nervous control of the retinal elements in this 
animal. 

Fick (90, p. 84), in a paper attacking Englemann’s assertion 
of the existence of efferent nerve fibers, takes the following 
position: 

Wenn dieser Schluss richtig ist, so kann man nur ruhig durch die 


ganze bisherige Sinnesphysiologie einen Strich machen und ihre Erfor- 
schung von Neuem beginnen; denn die rein centripetale und specifische 


EFFERENT FIBERS OF THE OPTIC NERVE 235 


Leitung in den Sinnesnerven gilt als die eigentliche Grundlage dieses 
Abschnittes der Physiologie. 


This statement may have been partially justified in view of 
the diversity of stimuli reported to be potent in causing changes 
in the retinal elements of the frog, but even if the presence of 
efferent fibers of the kind physiologically demonstrated in Amei- 
urus were found in all the vertebrate classes, it by no means 
follows that the fundamental principles of sensory physiology 
would be seriously endangered. In the normal Ameiurus the 
efferent function does not in the least interfere with the move- 
ments of the retinal elements, and stimulating agents which do 
not act directly upon these cells presumably are ineffectual in 
producing changes. Since, moreover, there is no evidence of 
any intrusion on the part of these elements upon any of the 
sensory processes, our ideas of retinal physiology scarcely require 
modification and certainly do not demand reorganization. 

From the study of Ameiurus experimental proof has been 
advanced showing the existence of an inhibitory mechanism, not 
associated with the optic nerve, which tends to prevent the 
movements of pigment cells and retinal pigment. Moreover, a 
second mechanism associated with the optic nerve was demon- 
strated, the integrity of which is necessary for the accomplish- 
ment of typical movements on the part of the retinal elements. 

It may now be fairly asked whether there is any evidence 
indicating the possible modus operandi of these two systems. 
Several schemes are readily suggested by which the facts hitherto 
presented could be explained. If, however, one conceives of the 
efferent fibers in the optic nerve as actively causing movements 
of the retinal elements, both in darkness and in light, either two 
kinds of efferent fibers must be postulated, or one movement of 
each of the elements is passive—a return to the unstimulated 
condition—and in some way is interfered with by the inhibitory 
mechanism when the optic nerve is cut. Nevertheless, either 
one of these explanations becomes discrepant when applied to the 
total behavior of the retinal elements. 

Since electrical stimulation of the cut optic nerve in the light 
induced changes in the retinal elements of an otherwise intact 


236 LESLIE B. AREY 


fish, it would seem reasonable to expect that the same would 
oecur in darkness, provided efferent impulses in the optic nerve 
directly stimulate the retinal elements to undergo positional 
changes. 

Gertz (11) tested the effect of electrical stimulation for one to 
two minutes in darkness and in hght upon the eyes of Abramis 
brama, both excised and in vivo, but only negative results were 
obtained. Englemann (’85) had previously asserted that the 
cones and retinal pigment of the excised or normal eyes of dark- 
adapted frogs responded to induction shocks in a similar way 
as to light, and Arcoleo (90) also claimed to have observed 
pigment migration under these conditions in a pithed toad and 
frog, as well as upon dark adapted excised eyes. The more 
recent work of Lederer (08) and of Fujita (11) upon normal 
frogs nevertheless, has not supported this view. 

When the cut stump of the optic nerve of a dark-adapted 
Ameiurus was stimulated with a faradic current in total dark- 
ness, no distinct changes occurred in the retinal pigment or rod 
cells, even though the stimulation continued for 45 minutes. Sim- 
ilar treatment of excised eyes also gave negative results. These 
observations indicate that a satisfactory explanation of the action 
of the efferent nerve fibers must be sought by viewing the situa- 
tion in a different way. 

It is instructive to adopt the point of view suggested by a 
study of the vasomotor nerves. In this case a tonic condition 
is presupposed either through the action of vasoconstrictor nerve 
fibers or possibly by the intrinsic properties of the muscles 
themselves. Dilation is believed to be accomplished by the 
action of dilator nerve fibers, whose impulses inhibit the tonic 
contraction of the musculature, thus indirectly causing relaxa- 
tion. To work out a detailed application for the condition found 
in Ameiurus would be both unprofitable and unwarranted. The 
simplest conception is that impulses* from the efferent compo- 

8 This view of tonicity differs greatly from that which Herzog (05) believed 
to exist in the frog. His statement that after destruction of the central nervous 
system the pigment and cones gave abnormally vigorous responses in the dark, 


as if released from an inhibition, does not agree with the observations of Ham- 
burger (’89), Arcoleo (90), Dittler (07) and Garten (’07). 


EFFERENT FIBERS OF THE OPTIC NERVE 2a0 


nents of the optic nerve block, i.e., counteract, the tonic inhi- 
bition produced by the second nervous mechanism, thereby allow- 
ing conditions of light and darkness to act directly upon the 
retinal elements.? This would explain why electrical stimula- 
tion of the cut optic nerves of dark-adapted fish, in the light only, 
led to the usual changes in the retina. Although no determina- 
tions were made, it follows that if this hypothesis is true, the 
converse experiment with light-adapted fish in the dark should 
result in the assumption of the dark phase on the part of the 
retinal elements. 

Assuming that the balanced action of a system like the one 
suggested is indeed a reality, it is evident, nevertheless, that 
questions relative to its adaptive significance are not easily 
answered. It is certainly difficult to explain the rationale of a 
situation whereby an animal possesses a mechanism the com- 
ponents of which act antagonistically, thus allowing photome- 
chanical influence to be exerted undisturbed. 

Since structure and function go hand in hand, the value of 
physiological evidence always becomes much enhanced by the 
coéxistence of a correlated structural basis. The possibility of 
double conduction in one set of nerve fibers is hardly to be con- 
sidered, hence there arises a pertinent query relative to the ana- 
tomical proof for the existence of efferent components of the optic 
nerve. 

Englemann (’85) first postulated ‘retino-motor’ nerve fibers 
to explain certain conditions which he asserted occurred in the 
frog, and further suggested that the anterior arcuate commissure 
of Hannover in the chiasma, the physiological significance of 
which had previously been unknown, served as an association 
tract through which the movements of the retinal elements of 
the two eyes were interrelated. 


9 Tt will be remembered in the experiments on Abramis and Fundulus (p. 233) 
that after severing the optic nerves the usual responses to light and darkness 
were maintained. Hence it seems probable (since the retinal elements undergo 
no changes in excised eyes) that the direct action of darkness as wel! as light is 
effective, provided normal circulatory conditions are maintained. 


238 LESLIE B. AREY 


By the use of Golgi methods and by methods of primary and 
secondary degeneration, centrifugal fibers originating in the verte- 
brate brain and extending to the retina can be demonstrated. 
The following quotation from Johnston (06, p. 265) summarizes 
the general results gained from studies of this kind: 


In fishes in which one eye has long been lost the optic tract of the 
opposite side degenerates with the exception of these efferent fibers, which 
persist and are stained by the Weigert method. In mammals, following 
section of the optic tract there occurs secondary degeneration of cells 
in the anterior quadrigeminum, and in the dorsal part of the genicu- 
latum laterale and pulvinar. These findings in mammals agree with 
those in fishes by the Golgi and degeneration methods, where the cen- 
trifugal fibers arise from the tectum opticum and geniculatum (Catois, 
02). The significance of these fibers is not understood but their pres- 
ence in all vertebrates seems to show that they have some constant 
function. 


A further anatomical difficulty is presented by the fact that a 
connection between optic nerve fibers and the pigment cells has 
never been demonstrated, for even the e%e:out fibers shown by 
Cajal (94) and others have their endin:s near the internal 
nuclear layer. Garten (07, p. 85) suggcsicd that an actual con- 
nection might not be necessary: ‘‘ Natiirlich liesse sich behaupten, 
die centrifugalen Opticusfasern rufen in der Stabchenzapfen- 
schicht einen Erregungsvorgang hervor, der sich ‘per contigui- 
tatem’ dem Pigmentepithel mitteilt.”” In view of the tentative 
conclusion reached by me concerning the balanced action of the 
‘retino-motor’ and the inhibitory nerve fibers, it is difficult to 
state what anatomical conditions would be imposed by such a 
system. Such matters of detail need not stand in the way of 
the fact of fundamental importance, which is the existence of 
demonstrably functional efferent optic nerve fibers. 

The significance of these efferent nervous elements in the light 
of the theory of nerve components may appear to demand no 
serious consideration, since the majority of neurologists prefer 
not to homologize the optic ‘nerve’ with true cranial nerves. 
Neither the afferent nor the efferent fibers of the optic nerve are 
comparable to the components of a true peripheral nerve, for both 
lie within the primary optic apparatus, which is itself a differ- 


EFFERENT FIBERS OF THE OPTIC NERVE 239 


entiated portion of the brain. Johnston (’06) and Herrick (’15) 
have assigned the optic nerve proper to the central system of 
tracts of the somatic afferent division. From the conclusion 
previously drawn concerning the function of the efferent optic 
nerve fibers, it is evident that they may, at least with as much 
propriety, tentatively be called visceral efferent elements. 


SUMMARY 


When the optic nerve only of Ameiurus is severed, the rods, 
cones and retinal pigment fail to execute their characteristic 
photomechanical responses. In other words, the movements of 
the retinal elements depend upon the maintenance of the integ- 
rity of this nerve. After hemisection ofthe nerve, movements 
of the elements occur only in that region of the retina adjacent 
to its intact side. It can not only be shown (since essentially 
normal responses ‘occur in excised eyes and in eyes attached to 
the body by the optic nerve alone) that a second mechanism 
exists in association with the muscles innervated by the oculo- 
motor nerve which inhibits these movements when the optic 
nerve is cut, but also that electrical stimulation of the peripheral 
stump of the optic nerve can overcome this inhibition. 

Hence experimental evidence has been advanced which offers 
physiological proof for the existence of functional efferent nerve 
fibers in the optic nerve of Ameiurus. Only by the balanced 
interaction of these elements with a second extrinsic set of nerve 
fibers (possibly the ciliary nerves), which independently exert 
an inhibitory effect upon the retinal elements, are normal photo- 
mechanical movements of the rods, cones, and retinal pigment 
accomplished. Although light is ineffectual after section of the 
optic nerve, temperature produces essentially normal responses 
in both the pigment and visual cells. It is probable that efferent 
impulses in the optic nerve fibers do not directly stimulate the 
retinal elements to motion, hut rather such impulses have an 
indirect action, possibly by counteracting, that is, blocking, the 
tonic inhibition exerted by the second system of nerve fibers. 
If these efferent optic nerve fibers fit at all into the scheme of 


240 LESLIE B. AREY 


‘nerve components,’ they may be designated as visceral efferent 
elements. 

Severance of the optic nerve of certain other fishes (Abramis 
and Fundulus) has no inhibitory effect upon the movements of 
the retinal pigment or of the rod- and cone-cells, and a similar 
relation has been reported for the frog by other workers. Hence 
it is impossible to state the extent to which the mechanism dis- 
covered in Ameiurus may be distributed throughout the verte- 
brate group, if, indeed, it is not peculiar to Ameiurus alone. 


BIBLIOGRAPHY 
Papers marked with an asterisk have not been accessible in the original 


Anaeuuccr, A. 1890 Untersuchungen iiber die Sehthatigkeit der Netzhaut und 
des Gehirns. Untersuch. zur. Naturlehre d. Menschen u. d. Thiere 
(Moleschott), Bd. 14, Heft 3, pp. 231-357, 2 Taf. 

1905 Physiologie générale de l'oeil. Encyclopédie frangaise d’oph- 
thalmologie, tome 2, pp. 1-141. 

*ArcoLEo, E. 1890 Osservazioni sperimentali sugli elementi contrattili della 
retina negli animali a sangue freddo. Annali d’Ottalmologia, Anno 
19, Fase. 3 e 4, pp. 253-262. 

Arey, L. B. 1915 The occurrence and the significance of photomechanical 
changes in the vertebrate retina—An historical survey. Jour. Comp. 
Neur., vol. 25, no. 6, pp. 535-554. 

1916 The movements in the visual cells and retinal pigments of the 
lower vertebrates. Jour. Comp. Neur., vol. £6, no. 1, pp. 121-202. 

Brrcn-Hirscureitp, A. 1906 Der Einfluss der Helladaptation auf die Struktur 
der Nervenzellen der Netzhaut, nach Untersuchung an der Taube. 
Arch. f. Ophthal., Bd. 68, Heft 1, pp. 85-111. 

Casau, S. R. 1894 Die Retina der Wirbelthiere. Bergmann, Wiesbaden, 4to, 
8 + 179 pp. 

Carpenter, F. W. 1906 The development of the oculomotor nerve, the ciliary 
ganglion and the abducent nerve of the chick. Bull. Mus. Comp. 
Zool., Harvard Coll., vol. 48, no. 2, pp. 141-229. 

Carots, E. H. 1902 Recherches sur l’histologie et l’anatomie microscopique 
de l’encéphale chez les Poissons. Bull. Sci. France et Belgique, Tom. 
36, pp. 1-166. 

*CurarIni, P. 1904 Cambiamenti morphologici, che si verificano nella re ina 
dei vertebrati per azione della ‘uce e dell’ oscurita. Parte 1. Retina 
dei pesci e degli anfibi. Boll. della R. Accad. Med. di Roma, Anno, 30, 
ppedo lO: 

Dirriter, R. 1907 Uber die Zapfenkontraktion an der isolierten Froschnetz- 
haut. Arch. f. ges. Physiol., Bd. 117, Heft, 5 u. 6, pp. 295-328. 


EFFERENT FIBERS OF THE OPTIC NERVE 241 


ENGLEMANN, T. W. 1885 Uber Bewegungen der Zapfen- und Pigmentzeilen der 
Netzhaut unter dem Einfluss des Lichtes und des Nervensystems. 
(Nach einem 1884 gehaltenen Vortrag.) Arch. f. ges. Physiol., Bd. 35, 
Heft 10, 11, u. 12, pp. 498-508. 

Fick, A. E. 1889 Uber die Lichtwirkungen auf die Netzhaut des Frosches. 
Bericht iiber d. 18. Versammlung d. Ophthalmol. Gesell. zu Heidel- 
berg 1889, pp. 177-183. 

1890 Uber die Ursachen der Pigmentwanderung in der Netzhaut. 
Vierteljahrsschrift d. naturf. Gesell. Ziirich, Jahrg. 35, pp. 83-86. 

1891 Untersu hungen iiber die Pigmentwanderung in der Netzhaut 
des Frosches. Arch. f. Opththal., Bd. 37, Abt. 2, pp. 1-20. 

Fusira, H. 1911 Pigmentbewegung und Zapfen kontraction im Dunkelauge des 
Frosches bei Einwirkung verschiedener Reize. Arch. *. vergl. Oph- 
thal., Jahrg. 2, Heft 2, No. 6, pp. 164-179. 

GarTEN, S. 1907 Die Veri&inderungen der Netzhaut durch Licht. Graefe- 
Saemisch Handbuch ges. Augenheilkunde, Leipzig, Aufl. 2, Bd. 3, Kap. 
12, Anhang, 130 pp. 

Gertz, H. 1911 Gibt elektrische Reitzung phototrope Netzhautreaktion bei 
Abramis brama? Arch. f. Ophthal., Bd. 78, Heft 1, pp. 224-226. 

*Griswns, G. 1891 Bijdrage tot de physiologie van den nervus opticus. Utrecht 
8vo., 3 + 76 pp. 

*HamBurGer, D. J. 1889 Dorsnijding van der nervus opticus bij Kikvorschen, 
in verband met de Beweging van Pigment en Kegels in het Netvlies, 
onder den Invléed van Licht en Duister Onderzoekingen d. Utrecht- 
sche Hoogeschol, Reeks 3, vol. 11, pp. 58-67. 

Herrick, C. J. 1915 An introduction to neurology. Saunders, Phila., Svo, 
335 pp. 

Herzoc,H. 1905 Experimentelle Untersuchung zur Physiologie der Bewegungs- 
vorgiinge in der Netzhaut. Arch. f. Anat. u. Physiol., Physiol. Abt., 
Jahrg. 1905, Heft 5 u. 6, pp. 413-464. 

Jounston, J. B. 1906 The nervous system of vertebrates, Blakiston, Phila., 
8vo., xx + 370 pp. 

Kout, C. 1892 Rudimentiire Wirbelthieraugen. Bibliotheca Zoologica, Bd. 4, 
Heft 13, pp. 1-140. 

Leprrer, R. 1908 Wirken elektrische Reize auf das Pigmentepithel des Frosch- 
auges? Centralbl. f. Physiol., Bd. 22, No. 24, pp. 765-766. 

*LopaTo, G.,& Prrrone, D. 1901 Sulle vie associative fra le due retine, studio 
sperimentale. Arch. di Ottalmol., vol. 8, pp. 465-489. 

Naummacuer, W. 1893 Uber den Einfluss reflectorische und centrale Opticus- 
reizung auf die Stellung der Zapfen in der Froschnetzhaut. Arch. f. 
ges. Physiol., Bd. 35, Heft 9 u. 10, pp. 375-387. 

6nop1, A. 1901 Das Ganglion ciliare. Anat. Anzeiger, Bd. 19, No. 5 u. 6, pp. 
118-124. 

*Prercens, EB. 1896 Action de la lumiére sur la rétine. Ann. Soc. Roy. Se. 
Méd. Nat., Bruxelles, tome 5, pp. 389-421. 

ScHWALBE, G. 1879 Das Ganglion oculomotorii. Jena. Zeitschr. f. Natur- 
wissensch. Bd. 13, pp. 173-268. 

Sparetu, R. A. 1913 The physiology of the chromatophores of fishes. Jour. 
Exp. Zo6l., vol. 15, no. 4, pp. 527-585. 


ABBREVIATIONS 


ell. bac., rod ellipsoid prs. dst. bac., outer member of rod 
ell. con., cone ellipsoid prs. dst. con., outer member of cone 
mb. lim. ex., external limiting mem-  rdx. n. opt., root of optic nerve 

brane rin., retina 
my. bac., rod myoid scl., sclera 
my. con., cone myoid st. bac. con., layer of rods and cones 
n. opt., optic nerve st. nl. ex., external nuclear layer 
pd. cl. pig., base of pigment cell st. pig., pigment layer 

PLATE I 


EXPLANATION OF FIGURES 


The figures of this plate are photomicrographs. Figures 1, 2, 3, and 4 are 
magnified 160 diameters, figure 5, 46 diameters, figure 6, 19 diameters, and fig- 
ure 7, 50 diameters. 

1 Shows the distribution of the retinal pigment of Ameiurus at 5° C. in the 
light. 

2 Shows the distribution of the retinal pigment of Ameiurus at 25° C. in the 
light. 

3 Shows the distribution of the retinal pigment of Ameiurus at 5° C. in the 
dark. 

4 Shows the distribution of the retinal pigment of Ameiurus at 25° C. in the 

dark. 
5 <A portion of a section passing through the retina of a previously dark- 
adapted Ameiurus. In the region marked X, close to the optic nerve, asmal cut 
had been made through the eyeball and retina. When the fish was subjected to 
daylight for 13 hours, the pigment peripheral to the incision did not expand, as 
is shown at the right of Xin the figure. The pigment in the regions of the retina 
the optic nerve fibers of which were not affected by the incision expanded essen- 
tially in a normal manner, as is shown in the left half of the figure. 

6 A section through the entire retina of Ameiurus. After the optic nerve of 
the previously dark-adapted animal had been one-half severed, the animal was 
exposed to daylight for 2 hours. In the half of the retina at the right of the 
figure, which was adjacent to the intact portion of the nerve, the pigment 
migrated to an extreme distal position, leaving an area behind relatively free 
from pigment. In the left half of the retina, which was adjacent to the cut 
side of the nerve, the pigment migrated, but not as completely as on the other 
side. This section passed close to the optic nerve but not through it. 

7 <A portion of a radial section through the retina of Ameiurus. After the 
optic nerve of a previously light-adapted animal had been two-thirds severed, the 
fish was subjected to total darkness for 2 hours. At the expiration of this time, 
the pigment in the half of the retina adjacent to the cut was found to have re- 
mained in the expanded position characteristic of light (shown at the right of 
the optic nerve in the figure), whereas in the half of the retina adjacent to the 
intact side of the nerve, normal pigment contraction occurred (shown at the 
left of the optic nerve in the figure). 


242 


EFFERENT FIBERS OF THE OPTIC NERVE PLATE 1 
LESLIE B. AREY 


ee eee 


--pd.cl.pig. 


st.pig.—-- 


--mbd.1Im.ex 


st.bac.con: 


--mb.lim.ex,--- 


L.B. A. Photo. 


PLATE 2 


EXPLANATION OF FIGURES 


All drawings were made at a magnification of 1400 diameters, a Leitz 75 
homogeneous immersion objective being used; in the reproduction figures 8, 9, 
and 12 are reduced to a magnification of 1159 diameters, and figures 10 and 11 
to 950 diameters. 

8 Shows the positions assumed by the visual cells at 3° C. in the dark. 

9 Shows the position assumed by the visual cells of Ameiurus at 27° C. in the 
dark. 

10 Shows the characteristic position of the visual cells of Ameiurus at room 
temperature in the dark. 

11 Shows the characteristic position of the visual cells of Ameiurus at room 
temperature in the hght. 

12. From the retina of a dark-adapted Ameiurus the optic nerve of which 
had been cut and the animal subjected to light. The rods remain shortened, as 
in darkness, although in this case the cones have shortened to a great extent, 
as is the characteristic response in light. 


244 


EFFERENT FIBERS OF THE OPTIC NERVE 
LESLIE B. AREY 


_---ell.con. iar ® 
my.con.------ y 
my.con 
ae -----prs.dst.bac. prs.dst,bac- 225 ae 3 
ee I] Flo} Cod 
ell bac. ---- 
Saeed my.bac 
een mb.lim.ex. my.bac------ 
---- St.nl.ex 
st.nl.ex.._ 


mb.lim.ex. ---- 


---ell.bac. 


------my.bac. 


-}|---prs.dst.con, 


P----= my.con 
-ell.con.----- 
Res prs.dst,bac. : ¢my.con 
ell.con. ---}--® | 

prs.dst.bac. ---- 
elfbac ell.bac--- 3 

Nur my.bac--- 

---my.bac 


~-~mb,lim.ex.) 


~sst.ni.ex. 


L. B. A. Del. 


245 


THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL. 26, NO. 3 


PLATE 2 


prs.dst.con. 


my 7on.----4/ Wat if 


Pee es ¥ N 

i is pt, a Me, 
Ny og td y vad otal eee 
: P nf 40.5 ve AD yim 


‘ H “> , a. f hs 
| : Mh oe SPN gt DPE MT 2 
; vf Wak iS staan ae i 
ye mai) oh | ea ti mtn 


Ps / ve 
« ; ‘ 
i e ox, oe 
nee eae 
, ; _ bs ; 
7 | , 
5 ¥ ) i ‘, 
+ 4 a Oe 
: Vee ft) 
fe i) 
. Os *) 


CORRELATED ANATOMICAL AND PHYSIOLOGICAL 


STUDIES OF THE GROWTH OF THE NERV- 
OUS SYSTEM OF AMPHIBIA 


Il. THE AFFERENT SYSTEM OF THE HEAD OF AMBLYSTOMA 
G. E. COGHILL 
Department of Anatomy, University of Kansas 
SEVENTY-NINE FIGURES 
CONTENTS 
fn iS ABADOUMICAN WARG es da oy ssc sss 6 id ape o's ao ve eodls «SEDO MMMR aS RSs bi 248 
PVE PRD DRLVO Los... sx 5 ancien wae oooh ee eRe SR Oy eelin oO 
a; Die MONAMOUWLS BLAGE. . oso: 2 «sis olen sp s aie es ed setae sine 2 sk 5 ha 248 
bu The early flexure stage...... 2... 6. ccs. ee ne continues mass se awe 251 
c, The: coiled-reaction stage. ... 55... .-1 sana wists Cailege ahs oe anes sie 253 
d. The early swimming stage............ 0-5. 6+ edesessecesesee. 200 
2. The facial and auditory nerveS..............cceeeee eee daneeeeeeee 257 
Br LNG MOMINOUIIC SLAB. .cbeccs ea uve sun cee oe PORE oh 258 
b. The early flexure stage............0..s0seneeresenwstectccees 259 
Go ENG COMEU-TEACLION SCARS... os cco eens oa:er oa bie Cinperee emia S aielsin (> 261 
d. The early swimming stage................eeeseeer enews eeeees 264 
3. The glossopharyngeal and vagus nerves.............000eee ee eee 266 
A. Dhe mon-motile Stages ....0.. 65. cco ene ens cueess Smee ens ces 266 
b. The early flexure stage........ 050. ..0 sendin an diwe espera es es 267 
c.. The coiled-reaction stage... ..0 2.026. c rene enener meses cress 268 
d. The early swimming stage..............0css cess eee e en eneees 270 
4. The eye and optic nerve............. cece cece nsec cence neceeeeeeee 273 
5. The olfactory organ and nerve.............eec cece cece ne eeeeeeeee 275 
RA SMITE Ee oh Geddy cis Fach sia Wg « niet «sis aie Sp i he pu Slo eR NDS oe be 276 
Pig tne Dhyslolopial PATE... .) vaccines So0 week Cea sxe eta smeNe Eye's > ch 4 280 
1. Response to tactile stimulation...............00. see e cece ener ees 280 
2. Respense to chemical stimulation. ............6---+e eee e erences 285 
Si. Response, fo Tights... a5 <5 nies pa sic tes on ie aie nen gM RSE Bn yalo Sie ye oe 285 
4. Response to olfactory stimulation.............--.0e0e ee eee e eens 288 
5. The auditory and lateral line organS............,..02-sseeeeeeees 288 
es SSUNINAEY sted. «<a ss nals os vs ve nlite Re hinkee Oni Manenis wdiuive "e's y+ « 289 
MR PPMEMBOGSSION<:. . cPad neice s cx call oot oie Us nie Sp ah Sa ARERR sama eae os 290 
1. Lateral line primordia, placodes and ectodermal thickenings...... 291 
2. The function of the auditory vesicle................02.00eeee eee + 292 
3. The sensory nerve roots in relation to later development......-.- 293 
4. The development of function in the neurone and in the reflex arc.. 296 
5. The cranial nerves of Amphibia...............cceeeer eres eerie 297 
GaPNICIITODIOUAKIS Canc... tein ehtere fe tie atest cin Cheiatmrere stars cele ete ace 1s Fai, Aol 299 
247 


248 G. E. COGHILL 


The first paper of this series, ‘‘The Afferent System of the 
Trunk of Amblystoma,’’ was published in Volume 24, number 2 
(1914), of this Journal. It will be referred to in this communica- 
tion as Paper I. In it (p. 168) were described the four physio- 
logical stages of development upon which this paper, also, is 
based. The next paper of the series will deal with the neurones 
of the second and higher orders in the reflex ares of the brain 
and spinal cord. 


IT. THE ANATOMICAL PART 


This part of the paper involves a rather detailed description 
of the sense organs and sensory nerves of the head. This de- 
scription, however, is not exhaustive, and may require supple- 
menting as the study progresses. Readers who desire an even 
briefer treatment of this phase of the subject may turn to the 
summary of this part of the paper (p. 276) and to the figures, in 
connection with which salient anatomical features are mentioned. 


1. The trigeminal nerve 


A. THE NON-MOTILE STAGE 


In embryos of the non-motile stage the ophthalmic ganglion 
and the Gasserian ganglion are situated widely apart and come 
into relation with each other only through their roots as the latter 
approach the brain. The double nature of the trigeminal nerve, 
therefore, is very obvious. As shown in figure 1, the ophthalmic 
ganglion already extends forward over the eye and the Gasserian 
ganglion extends ventrad behind the eye. 

The ophthalmic ganglion is directly continuous with the skin 
near its rostral end (figs. 1 and 7 to 10, Ad.,G.oph.). Favorable 
plane of section reveals two rather distinct portions of this gan- 
glion (figs. 10, 11, 12, G.oph.a,b), and, at least in some specimens, 
each part has an area of adhesion with the skin. Two such 
areas are indicated in figure 1 (Ad.G.oph.). 

No peripheral distribution of fibers from the ophthalmic gan- 
elion to the skin can be made out at this stage. The neuro- 
blasts of the ganglion are bipolar and have perceptible axone and 


THE NERVOUS SYSTEM OF AMPHIBIA 249 


dendritic processes. The structure of one of these cells, taken 
from the most distal portion of the ganglion, is shown in figure 
18. The proximal portion of the cell is not clearly seen in the 
section figured, but the dendritic pole extends well out into 
mesenchyme which surrounds the ganglion. In adjacent sec- 
tions it can be followed as a delicate fibrillated strand of proto- 
plasm about three times the extent of the process as shown in 
the figure. The course of this fiber, however, is not directly 
towards the skin and it becomes lost to view among the cells of 
the mesenchyme. I have been unable to find in this stage of 
the embryos any highly branched endings of naked fibers like 
the terminals of the fibers of the giant ganglion cells of the trunk 
(Paper I, figs. 17, 23 to 25). It may be that exceedingly fine 
filaments of protoplasm from the ganglion cells reach the skin 
here and there; but if they do, they escape the most exhaustive 
search through favorable preparations. The only perceptible 
connection, therefore, of the ophthalmic ganglion with the skin 
in the non-motile stage is through the areas of adhesion, or con- 
tinuity with the ectoderm as described above. 

The attenuation of the ophthalmic ganglion caudad into a 
slender root is illustrated in figures 1, 7 and 8 (G.oph.,R.oph.). 
Caudally of the region indicated in figure 8 the root becomes 
more slender still, so that it is very difficult to identify among 
the mesenchyme cells. In frontal sections, however, it is easy 
to trace the root to the brain in two closely associated divisions 
which correspond to the two portions of the ganglion (figs. 11 
and 12). As it nears the brain the root bends dorsad slightly 
and approaches the brain in close association with the root of 
the Gasserian ganglion but at a slightly more ventral level than 
the latter. 

As illustrated in figure 1, the Gasserian ganglion lies well out 
from the brain with its long axis directed latero-ventrad. There 
is at. this time no perceptible development of fibers to form a 
nerve trunk beyond the distal limits of this ganglion (fig. 1). 
As in the ease of the ophthalmic ganglion, there is here also an 
area of adhesion between the ganglion and the skin. ‘This con- 
tact is at the distal end of the ganglion (fig. 9, Ad.G.G.). It is 


250 G. Be (COGHILE 


not as extensive or as intimate as the adhesion between the 
ophthalmic ganglion and the skin. In a transverse section of 
the embryo this adhesion is seen in figure 9 to be ventral of the 
primordium of the preauditory lateral line organs. 

In the first section (10 «) ventrally of that of figure 12 the root 
of the Gasserian ganglion enters the brain, while in the next 

following section (20 ”).the root of the ophthalmic ganglion 
enters. The axones of neither of these roots have grown beyond 
10 or 20 uw from the immediate zone of entrance. Throughout 
their extent, however, they form a dense fibrillar mass imme- 
diately beneath the external limiting membrane of the brain. 
There is no clear evidence of bifurcation of the root fibers at this 
stage. The axones which compose the root of the Gasserian 
ganglion are illustrated in figure 19. The cells figured here occur 
a considerable distance out from the brain and the longer processes 
are the axones. The axones of the two cells of the figure come 
into such close relation with each other that they might be inter- 
preted as fusing together into a single root fiber, but they prob- 
ably do not anastomose. The perikarya of these neurones or_ 
neuroblasts, like those of all other ganglia of this age, are filled 
with yolk spherules. 

As the roots of the ophthalmic and Gasserian ganglia approach 
the brain a small strand of fibers enters them from a more dorsal 
position in the brain. The neurones from which these fibers arise 
lie just inside the external limiting membrane and, mostly at 
least, in a more dorsal position than the root entrance. They 
have large spherical nuclei and the other morphological charac- 
teristics of the Rohon-Beard cells of the spinal cord. In figure 
20 (DC) one of the these cells is shown in a transverse section 
of the medulla oblongata, and 15 « eaudad of this appears another 
such cell, which is reproduced in figure 21. The former neurone 
sends its process to the immediate region of the entrance of the 
trigeminal root while the latter (fig. 21) sends its process through 
the external limiting membrane into the root of the nerve. 
There are in this section two or three other neurones of this 
type. At this stage the cell boundaries are rather indistinct, but 
the fibrils which compose the emerging fibers of figure 20 con- 


THE NERVOUS SYSTEM OF AMPHIBIA 251 


verge towards the limiting membrane from among the yolk 
globules that surround the nucleus. These yolk globules are 
intracellular and certainly belong to the cell of this nucleus, so 
that this strand of fibrils interspersed with yolk globules must 
be a process of the cell. 

That these neurones with processes running out into the tri- 
geminal nerve are not motor is clearly shown by reference to 
figure 20, in which the motor nucleus of the nerve is easily estab- 
lished in a much more ventral position (Nuc.Vis.m.). The cells 
in question, therefore, must be giant ganglion cells, representing 
an extension of the Rohon-Beard cells of the spinal cord into the 
brain. As to the peripheral distribution of the fibers from these 
cells, nothing can at present be said excepting that they appear 
to enter both divisions of the trigeminal nerve. 

In this and in some of the later stages there appears just out- 
side the brain and about the roots of the trigeminal nerve par- 
ticularly, but in a less degree about the roots of the other nerves, 
a mass of cells which is quite distinct from the ganglia. It is 
made up largely of indifferent cells, but cells resembling neuro- 
blasts occasionally occur in it. This structure I have called the 
root mass. It is indicated (R.M.) in figures 1, 2, 12, 15, 20 to 
22, 34 to 37, 47 to 49, 69. There are no apparent grounds for 
assigning to it any physiological significance, and I have made 
no effort to determine its origin, history or morphological sig- 
nificance. 


B. THE EARLY FLEXURE STAGE 


By reference to figures 1 and 2 it will be seen that the oph- 
thalmiec ganglion has drawn somewhat nearer the brain. The 
position of the ganglion with reference to the eye is further illus- 
trated in figure 13 (G.oph.), and the attenuated nature of its 
root at this stage is shown in figure 14 (R.oph.). The ganglion 
has now broken away from its adhesion with the skin, but thick- 
ened regions of the ectoderm indicate the areas of earlier attach- 
ment (Hc.Th.). One of these ectodermal thickenings is shown 
in figure 13. The two parts of the ophthalmic ganglion are still 
perceptible. i 


252 G. E. COGHILL 


The ophthalmic trunk has grown considerably rostrad from 
the ganglion (fig. 2, 0.p.V), and follows a course in conformity 
with the contour of the eye. It forms two divisions as it leaves 
the ganglion, these representing, apparently, the two parts of 
the ganglion that have been noted above. The larger of these 
two nerves follows closely the contour of the eye to the immediate 
region of the olfactory epithelium. Along its course several fila- 
ments arise which penetrate the surrounding mesenchyme, with 
the cells of which they become so intimately associated that they 
very rarely can be traced continuously to the skin. From among 
the mesenchyme cells, however, naked filaments emerge beneath 
the skin and attach themselves to it. Such filaments are in all 
probability the terminals of the n. ophthalmicus profundus. 
They have the appearance of terminals of the Rohon-Beard cells 
in the trunk, as described in' Paper I. It is possible that they 
are processes from sensory neurones in the medulla oblongata 
such as we have here under consideration. The smaller of the 
two divisions of the N. ophthalmicus profundus passes more 
directly forward (morphologically dorsad) and can be followed 
only indefinitely in its distribution (fig. 2, 7). 

The Gasserian ganglion still projects ventrad caudally ofthe 
eye (fig. 2, G.G.), and dorsally of the primordium of the muscles 
of mastication (fig. 15, G.G.). An infraorbital trunk now arises 
from the extremity of the ganglion and extends at least two- 
thirds of the distance across the lateral surface of this pri- 
mordium (fig. 2, Mdb.). An occasional: fiber arises from this 
nerve and passes towards the skin, but no definite cutaneous 
distribution can be made out for the nerve as a whole. 

The root fibers of the trigeminal nerve, upon entering the brain, 
form an ascending and descending trigeminal tract (fig. 22, 
Tr. asc.V,Tr.des.V). These apparently arise by a bifurcation of 
the root fibers, although one can not be positive of this upon the 
basis of the observations of individual fibers since the fibers are 
very compact in their grouping. Figure 22 shows favorably the 
extension of the trigeminal root fibers into the brain. The plane 
of section may be interpreted from figure 16, which was made 
from the same section. It will be seen to pass longitudinally 


THE NERVOUS SYSTEM OF AMPHIBIA 253 


through the medulla oblongata and to be inclined ventro-laterad 
from the middle line on the side viewed in f gure 22. This plane 
of section reveals characteristic ventricular pits opposite the 
entrance of the nerves of the trigeminal and facial groups. The 
extent of the ascending trigeminal tract is probably fully shown 
in figure 22 (T'r.asc.V.). It extends only a short distance rostrad 
immediately beneath the external limiting membrane. The de- 
scending trigeminal tract (7’r.des.V) is considerably longer and 
of a more complicated structure. As illustrated in figure 22, 
this tract runs from the trigeminal root to the level of the middle 
of the auditory vesicle immediately next to the external limiting 
membrane. ‘The development of the tract is even slightly exag- 
gerated in the figure. The trigeminal root fibers themselves 
appear to extend only about to the level indicated in the figure 
by the reference line 7'7.Des.V. Beyond this the tract appears 
to be made up of short processes of neuroblasts along its course, 
corresponding, presumably, to neurones of an incipient sub- 
stantia gelatinosa. The fibrillar element of this tract can be 
recognized only in favorable longitudinal sections. Its full de- 
scription belongs to a later paper. 

A gaint ganglion cell component of the trigeminal nerve exists 
also in this stage, but nothing noteworthy appears concerning 
either its central or peripheral relations. 


C. THE COILED-REACTION STAGE 


The ophthalmic ganglion has receded along the root till its 
proximal portion has become massed against the brain. Its rela- 
tion to the eye, also, is more intimate than in the earlier stages 
(figs. 3 and 76, G.oph.). The n. ophthalmicus profundus is per- 
ceptibly more developed than in the early-flexure stage and the 
distribution of its fibers to the skin is more obvious. In addition 
to the two main divisions branching out in front of the eye there 
are now fibers that go directly from the ganglion to the overlying 
skin. The root fibers can be traced into the brain and differen- 
tiated from the root fibers of the Gasserian ganglion, the latter 
entering in a slightly more dorsal position. 


254 G. E. COGHILL 


The Gasserian ganglion, also, has migrated towards the brain. 
The infraorbital trunk can be traced approximately to the ventral 
margin of the primordium of the muscles of mastication (fig. 
3, Mdb.) It now passes through a distinct groove in this pri- 
mordium. As it leaves this groove it gives off the ramus maxil- 
laris (Ma.V) which passes only a short distance forward under 
the eye. No fibers from the r. mandibularis can be traced into 
the balancer at this stage. The fibers to the skin in the post- 
optie region are still less obvious than they are in front of the 
eye, but favorable preparations show that they are present. 

Figure 33 illustrates the trigeminal root as it enters the brain. 
The plane of section of this figure may be judged by figure 17, 
in which the area of figure 33 is blocked out and marked a. 
This plane of section will be recognized as inclined cephalo-ventrad 
in a longitudinal direction. The small ascending division of the 
root (Tr.Asc.V) can be recognized in the figure and the appear- 
ance of the ascending tract in transverse section is shown in 
figure 34, which is 56 u» rostrally of the entrance of the root. Here 
the ascending fibers are found scattered through a considerable 
-area dorso-ventrally. The descending division of the root (fig. 
33, T'r.Des.B) is much more massive than the ascending, and 
extends caudad to about the level of the facial root, as shown in 
figure 5, sketched in solid black. The tract which is labeled 
Tr.Des.V. in figures 38 to 49 is apparently composed at these 
various levels of processes of central neurones. Its fibers are 
scattered through a considerable area dorso-ventrally and as such 
a diffuse system may be followed on caudad into the spinal - 
cord. 

The giant ganglion cell component of the nerve is more obvious 
in this than in earlier stages. In figure 33 one of these ganglion 
cells (DC) appears in the angle of birfurcation of the root of the 
ganglionic component. The cell boundaries are here clearly 
defined. The plane of section, already mentioned as illustrated 
in figure 17, is such that the rostral (right) end of the figure is 
considerable more ventral then the caudal (left) end, and motor 
cells appear in the more ventral portion of the figure (Nuc.vis.m.). 
In figures 35, 36 and 37, three successive sections in the trans- 


THE NERVOUS SYSTEM OF AMPHIBIA 255 


verse plane, through the entrance of the trigeminal nerve, the 
same type of cell appears. In figures 36 and 37 the same cell 
appears (DC). Its typical, large nucleus has two nucleoli, one 
in each section. In both sections the peripheral process, which 
' projects into the root, can be distinctly recognized. In figure 
36 @ process runs ventrad towards the motor centers, and in 
figure 37 there appears the basal portion of a process which is 
directed dorsad. This structure conforms to the Rohon-Beard 
cells which, in early stages, have one peripheral process and two 
central. Figure 35 shows the presence of several other cells of 
this type in the immediate vicinity of the root. The distinctly 
more ventral position of the motor nucleus (fig. 36. Nuc.vis.m.) 
precludes any possibility of confusing motor with sensory cells 
in this region. In figure 34 there is represented one of these 
cells which is situated still farther dorsad, and 56 » cephalad of 
the entrance of the root of the nerve. The long process running 
_ventrad from this cell indicates that it has become unipolar, after 
the manner of the Rohon-Beard cells in the spinal cord. 


D. THE EARLY SWIMMING STAGE 


In embryos of this stage the ophthalmic ganglion has pressed 
still farther back upon the root against the brain. It now has 
a wide contact with the Gasserian ganglion (fig. 4, G.oph.). 
Cephalad it extends nearly to the rostral border of the eye (fig. 
78). From the caudal portion of the ganglion there now arises 
a small nerve which goes dorsad to the skin. In some speci- 
mens this seems to be represented by a single fiber. It appears 
as such in silver impregnations. This is probably my branch ‘x’ 
of the advanced larva and adult (’02). Near the rostral border of 
the eye (fig. 4, L.o.p.V), a branch, consisting of only a few 
fibers, passes laterad around the anterior surface of the eye to 
a position lateral and caudal of the external nares. This is pre- 
sumably my lateral terminal branch of the adult (’02, fig. 2, 
l.o.p.V). The remainder of the nerve passes along the mesial 
aspect of the olfactory epithelium to the skin of that region. 
This must be the mesial terminal branch of the adult. A con- 


256 G. E. COGHILL 


siderable number of fibers of this nerve attach themselves in a 
cluster to the supraorbital lateral line primordium, to which the 
course of the nerve has a definite relation. These branches from 
the ophthalmic nerve have all been identified in silver impreg- 
nations. 

The root fibers of the ophthalmic ganglion can be traced to the 
brain and can be seen to enter the brain a little ventrad and 
cephalad of the entrance of the root of the Gasserian ganglion 
(figs. 51 to 538, R.oph). The intimate relation between the proxi- 
mal portions of these two ganglia is shown in these figures. 

The Gasserian ganglion, like the ophthalmic, has become more 
consolidated about the proximal portion of its root than in any 
of the earlier stages. The infraorbital trunk has become still 
more deeply embedded in the groove through the primordium 
of the jaw muscles, across the entire lateral surface of which it 
now descends to the vicinity of the balancer. ‘To the balancer 
it sends a nerve (fig. 4, c), which can be traced in both stained 
preparations and silver impregnations, practically to the tip of 
the organ where it has cutaneous distribution. About where 
the main nerve leaves the groove it gives off the ramus maxillaris 
(fig. 4, Mx.V) which now passes a considerable distance forward 
under the eye. There is some evidence that this nerve arises 
from cells that le in the angle between the main projections of 
the ophthalmic and Gasserian ganglia. Soon after giving off 
the nerve to the balancer the ramus mandibularis divides into 
two branches which turn cephalad and mesad ventrally of the 
muscle primordia. These can be traced into close proximity 
with the oral plate. In silver impregnations of embryos of this 
age neurofibrils may be seen penetrating the skin along the 
course of this nerve. The cutaneous distribution of the nerve, 
however, is still less obvious than is that of the ophthalmic divi- 
sion, excepting in the balancer, which is a highly innervated 
organ at this stage. 

The root fibers of the Gasserian ganglion are shown in figures 
51 to 54 as they enter the brain to form the more dorsal portion 
of the ascending and descending trigeminal tracts. The extent 
of these tracts is shown in figure 6 (Tr. asc.V,Tr.des.V). Silver 


THE NERVOUS SYSTEM OF AMPHIBIA PAS 


impregnations, which differentiate these root fibers clearly trom 
the smaller fibrils of the region, show them running cephalad 
almost through the metencephalon and caudad beyond the level 
of the auditory vesicle. Along this course fibers of smaller caliber 
group themselves along the deeper aspect of the root fibers. The 
tract is, therefore, a composite of sensory neurones of the first 
and second orders. Just how far the root fibers extend can not 
be exactly determined, but they appear to reach the level of 
the general cutaneous root of the vagus, and some of them may 
even reach the spinal cord. The descending trigeminal tract as 
a whole is continuous with the sensory tract of the spinal cord 
(figs. 6, 66, DT, Tr.Des.V). 

In embryos of this age silver impregnations have been made 
which show the processes of giant ganglion cells extending con- 
tinuously from the cell body into the root of the trigeminal nerve. 
Cells of this description are situated dorsally of the root and in 
the root entrance zone in figures 51 to 54 (DC). In figure 50 
there is a representative group of these cells which aresituated 
35 w caudad of the root of the nerve. The ventrally directed 
processes of the cells are here shown in their relation to the 
deeper face of the descending trigeminal tract. 

In figures 51 to 54 there is illustrated, also, in a general way 
the degree of differentiation of neuroblasts and neurones in the 
marginal zone of the substantia grisea, and the elevated regions 
in the ventricular floor in some parts as compared with the rela- 
tive flatness in others. The study of the interrelations of such 
centers of differentiation and ventricular eminences belongs to 
alater paper. It may be added here, however, that fine terminals 
from the trigeminal root fibers ramify among the neurones which 
are grouped along the trigeminal tract. 


2. The facial and auditory nerves 


In all four stages of development here under consideration the 
various elements of this ganglionic complex can be recognized: 
the lateral line and visceral components of the facial and the 
auditory. 


258 G. E. COGHILL 
A. THE NON-MOTILE STAGE 


1. The lateral line component. ‘The position and extent of 
development of the primordia of the various lateral lines of the 
head are shown in Paper I, figure 56 (Pr.LL.). The primordium 
of the preauditory lines consists of a single boot-shaped patch 
of thickened epithelium: the toe of the ‘boot’ extending forward 
and a little dorsad as the beginning of the supraorbital line 
and the heel projecting ventrad as the beginning of the infra- 
orbital line. From the caudal end of this structure a narrow 
strip of thickened epithelium connects it with the auditory 
vesicle, which still has a broad attachment with the skin (fig. 
73, Ad.,Aud.V). This slender connective between the auditory 
vesicle and the lateral line primordium has the structure of the 
thickening of ectoderm which persists for a time after the ophth- 
almic ganglion detaches itself from the skin. It may represent 
the area of origin of a dorso-lateral placode, although I have no 
conclusive evidence of this. 

The facial lateral line ganglia of this stage are rather loose 
masses of cells to which only indefinite borders can be estab- 
lished (fig. 1). The mass of cells which represents the hyo- 
mandibular division of the nerve extends a considerable distance 
caudad under the auditory vesicle, with which it seems to be 
fused in a small area near the auditory ganglion. From this 
region it projects laterad and rostrad as a loose collection of cells 
which has a very indefinite relation to the skin. The other 
lateral line facial ganglion, which later gives rise to the r. ophthal- 
micus superficialis VII and r. bucealis VII has a more definite 
outline and is somewhat more highly differentiated. It holds a 
more dorsal position and extends rostrad to the lateral line pri- 
mordium, with which it is in direct contact. A considerable 
number of fibers from the ganglion attach themselves to the pri- 
mordium, upon which they can be traced for only a very short 
distance. 

The lateral line ganglia have root fibers which reach to the 
surface of the brain at about the level of the rostral border of the 


THE NERVOUS SYSTEM OF AMPHIBIA 259 


auditory vesicle, but within the brain these fibers have no per- 
ceptible extension. 

2. The visceral sensory component of the facial nerve. The gan- 
glion of this component, also consists of rather scattered cells 
which occupy a position ventrally of the lateral line ganglia. 
It projects laterad and ventrad to the ectodermal thickening 
which is associated with the spiracular pouch (Paper I, fig. 56 
Ec.Th.). With the cells of this thickening the ganglion is prac- 
tically continuous. There are no perceptible nerve fibers in the 
peripheral portion of the ganglion or arising from it to forma 
nerve. Centrally there appear to be a few root fibers which 
reach the brain, but it has so far been impossible to differentiate 
them certainly from the other root fibers of the complex or to 
establish that they are constantly developed at this time. 

3. The auditory nerve and vesicle. The auditory vesicle of this 
stage has a broad attachment to the skin (fig. 78, Ad.,Aud.V). 
This contact is with the lateral aspect of the vesicle. The endo- 
lymphatic appendage (Mnd.) is very imperfectly differentiated 
from the rest of the vesicle, but it can be readily recognized. 
The epithelium of the mesial, ventral and ventro-lateral regions 
of the vesicle is still very thick, and mitosis is particularly abun- 
dant in the more ventral portion. The ventro-mesial portion of 
the vesicle is continuous with the auditory ganglion (fig. 73, 
G.VIIT). As the ganglion projects rostrad it become free from 
the vesicle and a fiber or two from its proximal end reach the 
brain at a point more ventral and caudal than the root connec- 
tion of the lateral line ganglia with the brain. 


B. THE EARLY FLEXURE STAGE 


All the ganglia of this complex have acquired more definite 
- contours and greater compactness of structure. The interrela- 
tions of the ganglia can be most favorably studied in frontal 
sections; the root connections with the brain, in transverse 
sections. 

1. The lateral line component. The two, distinct lateral line 
ganglia I distinguish as a and b (fig. 2, G.L.L.VIJ,a and b). The 


260 G. E. COGHILL 


degree of separation of these two ganglia is somewhat exaggerated 
in figure 2. Ganglion a is distinctly farther advanced in develop- 
ment than ganglion b. In horizontal sections it exhibits a line 
of cleavage between two portions, and each portion gives rise to 
a distinct fiber bundle. These two bundles emerge from the 
distal end of the ganglion and obviously represent the r. oph- 
thalmicus superficialis VII and the r. buccalis VII. The latter 
is directed the more ventrad. ‘There are now sheath cells upon 
the fiber bundle from the ganglion to the lateral line primordium, 
and closely attached to the latter the naked fibers extend for 
some distance in two strands, representing the two nerves as 
indicated. The primordia of the lateral lines have shifted farther 
rostrad relative to the position of the auditory vesicle, from 
which they have become completely detached (Paper I, fig. 57, 
Pr.LL) ‘Their position, also, is somewhat more dorsal. Their 
ventral projection is distinctly larger and shows a tendency to 
form into two divisions. 

Ganglion b, of the hyomandibular division, is still very imper- 
fectly differentiated in its proximal portion from the auditory 
ganglion. Distally it is seen in frontal sections along the caudal 
and lateral border of the geniculate ganglion (fig. 2, G.gen.), 
the point of it extending ventrad behind the attachment of the 
entoderm of the spiracular pouch to the ectoderm. From this 
point of the ganglion a few fibers extend on ventrad and into 
the primordium of the hyoid arch. These fibers can be recog- 
nized only as they run in the plane of section. The fact that 
these fibers have no perceptible connection with the skin and 
that the primordium of the mandibular lateral line has not yet 
appeared leads me to interpret them as probably motor. 

The root fibers of the lateral line ganglia still enter the brain 
at a slightly more rostral level than the auditory vesicle, and in 
two divisions (figs. 23 to 28, R.L.L.VII,a and 6). The dorsal 
division seems to arise from ganglion a; the ventral division, 
from ganglion b. In figure 23 a few ascending fibers of each 
division can be seen (L.L.VII,Asc.,a,b). In figure 28 the corre- 
sponding descending divisions of these root fibers are repre- 
sented (L.L.VII,Des.,a and 6). Presumably the fibers bifurcate 


THE NERVOUS SYSTEM OF AMPHIBIA 261 


upon entering the brain at this stage, although this has not been 
actually demonstrated in individual fibers. 

2. The visceral sensory component. ‘The geniculate ganglion in 
its proximal portion holds a relatively ventral position in the 
ganglionic complex. Distally it reaches ventrad and rostrad to 
the region of contact between the ectoderm and entoderm of the 
spiracular pouch, where it becomes continuous with a cluster of 
ectodermal cells which has pushed in over the entoderm. ‘This 
cluster of cells is, presumably, the epibranchial placode of this 
nerve. No fibers can be seen issuing from the geniculate gan- 
glion at this stage. The root fibers of the geniculate ganglion 
are shown approaching the brain in figure 25 (R.VJJ,vis.) at a 
distinctly more ventral level than the lateral line root. They 
turn caudad immediately within the external limiting membrane 
and form an incipient fasciculus solitarius (figs. 26, 27). Their 
visible course is limited to the sections figured in this series. 

3. The auditory organ and nerve. The auditory ganglion has 
become completely detached from the vesicle. In figure 75, a 
single cell (VJII) represents the distal end of the ganglion. 
Towards the brain from this position the ganglion ean be fol- 
lowed distinctly (fig. 2, G.V/JJI). The auditory vesicle has 
detached itself entirely from the skin and has made very per- 
ceptible progress in differentiation since the non-motile stage, as 
may be seen by comparing figures 73 and 75. Mitotic figures 
are still very abundant in the inner portion of the epithelium. 
The root fibers of the nerve are seen in figure 27 (R. VIII) enter- 
ing the brain on the ventral aspect of the root of the geniculate 
ganglion. In figure 28, taken from the next section caudad of 
27 (10 u) no descending auditory root fibers can be definitely 
recognized. 


C. THE COILED-REACTION STAGE 


As in the case of the trigeminal ganglia, the ganglia of the facial 
complex have become more consolidated towards the brain and 
around the antero-ventral aspect of the auditory vesicle. 

1. The lateral line component. The lateral line primordia have 
become greatly extended since the early flexure stage (Paper I, 


THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL. 26, No. 3 


262 G. E. COGHILL 


fig. 58). In addition to the primordia shown in the latter figure 
there is now a primordium which runs mesad from the base of 
the balancer on the ventral surface of the head. In this epi- 
thelial thickening there is a lowly differentiated lateral line organ 
at the mesial end and possibly another at the base of the balancer. 
This primordium is not connected with the other preauditory 
primordia, unless it is by a very slight ectodermal thickening of 
an indifferent nature. 

The lateral line ganglia a and 6 are in close contact with each 
other proximally but project distally into distinct ganglia (fig. 3, 
G.L.L.VII,a,b). Ganglion a still reaches almost to the skin. 
In its distal portion it still has two parts, from each of which 
arises anerve. These nerves follow the primordia closely attached 
to the skin for a considerable distance. ‘The finer divisions have 
no sheath cells and their exact length is therefore difficult to 
determine. Ganglion 6 projects ventrad behind the spiracular 
pouch and from it nerve fibers pass to the skin at or near the 
lateral line primordium which has been described as occurring 
in this stage at the base of the balancer. 

The roots of the lateral line ganglia enter the brain as illus- 
trated in figures 38 to 41, figure 38 being the most rostral of the 
series. Here the root fibers of ganglion 6b (R.L.L.VII,b) are 
entering the brain and collecting into a bundle immediately 
beneath the external limiting membrance. In figures 39, 40, 
and 41, the next successive sections caudad, the root fibers of 

ganglion a are entering in a more dorsal position (PR.L.L.VIT/,a). 
- Within the brain these fibers appear to bifurcate. At least they 
form short ascending and descending tracts as shown in figure 
5 (L.L.Asc.,Des.,a and 6). Figure 38 shows the ascending tracts, 
and figure 41, the descending tracts in cross section, under high 
magnification. 

2. The visceral sensory component. The proximal portion of the 
geniculate ganglion has not changed its position noticeably (fig. 
3, G. gen.). Distally it projects to the spiracular pouch, where 
its appearance has materially changed. Along the caudal aspect 
of the entodermal wall of the pouch a fiber or two may be traced 
along the bases of the entodermal cells. Their course, however, 


THE NERVOUS SYSTEM OF AMPHIBIA ‘968 


is very short, for they can not be traced into the hyomandibular 
trunk. They take a course in conformity with this trunk, and 
probably represent an incipient r. alveolaris. This part of the 
ganglion apparently corresponds to the geniculate ganglion which 
has been described in the earlier stages. In addition to this 
there is now a very massive connection of the ganglion with the 
ectoderm running forward over the pouch and at the same time 
in close relation with the entoderm. The structure here gives 
one the impression that the ectoderm is contributing large num- 
bers of cells to the ganglion. This is probably the epibranchial 
placode, although my specimens have not been selected at short 
intervals and are not adequate for the deciding of this question. 
No fibers emerge from this portion of the ganglion. 

The root of the geniculate ganglion is shown entering the 
brain in figure 39 (R.VJJ,vis.). Within the brain they turn 
eaudad and form a distinct fasciculus solitarius (figs., 39 to 41, 
Fas.Sol.), which lies against the limiting membrane and wedged 
in between the lateral line tracts dorsally and the auditory tract 
ventrally. This tract can now be followed to the level of the 
glossopharyngeal root, from which it receives an increment, and 
then on to the level of the root of the vagus nerve (fig. 5). 

3. The auditory organ and nerve. The wall of the auditory 
vesicle has become greatly reduced in thickness and its lumen 
much enlarged since the early flexure stage (fig. 77). In the 
inner portion of the ventral wall mitosis is still going on rapidly. 
The auditory ganglion is closely applied to the vesicle ventro- 
mesially, while the endolymphatic appendage is closely pressed 
against the brain. This contact is obviously a secondary thing, 
brought about by the enlargement of the vesicle (compare figs. 
73, 75, 77, 79). The auditory nerve can be traced to the caudal 
extremity of the vesicle, closely applied to the vesicle in about 
the same relative position as it appears in figure 77. The root 
fibers (R.VIIZ) enter the brain as shown in figures 40 and 41 
ventrally of the root fibers of the geniculate ganglion. ‘Their 
course caudad on the ventral aspect of the fasciculus solitarius is 
indicated in figure 5 (VJITJ,des.). 


264 G. E. COGHILL 


D. THE EARLY SWIMMING STAGE 


The facial and auditory ganglia have in this stage become still 
more consolidated about the auditory vesicle and the roots enter 
the brain farther caudad relative to the vesicle. 

1. The lateral line component. ‘The preauditory lateral line 
primordia have advanced greatly in distribution and differentia- 
tion (Paper I, fig. 59). The supraorbital primordium can now 
be traced far down in front of the eye and the olfactory organ 
and spreads out at the end, in anticipation apparently of the 
line of organs that develops later around the rim of the snout. 
The suborbital primordium extends beneath the eye into close 
relation with the olfactory organ, while the proximal portion of 
this primordium is now continuous with the primordium of the 
hyomandibular region. The latter primordium now passes ven- 
trad behind the balancer, then mesad, where it appears on the 
ventral ‘surface behind the oral plate in two distinct ridges, in 
which the characteristic organ structure-is well pronounced. 

Silver impregnations of the lateral line fibers now show them 
extending practically the full length of the primordia, and follow- 
ing rigidly the course of the primordia as in the earlier stages. 
The nerve trunks are everywhere closely adhering to the epi- 
thelium and here and there fine fibrils can be followed in among 
the cells. 

The root fibers of the lateral line ganglia impregnate much 
more deeply with the silver methods than do the peripheral 
fibers. ‘They enter the brain as illustrated in figures 55 to 60, 
in two roots, R.L.L.VII, a and b). From a study of silver im- 
pregnations of these fibers it is impossible to be absolutely sure 
that all fibers from ganglion a enter root a, or that root 6 is 
exclusively made up of fibers from ganglion b. But if there is 
an intermingling of fibers from the different ganglia in the indi- 
vidual roots this must be very slight. The roots, upon entering 
the brain, form ascending and descending tracts, as shown in figure 
6. The descending tracts reach the level of the middle of the 
auditory vesicle, and the ascending tracts are of about the same 
length or a little longer. The relation of the descending tracts 


THE NERVOUS SYSTEM OF AMPHIBIA 265 


of the facial to the ascending tracts of the postauditory lateral 
line roots is shown in figure 61, where the tract VII a is most 
dorsal of the series. Between this and tract VJJ 6 is the dorsal 
division of the postauditory roots. Most ventral in the set is 
the ventral division of the latter. Good silver impregnations 
in longitudinal section make it appear that these tracts may all 
be considerably longer than figure 6 represents them. 

2. The visceral sensory component. ‘The geniculate ganglion is 
now in contact with the ectoderm by only a slender column of 
cells which rests also upon the entoderm. ‘This ectodermal con- 
tact of the ganglion lies about midway between two other pro- 
jections of the ganglion: the one contributing fibers to the r. 
hyomandibularis as the latter passes ventrad in very close rela- 
tion with the caudal wall of the spiracular pouch (fig. 4, e); the 
other, directed rostrad on the dorsal surface of the entoderm, 
giving off a well defined r. palatinus (fig. 4, d@). This nerve passes 
mesad and a little rostrad over the pharyngeal cavity almost to 
the middle line and forward to within abougt 50 u of the rostral 
end of the foregut. 

The root fibers of the geniculate ganglion enter the brain ven- 
trally of the lateral line roots, as seen in figure 56 (R.VII,Vis.) 
and then turn caudad as the fasciculus solitarius (fig. 6, Fas.Sol.). 
The tract still lies immediately against the limiting membrane. 
It is distinctly larger than in the coiled-reaction stage. 

3. The auditory organ and nerve. ‘The endolymphatic append- 
age is more sharply differentiated from the body of the auditory 
vesicle than it was, in the earlier stage, and is still closely pressed 
against the brain (fig. 79). The auditory ganglion and nerve 
extend to the caudal border of the vesicle in the position shown 
in figure 79. In silver impregnations the fibers of the nerve now 
appear as exceedingly fine filaments which become closely merged 
with the basal ends of the epithelial cells or with a basement 
membrane which is also slightly impregnated. Nerve ending 
upon the epithelial cells of the vesicle, however, have not been 
observed. 

The root fibers of the auditory ganglion, entering the brain as 
shown in figures 57 to 60, form a descending tract which extends 


266 G. E. COGHILL 


along the ventral aspect of the fasciculus solitarius to the level 
of about the middle of the auditory vesicle (fig. 6). 


38. The glossopharyngeal and vagus nerves 


This complex includes lateral line, general cutaneous and vis- 
ceral components. 


A. THE NON-MOTILE STAGE 


1. The lateral line component. ‘There are two postauditory 
lateral line primordia (Paper I, fig. 56, Po.LL.1 and 2). The one 
of these is closely associated with the ectodermal thickening over 
the first branchial pouch; the other lies along dorsally of the 
thickening of the ectoderm in the vicinity of the more caudal 
pouches. From the primordium over the first branchial pouch a 
slender strand of cells reaches to the brain (fig. 1, G.L.0.IX, 
R.L.L.IX,X). These cells are laden with yolk spherules and are 
mostly of an indifferent nature, although some of them have the 
appearance of neuroblasts. This is the lateral line ganglion of 
the glossopharyngeus. While the root reaches the brain, nothing 
can be determined concerning its fibers within the brain. 

The main lateral line ganglion of the vagus is a spindle-form 
body, situated opposite the first and second myotomes. It has 
a broad attachment with the lateral line primordium of that 
region. Its root is so slightly developed that it can be recog- 
nized only as it lies in the plane of section. A very few root 
fibers reach the brain in connection with the lateral line root of 
the ninth nerve (fig. 1, G.L.L.X). 

2. The general cutaneous component. On the lateral aspect of 
the lateral line ganglion of the vagus is a loose aggregation of 
cells which extends caudad into very close relation if not into 
contact with an ectodermal thickening which is practically con- 
tinuous with the lateral line primordium associated with the 
vagus nerve. This cluster of cells, representing the jugular 
ganglion of the vagus, is slightly more condensed in its most 
rostral portion, which bends mesad ventrally of the lateral line 


THE NERVOUS SYSTEM OF AMPHIBIA 267 


ganglion. This incipient root, however, does not reach the brain 
eng TG. J uge): 

3. The visceral sensory component. The visceral ganglion of 
the glossopharyngeus fuses with the ectodermal thickening which 
rests upon the entoderm of the first branchial pouch (Paper I, 
fig. 56, Hc.Th.). This structure is obviously the epibranchial 
placode of the glossopharyngeus. From it a strand of ganglion 
cells follows along the lateral line root to the brain. No definite 
relation of root fibers, however, can be made out (fig. 1, G.Vis.JX). 

The visceral ganglion of the vagus nerve is in a still more 
embryonic condition. It consists for the most part of a loose 
collection of cells situated on the ventral side of the lateral line 
and general cutaneous ganglia. The most highly differentiated 
part of it reaches out to the ectodermal thickening over the second 
branchial pouch, with which it fuses. Farther caudad there is 
loose connection of the ganglion with the ectodermal thickening 
over the third and fourth visceral pouches. ‘The ganglion at this 


time has no root connection with the brain (fig. 1, G.Vis.X). 


B. THE EARLY FLEXURE STAGE 


1. The lateral line component. The lateral line primordium, 
which is situated over the first branchial pouch, has become 
elongated dorso-ventrally, and two other primordia appear, the 
one situated over the second branchial pouch opposite the first 
myotome and the other opposite the second to fifth myotomes 
(Paper I, fig. 57). 

The lateral line ganglion on the ninth nerve has now become 
spindle form, but it still reaches to the primordium (fig. 2, 
G.L.L.IX). Its root fibers reach the brain in close contact with 
the root fibers from the lateral line ganglion on the vagus. ‘The 
latter ganglion now sends out a projection of cells which reaches 
to the primordium which is situated over the second branchial 
pouch, and extends caudad to the primordium of the trunk 
(fig. 2). 

The roots of these postauditory lateral line ganglia approach 
the brain together (figs. 29 and 30). In figure 30 this root 


268 G. E. COGHILL 


appears in two parts, a dorsal, from the ganglion on the vagus, 
and a ventral, from the ganglion on the ninth nerve. There are 
no perceptible ascending and descending divisions of these root 
fibers within the brain at this stage. 

2. The general cutaneous component. The jugular ganglion 
holds about the same position relative to the other ganglia as in 
the younger stage, although there is now no apparent connec- 
tion with the skin in its more caudal portion. It does’ connect 
with the skin, however, at a point near the second post-auditory 
lateral line primordium. ‘This strand of cells is in very close 
association with the ganglionic projection to this primordium. 
It forms an incipient r. auricularis vagi. 

3. The visceral sensory component. The epibranchial placode 
over the first branchial pouch is now massive and has a broad 
connection with the visceral ganglion of the ninth nerve (fig. 2, 
G.Vis.IX). It is now contributing cells, apparently, to the gan- 
glion. The latter extends along the lateral line division of the 
nerve to the brain. Its attachment to the brain is shown in 
figures 30 and 31 (R.Vis.LX). 

A projection of the visceral ganglion of the vagus still con- 
nects with a placode-like mass of cells situated over the second 
branchial pouch. The larger portion of the ganglion projects 
caudad and ventrad towards the other visceral pouches, near 
which it ends in indefinite clusters of cells. A slender strand of 
indifferent cells reaches from the ganglion to the brain, in which 
there may be a few root fibers (fig. 32, R.X). In figure 32 the 
strand of cells which represents the root (G@.X), extends caudad 
between the brain and the myotome (M). Here a single root 
fiber appears. This may, however, be a motor fiber. No root 
fibers can be positively recognized as arising from cells in the 
ganglion. 


C. THE COILED-REACTION STAGE 


1. The lateral line component. ‘The lateral line ganglion of the 
ninth nerve still reaches to the skin, but a relatively large strand 
of fibers without sheath cells now spreads across the primordium. 
This is an incipient r. supratemporalis IX. The ganglion has 


THE NERVOUS SYSTEM OF AMPHIBIA 269 


enlarged greatly and stretches along the root till it comes into 
close relation with the lateral line root of the vagus (fig. 3). 
There is, then, a considerable advance made towards the con- 
solidation of all the postauditory lateral line ganglia into a com- 
mon mass (compare Goghill, ’02, figs. 1 and 2). The ganglionic 
projection, however, which reaches out on the r. auricularis vagi 
now amounts practically to a distinct ganglion with its own root 
connection with the main lateral line root of the vagus. In the 
terminal portion of this smaller ganglion fibers connect with the 
primordium, but this nerve does not seem to be quite as well 
developed as the r. supratemporalis. It is a question whether 
these two nerves innervate a common primordium (Paper I, 
fig. 58) or whether there may not be two primordia here sepa- 
rated by an ectodermal thickening to which the jugular ganglion 
connects. 

The primorida of the trunk now appear in two distinct patches 
(Paper I, fig. 58, LL.). One of these lies opposite the third and 
fourth myotomes and the other opposite the seventh to the tenth 
myotomes. To both of these the main lateral line ganglion of 
the vagus sends nerves, which separate from each other a short 
distance from the ganglion. 

Figures 42 to 46 represent five successive serial section which 
show the lateral line roots entering the brain. ‘There are still 
two divisions of the roots, a dorsal and ventral, marked, respec- 
tively, a and b. Each division forms an ascending and a de- 
scending tract, the longitudinal extent of which is shown in 
figure 5. 

2. The general cutaneous component. The jugular ganglion still 
lies on the lateral aspect of the proximal portion of the lateral 
line ganglion of the vagus as an ill-defined group of cells. It 
reaches the skin as a ganglion rather than as a nerve, in close 
relation with the small lateral line ganglion already mentioned. 
Its point of contact with the skin seems to be between the lateral 
line primordium innervated by the latter ganglion and that inner- 
vated by the r. supratemporalis. It projects also caudad in 
loose relation with the visceral ganglion, but its relations in this 
region are very indefinite. A root-like projection of the ganglion 


270 G. B. COGHILL 


passes ventrally of the lateralis ganglion and through the V- 
shaped cleft of the first myotome, but no fibers can be traced 
from this region into the brain. 

3. The visceral sensory component. ‘The visceral ganglion of 
the ninth nerve has enlarged greatly since the last stage, in which 
it appeared to be receiving cells from the epibranchial placode. 
It now touches the auditory vesicle (fig. 3, G.Vis.JX). Distally, 
however, it still reaches to the entoderm, where it has a very 
indefinite border. Throughout its distal portion it shows two 
rather distinct portions, the more dorsal of which projects the 
farther over the entoderm. No nerve trunks are yet to be found 
arising from this ganglion. It has a root, however, which enters 
the brain as a relatively strong bundle of fibers on the ventral 
aspect of the lateral line roots. ‘This relation is shown in figures 
44 and 45 (R.Vis..X). The fasciculus solitarius (Fas.Sol.) is 
seen in figure 42 rostrally of the root. Caudally of the root it 
appears in figures 45 and 46, where it is larger. The most of 
the visceral sensory root fibers must, therefore, turn caudad in 
the fasciculus (fig. 5). 

The visceral ganglion of the vagus still reaches out to the 
branchial pouches. No peripheral nerves yet arise from it. Its 
root enters the brain as shown in figures 47 to 49. Figure 47 
shows ascending vagus fibers just ventrally of the glossopharyn- 
geal portion of the fasciculus solitarius (X). Figure 49 shows 
that the fasciculus solitarius (Fas.Sol.) is smaller on the caudal 
side of the root. It is apparent from this that the visceral sen- 
sory fibers are directed chiefly cephalad in the fasciculus at this 
time. The fasciculus solitarius extends only a short distance 
caudad of the root (fig. 5, R.Vis.X). 


D. THE EARLY SWIMMING STAGE 


1. The lateral line component. In addition to the lateral line 
primordia of the trunk there are now two distinct postauditory 
primordia, the one innervated by the r. supratemporalis, the 
other by the lateralis component of the r. auricularis vagi (fig. 4; 
Paper I, fig. 59). The lateral line can be traced clearly as far 


THE NERVOUS SYSTEM OF AMPHIBIA 271 


caudad as the seventeenth myotome, almost the full extent of 
the primordia. Fibers reach the primordium of the inferior line, 
also. The small lateral line ganglion which is associated with 
the r. auricularis vagi still stands out distinctly. The root 
fibers from the ganglia on the vagus pass to the brain on the 
dorsal surface of the root of the ninth nerve. ‘The root fibers 
have extended their tracts as indicated in figure 6 (L.L.IX,X 
Asc.,Des.). The relation of the root fibers to the tracts is shown 
in figures 62 to 65. The roots here are very compact bundles 
of fibers. There is no apparent interchange of fibers between 
the roots from the different ganglia. The dorsal tract seems to 
represent the vagus ganglion and the ventral tract the glosso- 
pharyngeal ganglion. The ventral tract runs rostrad in very 
close relation to the fasciculus solitarius and at the level of the 
middle of the auditory vesicle takes a position ventrally of the 
lateral line tracts of the facial nerve (fig. 61, L.D.1X,X,b). 

2. The general cutaneous component. The jugular ganglion 
holds about the same position relative to the other ganglia as in 
the earlier stages. Only in favorable preparations can it be dis- 
tinguished from the visceral ganglion, over which it lies in the 
form of a rather thin cap. It gives rise to the r. auricularis vagi, 
which is still a very short nerve. The exact distribution of the 
fibers can not be made out. General cutaneous fibers join also 
the projections of the visceral ganglion which reach out to the 
second and third branchial pouches. It has been impossible to 
demonstrate fibers from the jugular ganglion to the external gills, 
to the skin around the base of the gills or to the ventral surfaces 
of this part of the head. The cutaneous innervation, so far as 
the ganglionic system is concerned, is in a much lower state of 
development in the postauditory region than it is in the pre- 
auditory. 

The root of the jugular ganglion now reaches the brain, pass- 
ing across the dorsal surface of the visceral root from the rostral 
to the caudal side (fig. 4, R.G.Jug.). It enters the brain at a 
slightly more ventral level than does the visceral root, as shown 
in figures 67 to 69 (R.G.Jug., fig. 69). This root is continuous 
from the ganglion to the brain surface with the spinal accessory 


272 G. E. COGHILL 


root (fig. 4, R,X/), which in this stage extends caudad to the 
level of the third myotome (compare Coghill ’02, fig. 1, root 
X4,a,b,c,d). The latter root is probably largely motor in the 
early swimming stage, although cells that look like neuroblasts 
occasionally occur in it. The root of the jugular ganglion enters 
the descending trigeminal tract, as shown in figure 6. The diffuse 
nature of the tract in this region is illustrated in figures 66 to 
70 (Tr.Des.V). 

3. The visceral sensory component. 'The end of the glosso- 
pharyngeal ganglion still rests upon the entoderm, but there is 
now a postbranchial nerve arising from it which extends ventrad 
behind the pouch and runs a considerable distance rostrad in the 
primordium of the first branchial bar (fig. 4, G.Vis.LX,g). This 
is the ramus lingualis. 

No fibers can be traced from this nerve to the skin. Neither 
do fibers pass from it into the entoderm. Nevertheless it must 
be primarily visceral sensory, although it may at this time con- 
tain motor fibers also. ‘The nerve adheres in its course very 
closely to muscle primordia, just as do the r. mandibularis V, 
and the r. hyomandibularis VII of this age. 

The root of this nerve still enters the brain in close contact 
with the ventral surface of the lateral line component (figs. 4 
63, 64). Figures 62 and 65, respectively, show the condition 
just rostrally and caudally of the entrance of the root fibers, which 
join the fasciculus solitarius (Fas.Sol.). 

The projections of the vagus ganglion have not yet completely 
lost their connection with the entoderm of the first and second 
branchial pouches, while the caudal portion of the ganglion frays 
out into loose cells over the branchial region. The first n. 
branchialis vagi passes some distance into the primordium of the 
second gill arch behind the corresponding pouch, but it is not as 
well developed as the r. lingualis of the glossopharyngeus (fig. 
4, G.Vis.X,h). No other nerves can be found arising from the 
ganglion at this time. There appear to be no pharyngeal or 
prebranchial divisions of any of the postauditory branchial nerves 
in the early swimming stage. 


THE NERVOUS SYSTEM OF AMPHIBIA 273 


The root of the vagus ganglion is distinctly more developed 
than in the earlier stage. As it approaches the brain it turns 
distinctly dorsad from beneath the general cutaneous root (fig. 3, 
67 to 69, R.Vis.X). These root fibers must have a very short 
course within the brain, for the fasciculus solitarius (Fas.Sol.) 
diminishes considerably in thickness within the space of 20 u 
from the root entrance, as shown in figure 66. <A few fibers of 
the fasciculus, however, extend caudad for some distance (fig. 6). 
Dorsally of the fasciculus solitarius are found a few fibers of the 
sensory ascending tract from the spinal cord (figs. 66, 67, 68, 70, 
DT); and ventrally of it is the descending sensory tract of the 
trigeminus. At no place has the fasciculus solitarius become 
displaced from its earliest position beneath the external limiting 
membrane to the deeper position which characterizes the adult. 


4. The eye and optic nerve 


While the retina is recognized as essentially a part of the brain 
and the optic nerve as morphologically a tract of the central 
nervous system, the function of the eye as an exteroceptor requires 
that it be treated here with the afferent system. The structure 
of the eye at the four successive stages of development under 
consideration has been illustrated by figures 72 (non-motile), 74 
(early flexure), 76 (coiled-reaction), and 78 (early swimming). 
These are all taken from transverse sections of embryos at the 
level which shows the greatest extent of connection of the lens 
with the skin, or as nearly as could be determined, through 
corresponding points in the eye. 

In the non-motile stage (fig. 72) the pigmented layer is thick 
at the margins but soon tapers off into a relatively thin, simple 
epithelium, which has not yet acquired a high degree of pigmenta- 
tion. It fits down closely upon the retina, excepting in the 
ventral portion where clefts may occur here and there beneath 
it. The central ends of the retinal cells have well defined boun- 
daries, and have a considerable pigment deposited about them. 
Mitosis is going on rapidly throughout the entire peripheral 
(ventricular, with reference to the brain) border next to the pig- 
ment layer. The lens at this stage is simply a discoid thickening 


274 G. E. COGHILL 


of the epithelium, which fits into the optic cup. Its margin con- 
forms closely to that of the retina but there is a narrow space 
between them. In the ventral portion of this space mesenchyme 
cells occur. 

In the early flexure stage (fig. 74) the pigment layer has become 
thinner, particularly at the margins, and its pigmentation has 
increased considerably. The optic cup has greatly deepened. 
The cell boundaries along the central border of the retina have 
become less distinct and cells occur there which suggest the 
nature of neuroblasts. Mitotic figures are still abundant in the 
central portion of the ventricular margin of the retina. The 
lens has acquired the form of a hemisphere, or slightly in advance 
of this, since the line of contact with the skin does not equal the 
diameter of the lens. Mitosis is going on rapidly in the central 
portion of the lens and there is a faint suggestion of a cavity in 
it. The space between the lens and the retina has greatly in- 
creased, owing, apparently, to the increased curvature of the 
retina, and mesenchyme cells have pushed in almost to the center 
of this space (fig. 74, Mes.). Mesenchyma is invading the space, 
also, from dorsad. 

The retina of the coiled-reaction stage (fig. 76) has rounded: 
up into more nearly a spherical form. Cells which appear to 
be neuroblasts are now conspicuous in the middle region of the 
inner margin. The cells along the whole outer border (ventricu- 
lar surface) are still proliferating rapidly by mitosis. While the. 
lens is almost free from the skin, there is still a small area of 
close contact with it. A distinct eavity has appeared in the lens, 
on the outer side of which the epithelium is much the thicker. 
The skin over the lens is still thick, and mitosis continues in the 
lens. 

In embryos of the early swmming stage (fig. 78) the skin over 
the lens has become reduced to a thin epithelium, which, how- 
ever, is still pigmented in a considerable degree. The lens, 
though in contact with the skin, is perfectly detached from it. 
The cavity in it has increased greatly in size, and the inner epi- 
thelium is now much the thicker. The cells here are long and 
columnar and their borders well defined. Mesenchyma extends 


THE NERVOUS SYSTEM OF AMPHIBIA 20 


into the space between the skin and the retina, practically into 
contact with the lens both dorsally and ventrally. The pigment 
layer of the retina has now become highly pigmented, so that 
most of its nuclei are obscured from view. Mitosis is still going 
on in the middle of the outer layer of the retina, but it appears 
to be more abundant towards the margins. In the inner layer 
of the retina there now occur many ganglion cells, particularly 
in the middle region. Good silver impregnations show that the 
axones of these cells form an optic nerve which extends along 
the optic stalk into the brain, where they enter the most caudal 
portion of the chiasma ridge. The form and position of these 
cells are illustrated in figure 71. The axone of this cell passes 
into very close relation with the internal limiting membrane of 
the retina, where it mingles. with other axones. Nothing can be 
made out concerning dendritic processes of these cells. They 
appear still to be unipolar. 

- According to favorable silver impregnations of both A. punc- 
tatum and A. microstomum Cope, these optic fibers form a 
chiasma which is in intimate relation with postoptic commissure 
at this time. The optic fibers, being more deply impregnated 
than the other fibers of the commissure, may be followed clearly 
across the middle line of the brain. They appear to be a con- 
stant feature of the brain of the early swimming embryo, though 
they do not stand out clearly in stained preparation as they do 
in silver impregnations. 

In the same preparations which demonstrate the ganglion cells 
and their fibers clearly nothing of a nervous nature can be seen 
deeper in the retina. There are some round nuclei which sug- 
gest neuroblasts destined to become bipolar cells; but the layer 
of rods and cones seems very embryonic, since, as mentioned 
above, mitosis is still going on in its central region and no well 
defined cellular organization appears there. 


5. The olfactory organ and nerve 


In the non-motile stage the olfactory organ-consists of a greatly 
thickened patch of ectoderm which has externally a very slight 
concavity. It touches the brain only lightly in its dorso-caudal 


276 G. E. COGHILL 


region but not in the region from which the olfactory nerve of 
later stages arises. In the early flexure stage the external con- 
cavity has deepened slightly. The cells of the olfactory epi- 
thelium have become highly columnar and heavily pigmented 
peripherally. The organ has come into rather close contact with 
the brain in the rostral region, but no direct connection can be 
made out between them. In the coiled-reaction stage the external 
concavity has become a shallow pit with well defined margins, 
but the depth is not equal to the diameter of the aperture. There 
is now very intimate contact between the olfactory epithelium 
and the brain at the point where the olfactory nerve is about to 
appear, but no fibers have been observed entering the brain in 
this stage. In the transition from this to the early swimming 
stage the olfactory pit has deepened considerably but has not 
yet dilated perceptibly within the nares. The olfactory nerve 
has now become well established. It arises from near the rostral 
margin of the olfactory epithelium, where the latter is in actual 
contact with the brain. Silver impregnations indicate that the 
fibers arise from neurones situated near the base of the nerve, 
though the impregnations of the cell bodies has been very imper- 
fect and unsatisfactory. In frontal sections (parallel with the 
longitudinal axis of the embryo) the fibers appear running both 
caudad and rostrad within the brain into a zone of highly differ- 
entiated neurones which, presumably, are the mitral cells. There 
is a cell-free area here which obviously represents the glomerular 
zone. 


6. Summary 
A. THE NON-MOTILE STAGE 


The ophthalmic and Gasserian ganglia are both connected with 
the skin at their distal ends (figs. 1, 7 to 12). No fibers can be 
clearly traced from them to the skin. Their root fibers, however, 
reach the brain. Large cells, having the characteristics of the 
Rohon-Beard cells of the spinal cord, situated near the entrance 
of the trigeminal root, send processes out into this nerve (figs. 
20, 21). The general cutaneous ganglion of the vagus (jugular) 


THE NERVOUS SYSTEM OF AMPHIBIA 220 


is very indefinite in outline and has neither peripheral nor root 
fibers arising from it. The lateral line ganglia all have root 
fibers connecting with the brain. Their peripheral ends reach 
to the corresponding primordia of the lateral lines, excepting in 
the case of the ganglion of the hyomandibular system, the rela- 
tion of which to the skin is very indefinite. The auditory vesicle 
(fig. 73) is continuous with the skin and with the auditory gan- 
glion. A barely perceptible auditory root connects with the 
brain. The visceral sensory system is represented by the genicu- 
late ganglion, the glossopharyngeal and vagus. All of these 
connect with thickened patches of ectoderm which are associated 
with the visceral pouches. The geniculate has a very small 
root which reaches the brain. A strand of cells connects the 
glossopharyngeal ganglion with the brain but no fibrillar ele- 
ments appear in this. The vagus ganglion has no root. The 
structure of the eye at this stage is illustrated in figure 72. The 
olfactory epithelium has barely begun to invaginate and does 
not yet touch the brain at the point of future origin of the olfac- 
tory nerve. 


B. THE EARLY FLEXURE STAGE 


The ganglia of the trigeminal nerve have severed their connec- 
tion with the skin and now have peripheral fibers which connect 
obscurely with the skin (fig. 2). The root of the nerve forms a 
very short ascending tract and a descending tract which, aug- 
mented by neurones of the second order, extends to the level of 
the auditory vesicle (fig. 22). The jugular ganglion of the vagus 
has a very indefinite outline. It connects directly with the over- 
lying skin but has no perceptible peripheral or root fibers. The 
ganglia of the acustico-lateral system have acquired greater 
definiteness of outline and compactness of structure. They now 
send fibers to the various primordia, with which they are still 
in contact. Their root fibers enter the brain distinctly. The 
auditory vesicle has become detached from the skin and from 
the auditory ganglion (fig. 75). The visceral sensory ganglia 
are still connected with the epibranchial placodes and no pe- 
ripheral fibers arise from them. The root fibers of the geniculate 


THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL. 26, No. 3 


278 G. Bi COGHILE 


and glossopharyngeal ganglia enter the brain, but the vagus 
ganglion has no root. ‘The optie cup has deepened greatly and 
the lens projects deeply into its cavity. The invagination of the 
olfactory epithelium is more marked but there is not yet an 
olfaetory nerve. 


C. THE COILED-REACTION STAGE 


The distribution of the trigeminal fibers to the skin is much 
more obvious than in the earlier stage, the mesial terminal branch 
of the r. ophthalmicus profundus being well defined. A descend- 
ing trigeminal tract of a composite nature extends to the spinal 
levels (figs. 3, 5). The giant ganglion cell component of the 
trigeminus is now particularly clear (figs. 34 to 37). The outline 
of the jugular ganglion is still indefinite and no perceptible fibers 
leave the ganglion. The lateral line primordia have become 
much extended and a mandibular line makes its appearance. 
Lateral line nerves follow along the primordia closely attached 
to the skin, although no fibers have been traced in this stage to 
the mandibular line. All of the lateral line roots form ascending 
and descending tracts in the brain, and the auditory root forms a 
short descending tract (fig. 5). The auditory vesicle of this 
stage is illustrated in figure 77. The geniculate ganglion is still 
united with its epibranchial placode, but from it emerge several 
fibers which project ventrad behind the spiracular pouch as an 
incipient r. alveolaris. ‘The root fibers from this ganglion form 
the fasciculus solitarius which runs caudad immediately beneath 
the external limiting membrane to the level of the glossopha- 
ryngeal root which enters it. Peripherally the glossopharyngeal 
ganglion connects with its epibranchial placode. The vagus 
ganglion now has a small root which joins the fasciculus solitarius. 
Distally this ganglion connects with the epibranchial placodes, 
but gives rise to no peripheral fibers. The condition of the eye 
is shown in figure 76. The olfactory epithelium is now in close 
contact with the brain, but olfactory root fibers do not certainly 
enter the brain in this stage. 


THE NERVOUS SYSTEM OF AMPHIBIA 279 


D. THE EARLY SWIMMING STAGE 


Of the general cutaneous system the lateral and mesial ter- 
minal branches of the r. ophthalmicus profundus, the r. maxil- 
laris, r. mandibularis with a nerve to the balancer, and a nerve 
from the ophthalmic ganglion directly to the overlying skin are 
well defined (fig. 4). The r. auricularis vagi is perceptible but 
its distribution is uncertain. The root fibers of the trigeminus 
now extend in the descending tract to about the level of the 
root of the vagus nerve, and this tract is joined by the root of 
the vagus ganglion (fig. 6). The giant ganglion cell component 
of the trigeminal nerve is still present, but its exact distribution 
through the nerve can not be determined. The hyomandibular 
lateral line primordium is now continuous with the other pre- 
auditory primordia and is innervated by the r. hyomandibularis 
VII. The other lateral line primordia are now innervated prac- 
tically throughout their entire extent, the lateral line of the 
trunk extending into the level of the seventeenth myotome. 
The extent of the ascending and descending tracts of the lateral 
line roots is shown in figure 6, where the extent of the descending 
auditory fibers is also figured. The visceral sensory component 
is now represented by a well developed r. alveolaris VII, r. lin- 
gualis IX, and a less developed first r. branchialis vagi. There 
are no prebranchial or pharyngeal branches of the postauditory 
branchial nerves. The vagus root now connects with the brain 
and contributes fibers to the fasciculus solitarius. In the retina 
mitosis is still going on rapidly in the middle of the external 
layer, while in the internal layer ganglion cells are already 
developed which send their axones along the optic stalk into the 
brain. They decussate in the most rostral portion of the post- 
optic commissure. Only a suggestion appears of neuroblasts in 
the region of the bipolar cell layer. The olfactory nerve is well 
established and its fibers distribute themselves in a very definite 
glomerular zone of the olfactory lobe. 


280 GE. COGHILE 


Il. THE PHYSIOLOGICAL PART 
1. Response to tactile stimulation 


Early in my study of the responses of these embryos to tactile 
stimulation it became apparent to.-me that the skin of the pre- 
auditory region was less sensitive than was that of the post- 
auditory region. For conclusive evidence upon this point, how- 
ever, I am indebted to one of my graduate students, Miss Ruth 
Orcutt, who in March, 1911 carried out a series of experiments. 
two of which I introduce here as typical of her results. 

Ten specimens were selected which were approaching very 
near to the typical coiled-reaction stage, their movements being 
not quite extensive enough to be typical. These were touched 
with a hair one after the other in regular order in the following 
manner: At each trial an embryo was at first touched in the 
region of the eye. If this stimulus produced a response, the next 
embryo of the series was touched; if it did not produce response, 
the same embryo was, after an interval, touched over the post- 
auditory region in the territory of innervation by the vagus 
nerve. If response occurred, the next embryo was tested; if not, 
the same embryo was then touched over the myotomes of the 
cervical region, the territory of the Rohon-Beard cells of the 
spinal cord. The ten embryos were worked over in regular 
order in this manner till one hundred forty-three responses oc- 
curred. Of these five were in response to touch on the preaudi- 
tory region; 66, to stimulation of the territory of the vagus 
nerve; and 72, to stimulation of spinal territory strictly. 

Another group of embryos of a distinctly more advanced stage, 
being in the advanced coiled-reaction stage, were tested in the 
same manner. ‘These gave 94 responses, of which 12 were in 
response to stimulation of the trigeminal region; 41, to stimu- 
lation of the territory of the vagus; 41, to stimulation of the 
spinal region. 

These and other experiments upon this question give unequiv- 
ocal evidence that stimuli from the spinal field reach the motor 
centers with much greater facility and certainty than do stimuli 


THE NERVOUS SYSTEM OF AMPHIBIA 281 


from the territory of the vagus nerve; and that, as compared with 
stimuli from these areas, stimuli very rarely reach the motor 
centers from the trigeminal field in embryos of the typical coiled 
reaction stage. At the time these experiments upon this sub- 
ject were made we had no exact information upon the nature of 
the innervation of the regions under consideration. ‘The ana- 
tomical part of this paper, however, shows that the trigeminal 
nerve, throughout the whole period of development up to the 
early swimming stage, runs far ahead of the general cutaneous 
component of the vagus in the development of both peripheral 
and root fibers. The relative accessibility of motor centers, 
therefore, to stimuli from trigeminal and vagus territories can 
not depend upon the relative development of these nerves. ‘The 
certainty of responses, when the stimuli are applied to the spinal 
region, obviously depends upon the perfection of the giant gan- 
glion cell system of afferent neurones and their direct access to 
motor centers (Paper I). The neurones of this system occur as 
far rostrad as the second myotome, which is essentially within the 
limits of the medulla oblongata. They may, indeed, occur here 
and there farther cephalad as shown in this paper. Their pe- 
ripheral fibers, therefore, probably invade the region which was 
regarded in the above mentioned experiments as belonging to 
the vagus, but in a less efficient manner than in the trunk, so 
that response is not as certain from stimulation here as it is from 
stimulation farther caudad. The response to stimulation of the 
whole postauditory territory and trunk is in all probability 
effected through the giant ganglion cells system in the coiled- 
reaction stage. And since we now recognize cells of this type 
in the trigeminal nerve, such responses as occur to stimulation 
of this region in the earlier periods may take place through this 
system also. Granting, however, that all the giant ganglion 
cells in the trigeminus are distributed to the skin rather than to 
the muscle primordia, the innervation of the preauditory region 
by this system must be very sparse as compared with that of 
the postauditory and trunk regions, for there are only a few 
giant ganglion cells in the vicinity of the trigeminal root. Such 
sparseness of innervation, upon the hypothesis that the gan- 


282 G. E. COGHILL 


glionie portion of the trigeminus is not functionally connected 
with the motor centers, might explain the low degree of respon- 
siveness to stimuli in this region. On the other hand, if the 
responses are effected through the ganglionic system of the tri- 
geminus their infrequency in the coiled-reaction stage must depend 
upon the limitations of the zone of influence of the root fibers 
within the brain, for the peripheral distribution of the nerve at 
this time is very well developed. 

As a factor in this problem the descending ann tract 
naturally suggests itself. This tract has been described in the 
first part of the paper as being of a composite nature—made up 
of processes of neurones of the first and second, and possibly 
higher orders, in varying proportions according to the region. 
The root fibers of the trigeminal ganglia run some distance caudad 
in the coiled-reaction stage—probably almost to the level of the 
‘auditory vesicle. Beyond this the tract, now extending into the 
spinal region, consists of processes from central neurones along 
the course. If, then, stimuli reach the motor centers in the lower 
regions of the medulla oblongata through the descending tri- 
geminal tract they would have to traverse a series of relatively 
short neurones with intercalated synapses after leaving the tri- 
geminal conductors. Such a conduction path would certainly 
introduce relatively great resistance if it did not block the stimul 
altogether. This is, in all probability, the really significant fac- 
tor in determining the behavior of embryos of the coiled-reaction 
stage to tactile stimulation of the preauditory region. 

In the early swimming stage the condition is quite different. 
The trigeminal fibers now reach almost or quite to the level of 
the vagus root, that is, into the immediate vicinity of the centers 
that dominate the whole motor mechanism. Accordingly, re- 
sponse to stimulation of the trigeminal region has become prompt 
and regular. The data at hand, therefore, afford very good 
evidence that such a descending trigeminal tract as occurs in 
the early flexure stage (fig. 22) is at best a very inefficient con- 
ductor; and that before response can become prompt and con- 
stant the root fibers of the nerve must reach directly into the 
approximate levels of the motor centers. We have in this con- 


THE NERVOUS SYSTEM OF AMPHIBIA 283 


dition, then, the significance of the rapid growth of the descend- 
ing trigeminal tract between the coiled-reaction and early swim- 
ming stages, for there is very close correlation between the 
development of the power of locomotion and the introduction 
of the most anterior sensory field of the organism into direct 
connection with the motor centers so that it becomes effective 
in determining the direction of locomotion. 

The significance of this correlation in the growth of the motor 
and sensory systems finds a striking demonstration in the reac- 
tions of embryos of advanced swimming stage which have been 
transected in the upper portion of the trunk (Paper I, p. 200). 
This operation cuts out both the exteroceptive and proprioceptive 
stimuli from the greater portion of the trunk and leaves the 
motor centers more exclusively to the influence of the cutaneous 
field of the head. In head pieces of this kind, which have just 
enough of the trunk musculature attached to give unequivocal 
responses, the movements are almost universally away from the 
side touched. It is obvious, therefore, that this avoiding form 
of response to stimuli about the anterior end of the embryo 
‘ is a basic thing in the orientation of the swimming animal to its 
environment; and, judged by the anatomical results of this 
paper, its perfection and efficiency depend upon the growth of 
the trigeminal root fibers into proximity with the motor centers. 

In the above experiments upon the relative irritability of 
different areas it was observed that the regions under consider- 
ation differed materially with reference to the nature of the 
response they evoke. In the first set, three of the five responses 
to stimulation of the trigeminal area were away from the side 
touched, whereas 62 of the 66 responses to stimulation of the 
vagus field were away from the side touched, as were all of the 
72 responses to stimulation of the spinal field. In the second 
experiment mentioned the embryos were considerably older and 
all of the 94 responses were away from the side touched excepting 
two of the twelve responses to stimulation of the trigeminal area. 
These data are in accord with my general experience with these 
embryos, namely, that with increased responsiveness there is 
increased purity of type in response. In other words, embryos 


284 G. E. COGHILL 


that are not readily excited to motor response are much more 
likely to move towards the side touched than are embryos that 
are easily excited to movement. 

The conclusion of the last paragraph raises a question which 
has been mentioned in an earlier paper (’09, p. 252) as to whether 
movement away from the side touched is a fundamentally pri- 
mary thing in behavior or a secondary thing, derived by a process 
of selection out of a diffuse or irregular type of response. From 
the conclusion reached concerning the structural basis for prompt 
and certain response to stimulation of the trigeminal field, it 
might be inferred that this problem also hangs directly and 
solely upon the degree of development of the peripheral neurones 
in their relation to the motor centers. In other words, if a 
stimulus must pass through such a descending trigeminal tract 
as is shown in figure 22, representing the early flexure stage, 
there is a larger element of chance or accident in the direction of 
the response than there would be if the stimulus traveled into 
the immediate vicinity of the motor nuclei through the peripheral 
neurones. ‘This would be an attractive hypothesis, but, to my 
mind, it is only of secondary consideration. When, for instance, ° 
there is a high degree of responsiveness (a low threshold of stimu- 
lation for the entire reflex arc concerned) a stimulus such as is 
employed experimentally would act in a much more localized or 
restricted area than it would when there is a low degree of 
responsiveness. Under the latter condition the stimulus would 
probably be transmitted to relatively distant receptors, possibly 
even to the opposite side of the embryo and, finding here access 
to an are of lower threshold, it would excite a response that 
apparently violates the regular law of behavior. ‘This expla- 
nation gains credence from Hooker’s experiments (’11) which 
show that amphibian embryos of the ages under consideration 
are very sensitive to stretching of the skin by pressure upon 
relatively distant points. In fact his experiments show con- 
clusively that in order that a tactile stimulus act locally in any 
appreciable degree in these embryos there must be a very low 
threshold of stimulation for the reflex are as a whole. Undue 
resistance at any point in the are would certainly allow the 


THE NERVOUS SYSTEM OF AMPHIBIA 285 


stimulus to act on more distant receptors of ares with, possibly, 
a lower resistance. The cause of irregularity in the direction of 
response when responsiveness is low is, therefore, in all proba- 
bility to be found in the receptor system and adequate stimu- 
lation of an are as a whole rather than in an hypothetical diffuse 
form of conduction through the central nervous system. 


2. Response to chemical stimulation 


The question of the existence of a true irritability to sub- 
stances in solution as opposed to their destructive action on the 
skin has been discussed in Paper I. By spraying small jets of 
acid upon restricted regions of the embryo and by immersing 
head pieces of transected specimens in acid it has been shown 
that the skin of the head is like that of the trunk in its sensitive- 
ness to the action of acid. It appears, therefore, that the defini- 
tive, ganglionic nerves of the ages under consideration do not 
differ from the giant ganglion cell system in respect to chemical 
stimulation. 


8. Response to light 


During the season of 1910 I performed a series of experiments 
upon the reaction of larvae and embryos of A. opacum and A. 
punctatum to light. Sunlight, thrown from a mirror through a 
water jacket and condenser, was employed in such a way as to 
bring to bear very high illumination upon sharply restricted 
regions of-the animals. As a check on the adequacy of the 
stimulus a larva of A. opacum was used which was about ten 
days beyond the age when swimming begins. When intense 
illumination was brought to bear over the head and eyes of this 
larva responses were fairly prompt, as the following figures show, 
the reaction times being indicated in seconds, with the interval 
between one response and the next stimulus indicated in paren- 
tmeses: 5, (15); 10, (15); 8, (20); 8, (80);8, (20); 3, (80); 8, (80); 
7, (80); 8, (30); 8. The responses were always quick, total body 
movements and swimming away from the stimulus, slightly to one 
side rather than straight forward from the position when stimu- 


286 G. E. COGHILL 


lated. Immediately following this series of responses the larva 
lay with its trunk and tail only in the highly illuminated field 
for five minutes without movement, excepting one small move- 
ment of the head. In a fainter light, but with illumination 
sufficiently strong to cast a distinct shadow upon the dark sur- 
face on which the larva rested, it lay for ten minutes longer 
motionless. The responses recorded above, therefore, must have 
been to stimulation of the eyes and not to an action of the light 
upon the skin; and the irritability of the skin to light in this 
relatively late stage must be negligible. 

The methods of the above experiment were applied to five 
embryos of A. punctatum which had almost but not quite reached 
the swimming stage, but the results gave no evidence of irrita- 
bility to light. Four specimens which had reached the early 
swimming stage, however, gave the following results (the time of 
application of the stimulus being indicated in seconds, the plus 
sign indicating a reaction after the time indicated, the minus 
sign indicating no reaction): 


Specimen a. 20-+, 20+, 10+, 60 —, 30+, 30+, 10+: 
Specimen b. 60 —, 20 +, 120 +, 10 +, 60'-+-, 10 +, 60 —. 
Specimen c. 60 —, 70+, 60 —, 60 —, 20+, 15.—+. 

Specimen d. 60'—,45--, 50, (60—, 60 —, 60) —, 60 —- 


The specimens were stimulated one after the other so that 
there was a relatively long interval between successive stimuli 
applied to the same specimen. The illumination during this 
period may not have been absolutely constant but it was at all 
times very brilliant. As embryos of this age only rarely move 
without some form of external stimulation, it appears that the 
responses observed in this experiment were due to the stimula- 
tion of the retina by light, but the threshold of stimulation is 
exceedingly high. The shortest reaction time was ten seconds, 
as compared with three seconds in the older larva; whereas one 
specimen lay in one trial for two minutes before responding and 
in two other trials gave the shortest reaction time. 

Another set of specimens of the same age were tested in the 
same manner with the following results: 


THE NERVOUS SYSTEM OF AMPHIBIA 287 


Specimen e. 60 —,60—, 60 —, 60 —, 60 —, 60 —, 60 —, 60 —, 60 —. 

Specimen f. 60—,60—, 35+, 25+, 60 —, 55+, 60 —, 60 —, 254, 30+. 
Specimen g. 60 —,60—, 30+, 60 —, 60 —, 45 —, 50+, 604, 604+, 35-4. 
Specimen h. 55+, 60—, 60 —, 60 —, 60+, 60 —, 60 —, 60 —, 53+, 60-. 


In brief, specimen e gives no evidence of irritability to light, 
f and g show noteworthy evidence of it, while the other three 
show slight evidence of it. It may be said, therefore, that at the 
time swimming begins the retina is just beginning to be irritable 
to high illumination. 

‘The anatomical explanation for these results will not be fully 
before us till the central nervous system is better understood, 
but it is important to note here that it is just at this stage that 
the optic nerve makes its connection with the brain and that 
the retina is in a very embryonic condition in its outer layers. 
In the light of these reactions interest attaches to this order of 
development in the retina. The ganglion cells, it has been noted, 
are well differentiated and their fibers decussate in the incipient 
chiasma. ‘These neurones have as high a degree of differentia- 
tion as have other neurones that are known to be in the functional 
condition. The inference, therefore, is that the high and variable 
threshold of the retina at this stage is due to the embryonic con- 
dition of the more peripheral elements in the reflex circuit. On 
the other hand, it is interesting to know that with the layer of 
rods and cones and the bipolar cell layer in such an embryonic 
condition there could be any optic reflexes stimulated. One 
almost questions whether the ganglion cells themselves may not 
possess at this time a certain degree of irritability to light. 
In the early swimming stage, as described in the anatomical part 
of the paper, mitosis is till going on in the central zone of the layer 
of rods and cones and only suggestions of neuroblasts can be 
detected in the region of the layer of bipolar cells. It seems 
incredible that specialized photie receptors and conduction paths 
can be already established in such embryonic structures. Upon 
the basis of studies which are now in progress on the central paths 
of the optic reflexes it will be necessary to return to more exhaus- 
tive cytological study of the retina and to an amplification of 


288 G. E. COGHILL 


the experimental evidence of the development of function in the 
organ. 


4. Response to olfactory stimulation 


In the anatomical part of the paper it was noted that the 
olfactory nerve is well established in the early swimming stage. 
Preliminary experiments were made in 1912 to discover whether 
olfactory stimuli in any way stimulate movement. Pellets of 
various substances were enclosed in gauze and placed in the 
center of flat dishes about 9 by 13 em. In these dishes embryos 
somewhat older than the early swimming stage were distributed 
in cistern water. Pellets of pure gauze were used as checks. 
In other pellets were placed, egg masses, fresh algae and débris 
from aquaria which emitted a strong odor. ‘The embryos in all 
the dishes were frequently agitated to discover whether the 
olfactory organ had any influence in directing movements stimu- 
lated from other causes. Over thirty embryos in the different 
test dishes and the same number in the check dish were kept 
under observation for three hours without the manifestation of 
any tendency on the part of the specimens to collect in the 
vicinity of the odiferous substance. A number of experiments 
of this kind gave no further suggestion of olfactory reflexes. 
Experiments of this kind, however, should be carried further to 
determine whether other substances may not be found which 
would stimulate olfactory reactions, either directly or by rein- 
forcing or inhibiting other responses. It has been noted how- 
ever, that, although olfactory neurones have grown into the 
brain in considerable numbers, the olfactory epithelium is very 
poorly differentiated in the ages under consideration. 


5. The auditory and lateral line organs 


The relative development of the lateral line system of organs 
and nerves during this period arouses inquiry concerning their 
function. In the postauditory region this system develops dis- 
tinctly in advance of the general cutaneous component in both 
its central and its peripheral relations. Even in the preauditory 


THE NERVOUS SYSTEM OF AMPHIBIA 289 


region, where the general cutaneous system is more precocious, 
the connection of the lateral line nerves with the skin becomes 
readily recognizable earlier than in the case of the general cutane- 
ous nerves, and throughout the period under consideration the 
lateral line nerves have the more intimate and the more exten- 
sive contact with the skin. Judged upon the anatomical basis, 
therefore, the lateral line system would certainly be held as play- 
ing a larger réle in the life of the organism at this time than do 
the general cuitaneous nerves. I have been unable, however, to 
discover what this function may be and have further experiments 
planned. 

The auditory organ, being a simple dilated vesicle, gives no 
evidence of being a mechanism that would be adapted as a 
receptor to the function of equilibration, so that the auditory 
nerve, which has established quite extensive relations within the 
brain and intimate contact with the vesicular epithelium in the 
early swimming stage probably does not enter into the reactions 
of the organism till a later period. As is well known, the embryos 
of this stage lie on the side while at rest and hold the upright 
position only during locomotion. The embryos which I select 
as representing the early swimming stage move upon the sub- 
stratum, and their upright position during locomotion is prob- 
ably the result of the lateral thrust of the head, which would 
tend to prevent them from settling down on either side till 
motion ceases. 


6. Summary 


a. The preauditory portion of the ganglionic general cutaneous 
system develops far in advance of the postauditory portion. 
The former becomes a factor in quick and certain response only 
about the time the embryo begins to swim, although response 
to preauditory stimulation occurs irregularly and infrequently 
for a considerable time before this. 

b. The entrance of the receptor field of the trigeminal nerve 
into function as an important factor in the behavior is closely 
correlated with the extension of the root fibers of the trigeminal 


290 G. E. COGHILL 


nerve into the immediate vicinity of the motor centers in the 
lower portion of the medulla oblongata. 

e. Such irregularity as appears early in development in the 
direction of movement relative to the side touched depends upon 
the condition that the point of adequate stimulus is not the point 
actually touched, by reason of the variability in the threshold 
of stimulation of reflex ares as a whole, and not upon diffuse 
conduction through the nervous system. Reflex ares are from the 
first definite and fixed during the period under consideration. 

d. The afferent system of the head is like that of the trunk 
with reference to chemical stimulation. 

e. A slight responsiveness to high illumination of the retina 
occurs about the time the animal begins to swim, and this is in 
close correlation with the development of the first fibers of the 
optic nerve into the brain. Owing to the exceedingly embryonic 
condition of the retina at this time this topic demands further 
investigation. 

f. Although the olfactory nerve is well established at the end 
of the period under investigation there are no perceptible reac- 
tions to olfactory stimulation. 

eg. There is no available evidence that the auditory organ or 
lateral line nerves have any part in reactions, although the latter, 
judged upon the basis of its anatomical relations, would be re- 
garded as more efficient than the ganglionic general cutaneous 
system. : 

h. The physiology of the eye, olfactory organ, auditory organ 
and lateral line organs require further investigation in connec- 
tion with anatomical studies of the brain, which are now in 
progress. 


III. DISCUSSION 


It is the purpose to deal here with the work of other investi- 
gators which seems to have direct bearing upon the results of 
this paper. 


THE NERVOUS SYSTEM OF AMPHIBIA | 291 


1. Lateral line primordia, placodes and ectodermal thickenings 


In the use of the terms ‘lateral line primordia,’ ‘placodes’ and 
‘ectodermal thickenings,’ I am following Landacre and Conger 
in their work upon Ameiurus and Lepidosteus (Landacre, ’10, 
12; Landacre and Conger, 713). I wish, however, at the outset 
of this discussion, to emphasize the fact that my series of Ambly- 
stoma embryos is not adapted to a critical study of the histo- 
genesis of these structures, for my specimens have been selected 
at relatively long intervals according to physiological stages and 
with reference to a problem that is not primarily one of histo- 
genesis. It seems obvious, nevertheless, that there is a striking 
resemblance between Amblystoma and Lepidosteus (Landacre 
and Conger, ’13) in the early processes of differentiation of sense 
organs and nerves. 

While I have not undertaken to determine to what extent epi- 
branchial placodes contribute to the formation of the various 
cranial ganglia, it seems certain that the visceral ganglia of the 
facial, glossopharyngeal and vagus nerves all receive masses of 
cells from this source, particularly during the middle and latter 
part of the period covered by my studies. This tardy differen- 
tiation of the visceral as compared with the somatic sensory 
ganglia is correlated with a late entrance of the visceral sensory 
field into the behavior, and is particularly noteworthy in connec- 
tion with Landacre’s observation that the relatively late differ- 
entiation of the epibranchial placodes in Lepidosteus as com- 
pared with Ameiurus is correlated with a late development of 
the taste buds in the former as compared with the latter, accord- 
ing to particular growth periods. 

No lateral line placodes have been recognized in Amblystoma 
of my series, unless the ectodermal thickening which joins the 
auditory vesicle to the preauditory lateral line primordium be 
regarded as a vestige of a dorso-lateral placode. In section this 
thickening of the skin has the appearance of that which is formed 
in the early flexure stage over the ophthalmic ganglion where 
the latter has just become detached from the ectoderm. ‘The 
significance of this attachment of the ophthalmic ganglion to 


292 G. E. COGHILL 

the skin does not seem to be. understood. Landacre (’12) ob- 
served the same relations in Lepidosteus, but, apparently, did 
not satisfy himself concerning the meaning of it. 

The ectodermal thickenings which Landacre and Conger vie 
differentiated from the epibranchial placodes and the lateral 
line primordia certainly will be found also in Amblystoma, al- 
though my studies have not extended to the details of their 
relations. This subject deserves exhaustive study from mate- 
rial selected for the purpose. My object in touching upon this 
topic has been merely to bring the salient features in the differ- 
entiation of the nervous system which these structures represent 
into correlation with physiological processes in the organism as 
a whole. Limited as this study has been, however, I have been 
increasingly impressed with the results of Landacre’s studies on 
this subject as fundamental to a broadly biological view of the — 
development of the nervous system of the vertebrate head. 
These ectodermal thickenings and proliferations impress me not 
simply as vestiges of an ancestral metamerism or branchiomerism, 
but as centers of physiological processes which may be essential 
factors in the regulation of the growth of the nervous system. 
There is a field here for experimental studies, it seems to me, 
which might prove as illuminating as the well known observations 
upon the correlative development of the lens and the retina, 
and it is to be hoped that the purely embryological studies of 
these structures may be extended to other types of vertebrates. 


2. The function of the auditory vesicle 


Streeter (06) has studied the development of the ear as corre- 
lated with the development of the function of equilibration in 
the tadpole of the frog. He found that shortly before tadpoles 
acquire the ability to maintain an upright position ‘the labyrinth 
consists of a closed epithelial sac incompletely subdivided into 
compartments and possessing differentiated nerve endings which 
are connected with the brain by the acoustic nerve and ganglion.” 
The semicircular canals are ‘‘in the process of development, but 
are not completely pocketed off until after equilibration is already 


THE NERVOUS SYSTEM OF AMPHIBIA 293 
established. Consequently the semicircular canals as such are 
not an essential factor in equilibration.”’ He observed, further, 
that tadpoles from which he removed one auditory vesicle in the 
3 mm. stage began, when 9 to 9.5 mm. long, to differ from normal 
tadpoles in the manner of swimming—the operated specimens 
having a tendency ‘“‘to swim with the operated side under, and 
in the rolling movements around the long axis of the body it is 
from the operated side under to the opposite.”’ Streeter con- 
cludes that it is at about this time that ‘‘functional union is 
normally established between the ear vesicle and the spinal 
cord reflex centers, upon which the individual is dependent for 
maintaining its position in free water.’”’ Estimated by Streeter’s 
notes, this is about two days after the animals begin to swim, 
and, therefore, at a stage much beyond that which I eall the early 
swimming stage in Amblystoma. This difference in age is ex- 
pressed in the structure of the ear as well as in the efficiency of 
the motor system in locomotion, for the ear of the Amblystoma 
early swimming stage is a perfectly simple sac, except for the 
endolymphatic appendage—a much simpler organ than the ear 
of the frog tadpole when it acquires the power to swim upright 
in free water. Streeter notes that the auditory nerve is devel- 
oped at this time, but does not describe its central relations in 
detail. While it would be interesting to know these central 
relations from the point of view of functional mechanism as a 
whole, Streeter’s results seem to be conclusive on the question of 
the relation of the auditory organ as such to the function of 
equilibration, and they do not lead us to expect any functional 
union of the ear with the motor centers in embryos of Ambly- 
stoma of the early swimming stage. 


3. The sensory nerve roots in relation to later development 


The medulla oblongata of Amblystoma of lengths from 17 to 
38 mm. has been recently described by Herrick (’14). Such 
larvae are practically adult so far as reactions are concerned and 
in the structure of the medulla oblongata Herrick considers that 
all the fundamental relations of the adult have been acquired. 


THB JOURNAL OF COMPARATIVE NEUROLOGY, VOL. 26, No. 3 


294. G. EB. COGHILL 


A comparison of his descriptions with the anatomical results of 
this paper will show that the general plan of the sensory centers 
of the medulla oblongata is laid down by the growth inward of 
the root fibers of the peripheral nerves. The trigeminal and 
lateral line fibers have formed extensive ascending and descend- 
ing tracts, and the auditory and geniculate ganglia have formed 
descending tracts of considerable length by the time the animal 
begins to swim. These tracts are all closely grouped and all 
lie immediately beneath the external limiting membranes, and 
constitute the substantia alba of the dorsal portion of the medulla 
oblongata. They lie embedded in the marginal zone of the sub- 
stantia grisea, in which are neurones directed dorso-ventrad 
across the whole series of tracts of the various sensory modes. 
In the neurones of this group there can scarcely be any specificity 
of function with reference to the different sensory nerves (com- 
pare Herrick and Coghill 715). To this elementary plan the 
adult has added the longitudinal association tracts a and 6 of 
Kingsbury and Herrick, and an additional lateral line tract, 
while the fasciculus solitarius has become detached from the 
periphery and with its displacement inward the lateral line centers 
have developed greatly ventrad as well as dorsad. 

Without knowledge of intermediate stages between the early 
swimming stage and the younger of Herrick’s descriptions it is 
impossible to say just how the three root bundles of the lateral 
line VII of Herrick are derived from the two bundles of the 
younger stage. Since, however, the root bundles of the younger 
stage have the same relation to each other and to the root bundles 
of the postauditory system as do the two more ventral bundles 
of Herrick, and since the most dorsal bundle of Herrick has, in 
younger stages, smaller fibers than the other bundles, it is reason- 
able to infer that the most dorsal bundle of Herrick is a distinct 
and newly acquired tract, and not a detached portion of one or 
the other, or of both of the tracts of the younger stage. 

In the relation of the several root bundles to particular ganglia 
Herrick finds that, in the case of the facial division, each of the 
tracts receives fibers from both the preauditory ganglia. This 
is a very difficult relation to determine in the young embryos, 


THE NERVOUS SYSTEM OF AMPHIBIA 295 


for they must be studied under very high magnifications, and the 
roots come into very close relations with each other as they 
approach the brain. It seems unquestionable, however, that the 
more dorsal tract is derived chiefly from the dorsal ganglion (a) 
and that the ventral tract comes chiefly from the ventral gan- 
glion (b). In the case of the postauditory group, the dorsal 
tract is derived chiefly from the vagus and the ventral, chiefly 
from the glossopharyngeal ganglion. It is impossible to say 
that there is no interchange of fibers between the roots of either 
set, but if such anastomosis does occur it must involve only a 
very few fibers. ; 

If these lateral line tracts were each composed of fibers from 
one ganglion exclusively there would be grounds for assuming 
that they effect a certain degree of localization in the brain with 
reference to the particular areas with which they are severally 
connected. Herrick’s observation of the mingling of the fibers 
of the different roots, however, makes it necessary to seek some 
other explanation for this early differentiation of the lateral line 
system within the brain. ‘There is nothing in the sense organs 
to suggest that this has to do with different sensory modes. 
Its significance, therefore, must in some way have to do with the 

‘directness with which stimuli may pass to the motor centers. 
But that this mechanism as it is found in early swimming em- 
bryos could serve such a function seems incredible, particularly 
in view of the nature and arrangement of the neurones of the 
second order in the circuit. It is hoped that the study which 
is in progress on the association paths of the brain may offer 
something towards the solution of this problem. 

In view of Herrick’s observations of the bifurcation of the root 
fibers, it is of interest to note, further, that in the case of the 
trigeminal and lateral line nerves this takes place very early in 
development. In the case of the visceral sensory and auditory 
roots, however, there is little or no evidence of bifurcation of the 
fibers in the early swimming stage. 


296 G. E. COGHILL 


4. The development of function in the neurone and in the reflex are 


It has already been pointed out (Paper I) that the giant gan- 
glion cells of the spinal cord on the one hand and the motor 
neurones on the other have their connections made with their 
respective end-organs for a considerable time before they become 
integral parts of a functional reflex arc, and that, for this reason, 
these neurones can not be used as a standard by which to judge 
the structural changes in the neurone that mark the beginning 
of nervous function. Now we find that the cranial afferent 
nerves also are in the earlier stages developed in a degree that 
seems out of proportion to the part that they can be demon- 
strated to have in reflexes. We have noted, for example, that, 
in the coiled reaction stage, the trigeminal nerve has a very gen- 
eral distribution to the skin and that it has root connection with 
the brain from a much earlier period, whereas it becomes an 
important factor in behavior only about the time the embryo 
begins to swim. The neurones of this nerve must for a con- 
siderable time receive stimuli and conduct impulses that are 
never discharged through any reflex are to the effectors. The 
root fibers of the visceral sensory system also are well developed 
and the fasciculus solitarius is extensively developed for many 
days before the neurones of this system are subjected to physio- 
logical stimulation by the development of the mouth. None of 
these neurones, therefore, can give us a clue as to when they 
take on nervous functions. 

Our observations upon the development of the optic nerve in 
relation to the function of the eye show that the same uncertainty 
is attached to the neurones of the second or higher order in the 
are. In this case the retinal ganglion cells are highly differen- 
tiated and can, under very strong stimulation of the retina, 
become an integral part of a reflex are at a time when the more 
peripheral sensory elements of the are are in an exceedingly 
embryonic condition. The high threshold of the optic reflex 
must then be due to the undeveloped condition of the retina 
whereas the ganglion cells, optic nerve fibers, and the central 
neurones reaching caudad from the diencephalon to the lower 


THE NERVOUS SYSTEM OF AMPHIBIA 297 


portion of the medulla oblongata must be capable of conducting 
stimuli for a considerable period before they can be excited by 
normal stimulation of the retina. The development of the olfac- 
tory nerve and centers suggests that the same principle applies 
to much of the central nervous system in early development. 
We have, then, within our knowledge of the development of 
reflexes and reflex ares no rational basis for cytological studies of 
the differentiation of the neurone as correlated with the develop- 
ment of nervous function. It is hoped that the studies which 
are now in progress upon the central conduction paths will furnish 
such a basis for cytological investigations; and that the appli- 
cation of cytological methods to nerve cells of known physio- 
logical capacity may reveal the essential structural basis of con- 
duction in the neurone. 


5. The cranial nerves of Amphibia 


Some points of interest concerning the morphology and devel- 
opment of the cranial nerves of Amphibia deserve mention here. 

In my paper on the cranial nerves of Amblystoma (’02, p. 215, 
and figure 1) the lateralis ganglion which corresponds to VIIa 
of this paper is described as closely fused with the ganglion of the 
ophthalmicus profundus, and the latter as very intimately related 
with the Gasserian ganglion. In the younger embryos of this 
study these ganglia are widely separated, but in the latter part 
of the period under investigation these ganglia, particularly the 
Gasserian and profundus, approach each other rapidly, and 
finally the Gasserian and profundus ganglia establish a wide 
contact, the lateralis ganglion VII a still standing apart from 
them. A similar change, though of less degree, occurs in these 
ganglia in Rana pipiens, according to Landacre and McClellan 
(12), who describe the ganglion of the ophthalmicus profundus 
in the 8 mm. embryo as standing out ‘‘rather distinctly, indi- 
cating its definite character,’ and as being much more isolated 
in younger stages. In the early development, therefore, of both 
Anura and Urodela there is a process of consolidation going on 
between the various ganglia of the V + VII complex, a process 


298 G. BE. COGHILL 


which goes farther in Anura than in Urodela, that is to say, 
till all the ganglia become fused into a common mass, whereas 
they become two masses in the Urodela. 

This migration and fusion of ganglia involves not only a 
shifting of position relative to each other but also a general 
movement of the ganglia towards the brain along the path of 
their respective roots. There is also a shifting of the otocyst 
rostrad with reference to the point of entrance of the facial and 
auditory roots into the brain. Such a movement of the vesicle 
together with its expansion in growth, if carried far enough before 
chondrification of the auditory capsule sets in, might expiain the 
fusion of all these preauditory ganglia into a common mass in 
Urodela; but this could not explain the migration caudad of the 
ganglion of the opthalmicus profundus, for instance, as illus- 
trated in figures 1 to 4. There is apparent no purely mechanical 
cause for such movement on the part of these ganglia. 

A very similar migration and a consolidation of ganglia occur 
in the postauditory ganglionic complex as well. In adult Ambly- 
stoma there is a single postauditory ganglionic mass which can 
be resolved into its component gangha only with difficulty (Cog- 
hill, ’02, p. 231 and fig. 1). In the ages studied here (figs. 1 to 
4), however, there are two widely separated ganglionic masses, 
one upon the glossopharyngeal and the other upon the vagus 
nerve. Changes very like this occur in Anura according to 
Landacre and McClellan, for they describe a rather marked 
distinctness of the different ganglia of the postauditory complex 
in 8 mm. embryos of Rana pipiens. The fusion of these ganglia 
may be accounted for by the dilation of the auditory vesicle in 
later development, and the consequent pushing caudad of the 
glossopharyngeal ganglia upon those of the vagus. Before this 
takes place, however, there is a perceptible migration of some 
of these ganglia along their respective roots towards the brain, 
although this is not as clear as the migration of the trigeminal 
ganglia. The chief differences, therefore, between the cranial 
ganglia of Anura and those of Urodela depend, apparently, upon 
two factors—an active migration of the various ganglia towards 


THE NERVOUS SYSTEM OF AMPHIBIA 299 


. the brain and pressure exerted by the dilation of the auditory 
vesicle in its growth. 

In my discussion of the hyomandibular nerve of Amphibia 
(02, pp. 264-271) I pointed out differences between the Anura 
and Urodela which then seemed to me to warrant the hypothesis 
that the ramus alveolaris VII of the latter was a pretrematic 
nerve. The essential basis for that hypothesis was the apparent 
prebranchial position of the nerve in relatively advanced larvae 
‘“‘anteriorly of the deep pharyngeal evagination which represents 
the embryonic spiracular cleft”’ (’02, p. 228). This position, how- 
ever, must be secondary, for, as described in this paper, the first 
nerve that grows out from the geniculate ganglion enters the 
truncus hyomandibularis and passes behind the ectodermal-ento- 
dermal contact of the spiracular pouch. This position must be 
primary, as Emmel found it to be for the chorda tympani in the 
development of Microtus (’04), and the nerve must undergo very 
much the same kind of a shifting in position from an actual 
postbranchial position to an apparent prebranchial position just 
as the chorda tympani does in mammals, according to Emmel’s 
descriptions. Contrary to my earlier interpretation, therefore, 
the r. alveolaris VII of Urodela should be regarded as homolo- 
gous with the chorda tympani of mammals. 

; 6. Neurobiotaxis 

In numerous contributions (cited in his report which is men- 
tioned in the bibliography appended to the present paper) 
Kappers has presented an interesting array of data to show that 
the definitive position of the perikarya of motor neurones in the 
brain of vertebrates is determined by ‘‘a process of taxis or 
tropism occurring under normal conditions of nervous action, 
that is, under the influence of reception and propagation of 
stimuli.”” This process he designates neurobiotaxis. It involves 
the conception of a bodily migration of the motor neurones 
towards the chief source of stimulation by the afferent neurones 
or through the tracts of the brain. A discussion of this concep- 
tion as regards motor neurones does not belong in this paper on 


300 G. E. COGHILL 


the afferent system, but it is important to observe here that this 
paper describes specific instances of migration of cells masses in 
the form of several of the cranial ganglia. The clearest cases of 
this are the Gasserian and profundus ganglia, particularly the 
latter. This movement of the ganglia along the root towards 
the brain deserves more exhaustive and specific study, but some 
factors in the problem may be definitely stated. The ganglion 
of the ophthalmicus profundus V, for instance, in the earlier 
period is anchored to the skin far out over the eye. While this 
anchorage is intact root fibers make their connection with the 
brain. In embryos of the early flexure stage these fibers have 
established a firm anchorage to the brain by the bifurcation of 
the root beneath the external limiting membrane, and now their 
hold upon the skin is lost. Immediately a movement of the 
ganglion begins which finally carries it snugly up against the 
brain immediately around the entrance of the root (figs. 51 to 
54).. During this migration there are no apparent extrinsic 
mechanical factors introduced. A comparison of figures 1 and 
2 shows that the migration has begun before the ganglion comes 
into contact with the eye, and comparison of figures 76 and 78 
shows that after the ganglion comes in contact with the most 
dorsal portion of the eye it moves caudad over the surface of the 
eye—a median section through the eye in figure 78 showing 
only the tip of the ganglion while a similar section in figure 76 
shows the massive portion of the ganglion. The ganglion moves 
caudad, therefore, with reference to the eye as well as with 
reference to the brain and auditory vesicle. There must be, 
then, some intrinsic factor involved in this movement. 

As such a factor the root fibers with their firm achorage 
within the brain immediately suggest themselves (fig. 22). As 
has been described above, these fibers are growing very rapidly 
caudad towards the chief motor centers in the lower portion of 
the medulla oblongata and the upper portion of the spinal cord. 
This growth may be regarded not simply as an increase in mass 
but as an active stretching out of the whole fiber in an amoeboid 
fashion after the manner of the movements of nerve cells grow- 
ing in vitro as Harrison (710, 711) has described them. In other 


THE NERVOUS SYSTEM OF AMPHIBIA 301 


words, the entire axone may be thought of as creeping inward 
past its point of anchorage at the surface of the brain, as an 
amoeba would creep between obstacles, and as dragging the 
perikaryon after it. What the stimulus for such an action may 
be I shall not discuss here. I only wish to emphasize that this 
is a concrete case of migration of a group of neurones and that 
the rapid advance caudad of the root fibers within the brain and 
the concomitant shortening (not wrinkling) of the extracerebral 
root fibers convinces me that the axones are active factors in 
this. process. 

Another feature of the development of the sensory cranial 
nerves deserves mention in connection with the question of neuro- 
biotaxis, namely, the relative rates of development of the nerve 
trunks and their respective roots. In ease of the trigeminal 
nerve it has been noted in the anatomical part of this paper 
that the root connections with the brain are well established at 
a time when no peripheral fibers can be traced to the skin 
The lateral line primordia, also, are very small and are reached 
directly by the corresponding ganglia at a time when the latter 
have relatively long roots connecting with the brain. The most 
conspicuous case of this kind, however, is the visceral sensory 
system, the ganglia of which are connected with the placodes 
and are without peripheral nerves, excepting a few fibers of 
indefinite distribution from the geniculate ganglion, at a time 
when the root fibers have already formed a well defined fascic- 
ulus solitarius which reaches from the facial root to that of the 
vagus. 

It is apparent, then, that in the development of the sensory 
cranial nerves the axone runs far ahead of the dendrite. This 
suggests the inference that, from the point of view of taxes or 
tropisms, normal peripheral excitation does not stimulate the 
growth of the axone inward. This is not meant to imply that 
there can be no nervous processes in these neurones before the 
. dendritic terminals are established, but that, if there are such 
processes, they can have no reference to such stimuli as nor- 
mally act upon the corresponding definitive nerves and have 
significance in the reactions of the animal. Similar conditions 


302 G. E. COGHILL 


prevail in the development of the reflex mechanism of the trunk, 
where the Rohon-Beard cells have their typical connection with 
the skin and the motor cells have their terminals established 
upon the muscles for a considerable time before they become 
incorporated into a functional reflex are and are, therefore, of 
any value to the animal as regards behavior. From these facts 
it appears, then, that the behavior-value of nervous processes, 
if there be such in these nerve cells of early stages of develop- 
ment, has no regulating influence in establishing the primary 
plan of the integrating mechanism of the organism. What this 
regulating agency may be is one of the chief topics of interest 
in my study, but it can be best discussed upon the basis of an 
exhaustive anatomical and physiological analysis of the develop- 
ment of the various elements of the central nervous system dur- 
ing this early period. Such a study I already have well under 
way. 


Since this paper was submitted for publication new evidence 
of a ‘general chemical sense’ in the skin of animals has been pub- 
lished by Crozier (16) with particular reference to my sugges- 
tion (14, pp. 205-207) that the stimulating effect of substances 
in solution may be accounted for by their destructive action upon 
the epithelial cells. He concludes that there is ‘‘no ground for 
Coghill’s assumption that the cells of the germinative layer of 
the epithelium of fishes and amphibians are exposed to the action 
of the stimulating agent and thereby disrupted; and there is no 
histological evidence of disruption.” 

Without entering into further discussion or expression of 
opinion I wish here merely to raise the following questions, which 
seem to me to require settling before the main question at issue 
can be regarded as closed: 

1. Since the skin of the invertebrate is very different anatomi- 
cally from that of the vertebrate (the earthworm, for instance, 
having sensory nerve cells in the skin with endings free on the | 
surface) may the reactions of the one be legitimately Judged upon 
the basis of the structure and function of the other? 


THE NERVOUS SYSTEM OF AMPHIBIA 303 


2. Since, according to Crozier the ‘concentrations of irritants 
employed by Parker and others” ‘‘do not penetrate at all’ are 
there free nerve endings at the surface of the skin of vertebrates 
to function as receptors for these chemical stimuli (see Herrick- 
Coghill, 798)? 

3. Since the-epidermis of fishes and amphibians may be thought 
of as a system of three colloidal agglomerations, normally in 
equilibrium with each other—(1) the cellular protoplasm, and (2) 
surface films of mucus-like substance, which bathe opposite 
surfaces of (3) the cell membrane—may not very great dis- 
turbances of equilibrium with consequent violent effect upon 
neighboring parts result from the addition of a foreign solvent 
or solute to the surface film actually without the added solvent 
or solute penetrating the cell membrane (Nelson, °13, particu- 
larly table 5)? 

4. In order to stimulate the tactile nerve endings mechanically 
through the reaction of the epithelial cells must the action of 
the reagent be so violent as to produce histologically or chemically 
demonstrable effect upon these cells? May not relatively slight 
increase or decrease in the turgor of the epithelial cells be suffi- 
cient to stimulate? 

5. Why should the stimulating effect of water when ap- 
pled after the application of a chemical stimulus be evidence 
of destruction of tissue inone case (‘If any serious disintegration 
were produced by these solutions, it would be reasonable to expect 
the continuance of activity after external supply of the stimu- 
lant had been removed.’’ Crozier, 716, p. 3) and in another case 
be evidence of a ‘general chemical sense’ (‘‘Washing the foot 
with distilled water does not lead to cessation of the contraction, 
because, after exposure of the foot to certain solutions (Loeb 
05), water stimulates’? Crozier, 716, p. 5)? 

6. How can this last mentioned irritability to water after 
chemical stimulation be explained upon the hypothesis of a “‘gen- , 
eral chemical sense,’”’ and if it can be so explained why should it 
appear so conspicuously in the behavior of amphibian embryos 
(Coghill, 714, pp. 197, 198) the responses of which, Crozier con- 


304 G. E. COGHILL 


cedes, are probably ‘“‘not at all due to sensory stimulation by 
acid?”’ 

7. Granting that the so-called chemical stimuli act indirectly 
through the epithelial cells upon the nerve endings mechanically, 
may not the apparent physiological separation of the tactile 
and chemical receptors by use of cocaine still be accounted for 
upon the basis of the comparative strength of stimuli (Coghill, 
‘14, pp. 206, 207)? 

8. Regarding psychological differentiation of the two forms 
of stimuli as applied to the inner surface of the cheek of man, 
are the components of the cranial nerves of man sufficiently 
known to warrant the inference that morphologically equivalent 
receptors are concerned in the oral cavity of man and in the skin 
of the fish or frog? May not the visceral sensory system be 
a factor in the irritability in the oral cavity? 

Some of these questions I have under investigation and I 
shall probably discuss them further in a later paper. Meantime, 
in the interest of the leading problem of my investigations, I 
shall hope to see Crozier’s methods of inquiry projected further 
into the question concerning the existence of a ‘general chemical 
sense’ in the skin of vertebrates. 


THE NERVOUS SYSTEM OF AMPHIBIA 305 


BIBLIOGRAPHY 


Coeuitt, G. E. 1914 Correlated anatomical and physiological studies of the 
growth of the nervous system of Amphibia. I. The afferent system 
of the trunk. Jour. Comp. Neur., vol. 24, no. 2. 

Crozier, W. J. 1916 Regarding the existence of the ‘common chemical sense’ 
in vertebrates. Jour. Comp. Neur., vol. 26, no. 1. 

Emmet, V. E. 1904 The relation of the chorda tympani to the visceral arches 
in Microtus. Jour. Comp. Neur., vol. 14. 

Harrison, Ross GRANVILLE 1910 The outgrowth of the nerve fiber as a mode 
of protoplasmic movement. Jour. Exp. Zool., vol. 9, no. 4. 
1911 On the stereotropism of embryonic cells. Science, N.S., vol. 34, 
no. 870. 

Herrick, C. Jupson 1914 The medulla oblongata of larval Amblystoma. 
Jour. Comp. Neur., vol. 24, no. 4. 

Herrick, C. Jupson, and Cocuiti, G. E. 1915 The development of the reflex 
mechanism in Amblystoma. Jour. Comp. Neur., vol. 25, no. 1. 

Herrick, C. L., and Coeuitt, G. E. 1898 The somatic equilibrium and the 
nerve endings in the skin. Jour. Comp. Neur., vol. 8. 

Hooker, Davenrort 1911 The development and function of voluntary and 
cardiac muscle in embryos without nerves. Jour. Exp. Zool., vol. 11, 
no. 2: 

Kapprers, C. U. Arriins, 1914 Phenomena of neurobiotaxis in the central 
nervoussystem. XVIIth International Congress of Medicine, London: 
1913, Anatomy and Embryology. 

Lanpacre, F. L. 1907 On the place of origin and method of distribution of 
taste buds in Ameiurus melas. Jour. Comp. Neur., vol. 17. 
1910 The origin of the cranial ganglia in Ameiurus. Jour. Comp. 
Neur., vol. 20, no. 4. 
1912 The epibranchial placodes of Lepidosteus osseus and their rela- 
tion to the cerebral ganglia. Jour. Comp. Neur., vol. 22, no. 1. 

Lanpacre, F. L., and Conger, A. C. 1913 The origin of the lateral line pri- 
mordia in Lepidosteus osseus. Jour. Comp. Neur., vol. 23, no. 6. 

Lanvacre, F. L.. and McCietian, Marie F. 1912 The cerebral ganglia of the 
embryo of Rana pipiens. Jour. Comp. Neur., vol. 22, no. 5. 

Logrs, J. 1905 Production of hypersensitivity of the skin by electrolytes. In: 
Studies in General Physiology, pt. II, Chicago, pp. 748-765. 

Netson, C. FerpInanp 1913 Studies in osmosis, Jour. Amer. Chem. Soc., 
vol. 35, No. 6. 4 §Qp 

STREETER, GEorRGE L. 1 Some experiments on the developing ear vesicle 

_ of the tadpole with relation to equilibration. Jour. Exp. Zool., vol. 3. a re ‘* 


306 G. E. COGHILL 


EXPLANATION OF FIGURES 


All of the figures are taken from Amblystoma punctatum Cope, excepting 
figure 71, which is from A. microstomum Cope, and figures 11 and 12, which are 
from another species of Amblystoma, probably jeffersonianum. 

Methods. The methods followed in the preparation of Nos. 444, 449, 467 and 
473 were described in Paper I. In Nos. 548, 546, 545, 550, 557 and 593 the fixa- 
tion was with Van Gehuchten’s fluid, and staining with erythrosin and toluidin 
blue. For No. 455 fixation was with formol-Zenker solution; staining with 
Bomer’s haematoxylin and slightly acidulated orange G. For Nos. 561, 562 and 
567 fixation was with Zenker’s solution and staining with iron haematoxylin and 
orange G., excepting that the latter was omitted in No. 567. For 447 and 496 
the fixation was With corrosive sublimate-acetic acid mixture and staining with 
alum cochineal (447) and alum carmine (496) in toto with Lyon’s blue in 95 per 
cent alcohol on the slide. No. 635 was prepared with Paton’s silver method: 
neutral 10 per cent formalin about three months [formalin (40 per cent formal- 
dehyde) is saturated with sodium carbonate or lithium carbonate and the fixing 
fluid is made by mixing one part of this stock solution with nine parts of distilled 
water]; washed in running water over night; { per cent silver nitrate five days 
and four hours; 2 per cent nitrate of silver to which four drops of sodium 
hydroxide have been added for every 20 cc. of the solution, and then about 
twelve drops of ammonia, two hours; bath in distilled water to which one 
drop of glacial acetic acid has been added for every 2 cc., twenty minutes; 
1 per cent freshly dissolved hydrochinone, sixteen hours; washed in distilled 
water; dehydrated through graded and absolute alcohol in about 6% hours; 
cleared and embedded through cedar oil, cedar oil one part and paraffin one 
part, cedar oil one part and paraffin two parts, then pure paraffin, in two 
hours or a little less. My experience is that this method of clearing and embed- 
ding is very satisfactory in use upon amphibian embryos, in which the yolk is 
a very refractory element. The silver-impregnated material does not suffer 
from treatment with cedar oil for a day or more. 


ABBREVIATIONS 


Ad., area of adhesion between the skin 
and ganglion or otocyst as indicated 

Aud.V., the auditory vesicle or otocyst 

c, the branch of the n. mandibularis V 
to the balancer 

d, ramus palatinus VII 

DC, the Rohon-Beard cells of the spinal 
cord and cells of like character in the 
brain 

DT, the dorsal sensory tract of the 
spinal cord and its extension into the 
brain 

e,n. hyomandibularis VIT 

Ec.Th., thickened regions of ectoderm 

End., the endolymphatic appendage 
of the otocyst 


Ent., the point of entrance of the nerve 
indicated into the brain 

Fas.sol., the fasciculus solitarius 

g, ramus lingualis of the glossopharyn- 
geal nerve 

G.G., the Gasserian ganglion 

G.gen., the geniculate ganglion 

G. Jug., the jugular ganglion of the 
vagus 

G.L.L.VII, the lateral line ganglia of 
the facial nerve 

G.L.L.IX, the lateral line ganglion of 
the glossopharyngeal nerve 

G.L.L.X, the lateral line ganglion of 
the vagus 


THE NERVOUS SYSTEM OF AMPHIBIA 


G.oph., the ganglion of the ophthalmi- 
cus profundus 

G.VIIT, the ganglion of the auditory 
nerve 

G.Vis.IX, the visceral ganglion of the 
glossopharyngeus 

G.Vis.X, the visceral ganglion of the 
vagus 

h, the postbranchial division of the first 
ramus branchialis vagi 

L.o.p.V, the lateral terminal branch 
of the ophthalmicus profundus 

L.L.VII, \ateral line fibers of the facial 
nerve within the brain 

L.L.VII,Asc., the ascending divisions 
of the lateral line root fibers of the 
facialis 

L.L.VII,Des., the descending divisions 
of the lateral line root fibers of the 
facialis 

L.L.IX,X, the lateral line root fibers 
of the glossopharyngeus and vagus 
within the brain 

L.L.IX,X,Asc., the ascending divi- 
sions of the lateral line root fibers of 
the glossopharyngeus and vagus 

L.L.IX,X,Des., the descending divi- 
sions of the lateral line root fibers of 
the vagus 

M, myotome, the third in figures 1 to 4 

Mdb.V, ramus mandibularis trigemini 

‘Mes., mesenchyma 

M.o.p.V, the mesial terminal branch of 
the ophthalamicus profundus 

Mz., ramus maxillaris trigemini 

Neu. II, neurones of the second order 
in the reflex circuits 

Nuc.vis.m., the visceral motor nucleus 

Olf., the olfactory organ 

Op.St., the optic stalk 

o.p.V., the ramus ophthalamicus pro- 
fundus 

P.L., the pigment layer of the retina 


307 


R.G.G., root of the Gasserian ganglion 


- R.G.Jug., root of the jugular ganglion, 


or the somatic sensory root of the 
vagus ; 

R.L.L.VII, the lateral line root of the 
facialis 

R.L.L.IX,X, the lateral line root of 
the glossopharyngeus and vagus 

R.M., the root mass, a collection of 
chiefly indifferent cells about the en- 
trance of the nerve roots into the 
brain 

R.oph., the root of the ophthalmicus 
profundus 

R.V., root of the trigeminus 

R.V,m., the motor root of the trigemi- 
nus 

R.VII,m., the motor root of the facialis 

R.VII,vis., the visceral sensory root of 
the facialis 

R.VIIT, the root of the auditory nerve 

R.Vis.IX, the visceral sensory root of 
the glossopharyngeus 

R.Vis.X, the visceral sensory root of 
the vagus 

R.X,m., the motor root of the vagus 

R.XI, the spinal accessory root, or the 
most caudal motor root of the vagus 

t, a dorsal branch of the ramus oph- 
thalmicus profundus 

Tr.Asc.V, the ascending trigeminal 
tract 

Tr.Des.V, the descending trigeminal 
tract 

Tr.S, a tract in the motor zone, prob- 
ably bulbo-spinal 

V, nervus trigeminus 

VC, the somatic motor neurones of the 
ventral column 

VII, nervus facialis 

VIII, nervus acusticus 

VT, the somatie motor tract 

X, nervus vagus 


308 G. E. COGHILL 


Figs. 1 to 4 Graphic projections from serial, transverse sections upon the 
median plane to show the ganglia and sensory nerves of the head in the four 
physiological stages: 

Fig. 1 (No. 467), the non-motile. 

Fig. 2 (No. 473), the early flexure. 

Fig. 3 (No. 449), the coiled-reaction. 

Fig. 4 (No. 444), the early swimming stage. X 50. 

The brain is drawn in heavy lines; the myotomes and sense organs, in light 
lines; the general somatic sensory ganglia, in stipple; the lateral line and auditory 
ganglia, in diagonal lines; the visceral sensory ganglia, in rectangular cross- 
hatching. The lateral line nerves are not fully represented here. ‘They follow 
the course of the lateral line primordia, the distribution of which is shown in 
Paper I, figures 56 to 59. These figures with the descriptions in the text of this 
paper will give the important features of the lateral line nerves in the stages 
described. Silver impregnations of the early swimming stage show that these 
nerves are practically coextensive with the primordia of the organs. This is 
probably true of the earlier stages also, although satisfactory impregnations of 
the fibers have not yet been secured. The ganglia and sense organs have been 
projected mechanically with great care. The nerve trunks, which are too small 
to project distinctly at the magnification which has been used, have been sketched 
free-hand. 


THE NERVOUS SYSTEM OF AMPHIBIA 309 


RLLVi_ RLLIXX GLLX GJug. 


\ 
oe Be e.LLvi\RVis i eee 
- : \ 


ee aia 
| 


\ 


Brain G.VII | | GNis.X 
| | 
| G.L.LIX 
‘ 
GNis.IX 
| 
\ . RELAX LX 
EeTh, EnthV RM: cov GiuLix! RV s.X G.LL. 


Brain 


G.L.L.VII 
| G.gen. 


ela 


Mdb.V, 


THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL. 26, No. 3 


310 G. E. COGHILL 


G.LLVIl 
G.VII 


THE NERVOUS SYSTEM OF AMPHIBIA 311 


Des VNC LLIX,X,Ase. 
nt.V L.L.VilDes. Ent. b.L.1X,X 
Tr.Asc.V Ent. Vil L.L.IX,X, Des. 


L.L.VIl Des. L.L.IX,X, Asc. 
L.L. VII Ase. | | Ent.L.b.1X,X 
Tr.Asc.V | L.L.IX,X, Des. 
| 


Ent.V |Ent. Vil Ent. X Vis. OT 


R. VII. Vis. 
Ent. Vill 

Vill Des. 

Tr.Des.V 


Ent. 1X Vis. 


6 


Figs. 5 and 6 Projections from serial, transverse sections upon the sagittal 
plane, of the sensory root fibers after they have entered the medulla oblongata, 
in the coiled-reaction stage (fig. 5, No. 449) and in the early swimming stage 
(fig. 6, No. 444). & 50. The most rostrally located Rohon-Beard cell of the 
spinal cord is represented in each case, and the ascending fibers of these cells 
are shown as far cephalad as the root of the vagus nerve. The approximate 
extent of the trigeminal root fibers caudad is indicated in solid black, while the 
whole fiber system which I have called in the text the descending trigeminal tract 
but which is made up in large part of neurones of the second order is indicated 
in light lines about and beyond the root fibers proper. These figures are in a 
measure schematic, as will be seen by comparing them with figures 38 to 70, 
with which they should be studied. 


312 G. E.. COGHILL 


Figs. 7, 8,9 Non-motile stage (No. 467). X 50. 

Fig. 7 (section 1-3-12) illustrates the adhesion of the ophthalmic ganglion 
to the skin (Ad.G.oph.) dorsally of the eye, and the relation of the eye to the 
olfactory epithelium (OI/f.). 

Fig. 8 (section 1-4-8) is 120 caudad of figure 7 and shows the constricted 
nature of the root of the ophthalmic ganglion, although at this level it is still 
ganglionic, also the relations of the retina, optic stalk and brain. 

Fig. 9 (section 1-4-8) shows the adhesion of the Gasserian ganglion (G.G.) 
to the skin at Ad., just ventrally of the lateral line primordium (Sup.L.L.). 
These drawings are made from the same embryo as is figure 1, with which they 
should be studied. 

Fig. 10 Non-motile stage (No. 546, section 4-3-1). X 50. The section. lies 
approximately in the frontal plane, and shows favorably the two divisions of the 
ophthalmic ganglion (G.oph.) and the adhesion of one of these divisions with the 
skin (Ad.). Farther caudad the Gasserian ganglion appears in section (G.G.). 
A little farther caudad is the ectodermal thickening that is associated with the 
spiracular pouch (Ec. Th.). 

Figs. 11 and 12 Non-motile stage (No. 455, sections 2-1-10 and 13). X 50. 
The plane of section is longitudinal and latero-ventral, favorable to show the 
relations of the ophthalmic and Gasserian ganglia and their roots. 

Fig. 11 shows the two divisions of the ophthalmic ganglion, a and b. Fig. 12 
is 30 farther ventrad and shows the root connections of these nerves with the 
brain (R.V). Around the base of these roots is amass of cells (R.M.) which does 
not seem. to be a part of the ganglia at this stage. I have called this the root 
mass. The preauditory lateral line ganglion and the visceral sensory ganglion 
of the facial nerve (G.VIZ) are represented as a single mass. Its rostral end is 
in contact with the entoderm of the spiracular pouch. 

Fig. 13,14and15 Early flexure stage (No 473, sections 1-3-14, 1-4-6 and 1-4-12 
respectively.) x 50. These sections are selected from the series from which 
figure 2 was made, and should be compared with it. 

Fig. 13 illustrates the ophthalmic ganglion after it has detached itself from the 
ectoderm, leaving an ectodermal thickening at Ec.Th. 

Fig. 14 shows the attenuation of the root of the ophthalmic ganglion (R.oph.) 
160 » caudad of the last figure and at the level of the caudal portion of the eye. 


THE NERVOUS SYSTEM OF AMPHIBIA 313 


5 


zy 


314 G. E. COGHILL 


Fig. 15 is 60 u still farther caudad, showing the Gasserian ganglion (@.G.) con- 
nected by a slender root with the root mass (R.M.) at the surface of the brain, 
and the relation of this ganglion to the lateral line primordium (Sup.L.L.). 

Fig. 16 Early flexure stage (No. 550, section 1-2-14). xX 50. This figure is 
introduced to show the plane of section of figure 22, which is taken from the same 
section. ; 

Fig. 17 Coiled-reaction stage (No. 561, section 1-3-14). X 50. This figure 
is introduced to show the plane of section of figure 33, which was taken from the 
region indicated between the lines at a. 

Fig. 18 Early flexure stage (No. 545, section 5-2-14). x 500. A ganglion 
cell with its peripheral, dendritic process, in the distal portion of the ophthalmic 
ganglion. The spherules in the cytoplasm are yolk. 

Fig. 19 Early flexure stage (No. 598, section 2-5-7). > 500. Ganglion cells 
of the Gasserian ganglion, with axones reaching towards the brain. Other indif- 
ferent cells appear around the ganglion cells. The yolk is indicated in the cyto- 
plasm as in the last figure. 

Figs. 20 and 21 Non-motile stage transverse section (No. 543, sections 3-3-16 
and 13, respectively). > 500. 

Fig. 20 is 15 caudad of figure 21, and shows root fibers of the trigeminal 
nerve within the brain (7'r.Des.V). Both figures show cells of the root mass 
(R.M.) Fig. 21 a large cell (D.C.) which has the anatomical features of Rohon- 
Beard cells of the spinal cord sends a process out of the brain into the trigeminal 
root. Asimilar cell (D.C.) appears in figure 20, in which figure the motor nucleus 
of the trigeminus (Nue. vis.m.) is seen to be in a much more ventral position. 


THE, NERVOUS SYSTEM OF AMPHIBIA ald 


Nuc.vis.m. 


F.D 
- Be 


316 G. E. COGHILL 


R.L.L.VII 
Asc. 


AX RM. 
= y 4 Be 


Fig. 22 Early flexure stage (No. 550, section 1-2-14). X 200 The plane of 
section of this figure is illustrated in figure 16. It illustrates a longitudinal sec- 
tion of the medulla oblongata, in which the entrance of the trigeminal root ap- 
pears (R.V). There is a short ascending trigeminal tract (Tr.asc.V) and a 
descending trigeminal tract (T'r.des.V) which reaches to the level of the middle 
of the auditory vesicle (Aud.V) The ventricular pits opposite the entrance of 
the roots are conspicuous features of such a section. 

Figs. 23 to 28 Early flexure state (No. 473, sections 1-5-14 to 19). X 500. 
These figures are from six successive sections of the series from which figure 2 
was made, and show the relations of the root bundles of the seventh and eighth 


THE NERVOUS SYSTEM OF AMPHIBIA 317 


nerves as they enter the brain. Figures 25 and 26 show the root fibers of the 
lateral line component (?.L.L.V/I, a and b) and figures 24 and 23 show the ascend- 
ing divisions of these root fibers (L.L.VII,Asc.). Figure 25 shows also the root 
fibers of the visceral sensory component (R.VI/.vis.) of the facial nerve approach- 
ing the brain amid cells of the root mass, and figure 27 shows this component 
entering the brain. Immediately ventrally of this is the root of the eighth nerve. 
In the next section caudad, figure 28, descending divisions of the lateral line 
root fibers (L.L.VII,Des.) appear, but nothing can be seen of the other com- 
ponents of this complex. The entire cephalocaudal range through which the 
lateral line component can be recognized in this series within the brain is only 
60 mu. ; 


318 a. E. COGHILL 


Figs. 29 to 31 Early flexure stage (No. 473, sections 1-6-16 to 18). X 500. 
three successive transverse sections through the entrance of the root of the 
glossopharyngeal nerve into the brain. The lateral line component (R.L.L.1X,X) 
enters in figures 29 and 30, and the visceral sensory component (R.Vis./X) enters 
in figures 30 and 31 at a slightly more ventral and caudal level than the entrance 
of the lateral line component. No ascending or descending fibers of this complex 
can be made out beyond the limits of these three sections, a range of 30 u. 

Fig. 32 Early flexure stage (No. 473, section 2-1-12). .x 500. This figure 
shows the only perceptible connection of the vagus nerve other than the lateral 
line root with the brain at this time, a single nerve fiber (R.X.), which is of doubt- 
ful nature. It may be motor. 


THE 


NERVOUS SYSTEM OF 


AMPHIBIA 


319 


320 G. E. COGHILL 


Fig. 33 Coiled-reaction stage (No. 561, section 1-3-14). X 500. The plane’ 
of section is shown in figure 17, in which the region of this figure is included 
between the lines and indicated by a. The caudal end of the figure is to the left, 
and is a great deal farther dorsad than is the rostral or right end of the figure. 
The root of the trigeminal nerve (R.V) here approaches the brain in two divi- 
sions, the root from the Gasserian ganglion (R.G.G.) and that of the ophthalmic 
ganglion (R.oph.). The latter is entering the brain and bifurcating into a short 
ascending tract (7’r.Asc.V) and a massive descending tract (T'’r.Des.V). In the 
angle of this bifurcation is a large neurone (D.C.), the very large nucleus of which 
is in sharp contrast with the surrounding nuclei. This cell lies in the dorso- 
ventral level of the sensory root and is far removed from the motor nucleus, 
several cells of which can be recognized in a more ventral position (Nuc.vis.m.). 

Fig. 34 Coiled-reaction stage (No. 562, section 1-4-2). X 500. This is from 
a transverse section located 56 u rostrad of the entrance of the trigeminal root. 
It shows the presence at this level of ascending root fibers of this nerve (T'r.Asc.V). 
In a more dorsal position there is a cell of the Rohon-Beard type (D.C.) which 
has a long process extending into the immediate vicinity of or actually into the 
ascending trigeminal tract. 

Figs. 35 to 37 Coiled-reaction stage (No. 562, sections 1-4-9, 10 and 11). 
< 500. Three successive transverse sections through the brain at the entrance 
of the trigeminal root (R.V). 

Figure 35 is the most rostral in position, and shows two or three cells of the 
type described in the last two figures (D.C.). Another cell of this type appears 
in figures 36 and 37 (D.C.), two sections through the same cell. In both figures 
the peripheral process of the cell is shown entering the trigeminal root. In figure 
37 there is a stump of a dorsally directed process, and in figure 36, evidence of a 
ventrally directed process. Cells of the root mass (#.M.) appear in all these 
floures. 


321 


THE NERVOUS SYSTEM OF AMPHIBIA 


Nuc.vis.m. 


NY 


R. oph. 


he 


bas 


: eben 


322 G. E. COGHILL 


Figs. 38 to 41 Coiled-reaction stage (No. 447, sections 1-5-10 to 13). > 500. 
Four successive, serial, transverse sections through the entrance of the seventh 
and eighth nerves into the brain. These figures are from the same specimen as 
are figures 3 and 5, with which they should be studied. Figure 38 is the most 
rostral of the series. In this figure the root of the lateral line ganglion on the 
hyomandibular division of the nerve (R.L.L.VIT,b) is entering, and the ascending 
fibers of the other division (L.L.VII,Asc.a) appear in a slightly more dorsal 
position. These fibers are seen entering the brain in the three following figures 
(R.L.L.VII,a). The visceral sensory component of the facial nerve is entering 
the brain in figure 39, and its fibers appear as the fasciculus solitarius (Fas.Sol.) 
in the two following figures. The auditory root (R.V/I/) enters in figures 40 and 
41. All of these root systems lie dorsally of the descending trigeminal tract 
(Tr.Des.V), which consists here of scattered fibers through a rather extensive 
zone. 


THE NERVOUS 


= 


f 
mr 


] 


a 
§ 
a, 


SYSTEM OF 


AMPHIBIA 


324 G. E.. COGHILL 


Fas.Sol.pe / 


Des, ‘ ' 


Figs. 42 to 46 Coiled-reaction stage (No. 447, sections 2-1-6 to 10) X 500. 
Five successive serial, transverse sections through the entrance of the glosso- 
pharyngeal nerve into the brain. Figure 42 is the most rostral. In it can be 
readily recognized the ascending fibers of the lateral line component of the ninth 
and tenth nerves (L.L.IX,X,Asc.a and b), and ventrally of these the visceral 
sensory component of the facial nerve as the fasciculus solitarius (Fas.Sol.). 
The lateral line components are entering the brain in the three successive figures, 
while the visceral sensory (R.Vis.X) enters in figure 44. In the two following 
figures this component appears with the facial fibers as the fasciculus solitarius, 
and in figure 46 the descending divisions of the lateral line roots occur (L.L. 
IX,X,Des.a,b). The descending trigeminal tract (Tr.Des.V) appears as scattered 
fibers immediately beneath the external limiting membrance throughout this 
region. Neurones of the second order (New. I/) stretch across the whole mesial 
face of this system of root fibers. 


325 


THE NERVOUS SYSTEM OF AMPHIBIA 


»') 
Ace 3°, 
So fe 


ys 
nS 
7 " 


——— 


& 


Fas.Sol 
Nuc.vis.m. 


ele ho eae 


26, No. 3 


THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL. 


326 G. E. COGHILL 


Figs. 47, 48,49 Coiled-reaction stage (No 449, sections 1-6-17, 18,19). X 500 
Three successive serial transverse sections through the brain at the level of the 
entrance of the root of the vagus nerve. Figure 47 is the most rostral of this 
set: In it the visceral sensory root of the vagus (R.Vis.X) is approaching the 
brain surrounded by cells of the root mass (R.M.), and the fasciculus solitarius 
appears in two clusters of fibers (Fas.Sol.), a more dorsal representing the facial 
and glossopharyngeal portion and a ventral consisting of ascending fibers of the 
vagus root. ‘This figure, as well as the following, extends ventrad to show the 
relation of the sensory centers to the motor column (V 7’), and the visceral motor 
nucleus (Nuc.vis.m.). The root fibers of the vagus enter in figure 48, and in the 
following figure a few descending root fibers can be recognized (Fas.Sol.). 

Fig. 50 Early swimming stage (No. 567, section 2-1-19). > 500. From a 
transverse section through the medulla oblongata 35. caudad of the entrance 
of the trigeminal nerve. Around the dorsal margin of the descending trigeminal 
tract (7’r.Des.V) is a cluster of three or more large neurones (DG) of the giant 
ganglion cell type. Their processes run ventrad towards the motor nucleus 
of the trigeminus (Nuc.vis.m.). 


THE NERVOUS SYSTEM OF AMPHIBIA 327 


Tr.Des.V 


328 G. E. COGHILL 


Figs. 51 to 54 Early swimming stage (No. 557, sections 2-1-1,5,7,11). X 200. 
Four sections selected from a transverse series through the brain in the vicinity 
of the entrance of the root of the trigeminal nerve. The massing of the 
ganglia against the brain is shown here, a relation which should be noted in the 
study of figures 1-4. Figure 51 is the most rostral in this set, and figure 54 the 
most caudal. In figure 51 the ophthalmic ganglion (G.oph.) appears with a 
crescentic piece of the Gasserian ganglion (G.G.) on its. ventral surface. The 
fibers from this ganglion enter in figure 52 (R.oph.), while the root of the Gasserian 
ganglion enters in figure 53 (R.G.G.) more dorsally. This section is 28 4 caudad 
of the last. In figure 54 the motor root of the nerve appears. 


THE NERVOUS SYSTEM OF AMPHIBIA 329 


Tr.Asc,. V— ) 


— 


330 G. BE, COGHILL 


Figs. 55 to 60 Early swimming stage (No. 496, sections 1-5-4 to 9). X 500. 
Six successive serial transverse sections through the entrance of the roots of the 
seventh and eighth nerves into the brain. Figure 55 is the most rostral of the 
set. In this figure the ascending divisions of the lateral line component (L.L. 
VII,Asc.a,b) and the roots of the visceral sensory component (R.VIT, Vis.) and 
the auditory root R.VIII) are beginning to enter. The lateral line roots (R.L. 
L.VII,a,b) enter in figures 56 to 59. In figure 58, 59 and 60 the fibers of the vis- 
ceral sensory component have become the fasciculus solitarius (Fas.Sol.), and 
the descending auditory root fibers (VIJI,Des.) appear along the dorsal margin 
of the descending trigeminal tract (7r.Des.V). Neurones of the second order 
(Neu. IT) here lie across the entire sensory zone. These figures should be com- 
pared with figure 6. 


331 


THE NERVOUS SYSTEM OF AMPHIBIA 


L.L.VIlAse. 


Nuc.vis.m. 


332 G, E. COGHILL 


Fig. 61 Early swimming stage (No. 496, section 1-5-15). X 500. From the 
same series as figures 55 to 60, at the level of the middle of the auditory vesicle 
and endolymphatic appendage (Hnd). This should be compared with figure 6. 
It shows the relations of the descending facial and ascending postauditory lateral 
line components in this early stage; also the fasciculus solitarius (Fas.Sol.) and 
the descending auditory fibers (VIII Des.). Mesenchymal cells are now pressing 
in between the brain and the otocyst. 


THE NERVOUS SYSTEM OF AMPHIBIA 333 


Fas.Sol.__ 
~S 
RVI ses 
a) e 


a 


ViliDes. 


334 G. E. COGHILL 


Figs. 62 to 65. Early swimming stage (No. 496, sections 1-5-24 to 1-6-2). 
< 500. Four successive transverse serial sections through the entrance of the 
glossopharyngeal nerve into the brain. Figure 62 is the most rostral. The 
lateral line roots (R.L.L.JX,X) enter in figures 63 to 65, and on their ventral sur- 
face the visceral sensory root (R.Vis./X ) enters in figures 63 and 64. These figures 
should be compared with figure 6. 


THE NERVOUS SYSTEM OF AMPHIBIA 335 


= ~~ 
Nuc.vis.m. a! Bs Ao ec 
a Sy ma RAN 
Z Se a 


RiVisIX &, é 


Tr.Des.V \ 


Nuc.vis.m, 


336 G. EB, ‘COGHILE 


Figs. 66 to 70 Early swimming stage (No. 496, sections 1-6-19 to 23). > 500. 
Five successive serial transverse sections through the entrance of the root of 
the vagus nerve into the brain. Figure 66 is the most rostral in position. In 
this figure the fasciculus solitarius (Fas. Sol.) can berecognized between the 
descending trigeminal tract (T'r.Des.V) ventrally and the ascending fibers of the 
spinal Rohon-Beard cells dorsally (DT). The visceral sensory root of the nerve 
(R.Vis.X) enters in figures 67 and 68 and the somatic sensory root (R.G.Jug.) en- 
ters in figure 69; the visceral sensory root entering more dorsally. Figure 70 shows 
fasciculus solitarius fibers caudally of the entrance in a relation like that in 
figure 66, which is on the rostral side of the root. The somatic motor column is 
shown in 67, 69 and 70 (V7). The visceral motor nucleus appears in all the 
figures of this set, and lying between it and the external limiting membrane is a 
tract (Tr.S) which can not be described in detail in this paper. It is probably 
a bulbo-spinal tract. The figures of this set should be compared with figure 6. 


THE NERVOUS SYSTEM OF AMPHIBIA BBYA 


338 G. E. COGHILL 


Nuc.vis.m, 


Retinal Ganglion Cell 


TI 


Fig. 71 Early swimming stage of Amblystoma microstomum Cope (No. 635, 
section 1-3-7). X 500. A transverse section of the embryo through the eye, 
to show the nature of the ganglion cells of the retina and their fibers, which at 
this time decussate in the most rostral portion of the postoptic commissure. 
Silver impregnation after fixation in neutral formalin. 


THE NERVOUS SYSTEM OF AMPHIBIA 339 


gh 
SAG. VIII 


Se 
IS 


Figs. 72 to 79 From transverse sections through the eye and ear of the four 
stages. 


Figs. 72, 73 non-motile (No. 467). 
Figs. 74, 75 early flexure (No. 473). 


340 G. E. COGHILL 


ows NS « 


Ma reinks 
eer 
oO 


Figs. 76, 77 coiled-reaction (No. 449); early swimming, figures 78, 79 No. 
444). The sections through the eye were selected at levels which show the most 
extensive area of the lens, and those of the ear were.selected at levels which 
show the clearest differentiation of the endolymphatic appendage. 


THE EFFECT OF ACTIVITY ON THE HISTOLOGICAL 
STRUCTURE OF NERVE CELLS 


R. A. KOCHER 


From the Henry Phipps Psychiatric Clinic, Baltimore,’ and the George Williams 
Hooper Foundation for Medical Research, University of 
California, San Francisco 


INTRODUCTION 


The study of the effect of functional activity on the morphology 
of nerve cells has been the subjectof numerous investigations 
in the past. These researches have covered a considerable range 
as to kinds of cells studied, degrees of fatigue, and as to methods 
of treatment of the material. There is almost complete agree- 
ment on one point, namely, fatigue results in appreciable changes 
in cell structure. There is, however, the utmost divergence of 
opinion as to the nature of the changes, and one who tries to 
correlate the findings of the different workers in this field is 
utterly confused.. The chief findings relate to*(a) size of cell 
body and nucleus and (b) amount and distribution of chromatic 
material. While one group of workers concludes that fatigue 
results in decrease in the size of cell and nucleus (Hodge (1), 
Van Durme (2), Luxemburg (3), Legendre and Piéron (4), Lugaro 
(5), and Nissl (6) ), another group finds an increase in the size 
or a change in the nucleus-plasma relation (Crile-and Dolley 
(7), Mann (8), Vas (9) ). Lambert (10) noted no change in 
the size of cell or nucleus. Increase in chromatic substance 
-(hyperchromatism) as the chief result of fatigue has been noted 
by Nissl. Most workers (Van Durme (2), Vas (9), Lugaro (5), 
Mann (8) ) have deseribed a decrease in chromatic material as 


1 These experiments were carried out at the suggestion and under the direc- 
tion of Prof. Adolf Meyer, and completed in 1912. The computations were in 
part made after leaving Baltimore. It g ves me great pleasure to express here 
my thanks and appreciation to Doctor Meyer for his keen interest in the experi- 
ments and for many helpful suzgestions. 


341 


THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL. 25, NO. 3 


342 R. A. KOCHER 


the result of exercise. Luxemburg (3) described a breaking up 
of the chromatic bodies into granules, and Dolley tried to recon- 
cile these contradictions by regarding the hyperchromatism as a 
result of moderate activity, and the hypochromatism the result 
of excessive activity or exhaustion. They find that both stages 
may be present in the same animal simultaneously. Other 
changes, such as wandering of the nucleus toward the periphery 
of the cell (Magini (11), Lambert (10), Vas (9) ), rupture of the 
nuclear and cell membranes (karyolysis and karyorhexis) have 
been described. Dolley noted thirteen stages of cell change 
from hyperchromatism to disintegration or death of the cell, 
corresponding to the degrees of moderate activity up to com- 
plete exhaustion. Eve (12), in a study of sympathetic nerve 
cells, concludes that “there are usually some small differences 
before and after stimulation, but these are nearly all inconstant 
and generally reversible. Such divergence in the results is not 
so surprising when we stop to consider the complicating factors 
necessarily attending these experiments, such as, (1) difficulty 
of separating the effects of normal activity from unavoidable 
shock or injury to the nervous system in killing the animal (the 
nervous system does not ‘die’ as soon as the heart staps beating) ; 
(2) postmortem changes ensuing between the time of death and 
complete penetration of the tissue by the fixing agent due to the 
action of autolytic enzymes present in all tissue; (8) varying 
- chemical action of fixing agents; for example, formaldehyde coag- 
ulates protein by combination with the amino groups, alcohol 
by dehydration, sublimate by formation of salts, ete.; (4) the 
solvent action of materials used in fixation and in imbedding, 
for example, alcohol, xylol, paraffine; (5) varying effects of chemi-_ 
cal reaction between basic or acid dyes used in staining and the 
differént cell structures; (6) effect of subjecting tissue to tempera- 
tures of 50° to 54° in the paraffine oven for a period of several 
hours. 

In the present series of experiments I attempted as much as 
possible to minimize the formation of artefacts, having in mind 
the above mentioned considerations. It was hoped that by 
using special care in the handling of material, the use of im- 


EFFECT OF ACTIVITY ON NERVE CELLS 343 


proved technic in fixation and staining, and by varying the kind 
of degree of activity in a long series of experiments, using differ- 
ent animals and studying various kinds of nerve cells, that some 
degree of uniformity of results might be obtained. 


METHODS 


A resting control animal was used in each experiment. The 
animals were killed in most cases by bleeding after ether anaes- 
thesia, the nerve material removed as quickly as possible, cut 
into small pieces, and the control and fatigue specimens placed 
in the same fixing solution, imbedded side by side in the same 
block of paraffine, cut with the same stroke of the knife, mounted 
and stained together on the same slide. The detailed data of 
experiments showing animals used, kind of stimulation, length 
of time, microscopic technic, ete., are shown in table 1. In all, 
fifteen experiments were performed. Examination of the table 
reveals the fact that almost every form of activity was used; 
normal activity, forced activity, activity resulting from electrical 
stimulation, both faradie and galvanic, chemical stimulation, and 
shock being applied in the experiments. The kinds of animals 
used were dogs, cats, rats, sparrows, pigeons, and frogs. The 
microscopic technic used was varied considerably, not only as to 
fixative but also as to staining fluid. The stain most frequently 
used was Held’s modification of Nissl’s method. 

The examination of the material for comparison was facili- 
tated by having both control and fatigue specimens mounted on 
the same slide. Changes such as have been previously described 
as resulting from fatigue were carefully examined for; namely, 
comparative amounts and distribution of chromatic substance, 
size of granules, nucleus-plasma relation, relative size of cells 
and nuclei, ete. In order to determine whether any change in 
the size of the cell had resulted from activity, a large series of 
camera lucida drawings were made. These drawings were made 
of cells without selection. A field was taken, and every cell 
showing a nucleolus was included. This precaution was neces- 
sary in order to be assured that the cell was cut through a com- 


KOCHER 


R. A. 


344 


. 
. 


UIN]]9qQe.100 *9 
BIpsuvs 
poyjeul s ple Ivquin] puv [BoTAIe. *q ; 
an]{q eue,Aqyour pIng s,pleH BIjsuvs Iequint [[upeery 
emoiyoA[Odg |} pmp- s,zoyuez | puv [ewotaseo ‘Jessop ‘pv | ‘say ¢ pus 2z ‘T ssop | ul suruund 
yoyoo - urlBiq °9 [oom 
an]q xUIpINjO], “q |-]e yued s9d CG queuIesi1B[Ue IequIny *q ® UL SulUUNI APIATIOR 
S.Pp[PH “Y | pue ping s.pley | JUeUeSIe[Ue ,woTAIOD “D ‘say ZI “BO sjB1 | puv SurmuIMs | poo010 qF (q) 
joyoo UIBIG OTOYM *9 
onjq auIpNjOL “q | -[e yued aed cg BISuUBS [BIyOVAq “q 
S,Pp[PH “Y | pue ping s,pjex | UeWIESIv]UD [VIyORAG ‘Dp ‘say ZT “Bo | suoosid 
yoyoo UIBIG BJOYM *9 
an[q IUIPINjOT, “gq | -[e yueo sed GG juUSUIESIv[US [BIyoRIq *q 
S.pPeH “P| pue ping s.pjeH BI[SUBS [BIYOVIg “Dd ‘say ZT “Bo | smorieds sulAy 
ISSIN 
jO UOl}RBOyIpoUl SN4IAS 9}RIONIO *9 
S.PlPH 19}je joyoo wIn[[9qa1e0 *¢. 
auIsOIyyAII puB -[8 quad aed GG pioo jo JUuSUes1e[Ua suluuni AYIATIOB 
enjq osuejAyJoyy | puw ping s,ppey{ | [BorAte. puv azaequiny *p ‘say Ze ssop |pue SuUly[VM | [VUION (¥V) 
NIVLS NOILVXI4 aqaidows AOSSiL os eoreae IVWINV NOILVTOWILS 40 WHO 
T ATaVL 


345 


EFFECT OF ACTIVITY ON NERVE CELLS 


8 


anyiq sue,Aqyyour 


aMoIyoALO,g 


anjq sue, Aqjour 
aemorgoAl[og 
[ABIN, $0: GOR 


-BoyIpou SPOT] 


ISSIN JO WOT} 


“BOyIpour §,Pl°H 


eny{q eue|Aqyour 
aumor1yoAyTog 
Ureqs 8 PlPH 


uIsod pu 
anjq euIpmnjoy, 
anyq ouIpmyoy, 


ulyeur 
-10j jus Jed QT 


be 
-BULIOJ pu [OY 


-Oo]8 PING 8,PIPH 


PINY 8,PIOH 


yoyoo 
-[8 pus Ul[VULIO,T 


X9}100 IB] [9qQ9I109 
pue sniAd ayeronso 


p4109 
iequn, pue [BorAreo 


BIlsuvD rvquiny “q 
JUSWIOZIB[Ue LequIN] “Dp 


p4ioo requiny “q 
BYZUBS O1YBIOS “D 


eysues 
[vsiop pus sequmy *¢ 
p1oo jeurds jo 


lOyooTy | quowlesie[ue Lequiny, “Dp 
a a ee ee ee ee) ee ee eee 


980} BUI0) ‘sry ¢ 


“SI OI-T 


sry 9 
"sil ff 


(Yo “908 
ch pues uo 


‘008 CT) “sIy ¢ 


(‘oes aad 
“UIT}S) “SI TF 


ssop 


sd01J 


sdo1y 


syeo 


uossid 


sop 
‘DY Pp ‘poaour 
-91 poojq “00 
00% ‘“Brutevue yooyug 


#()—U01yN]OS 
4/88 Jo uoIsny 

-1od Y4IM “13 uoly 
f-} ourayoAys | -vjnurnys Snug 


AAIOU OTYBIOS 
jo woryeyn 
-UITYS OIpBIBy 


Se] Jo uolRN u0l}B[NUITYS 
-UIIYS OIUBATVS | [VITIZIOITY 


346 R. A. KOCHER 


parable plane. Cells were first drawn using a low power for 
orientation, each cell numbered on the paper, then the one-sixth 
objective or the one-twelfth objective was used to project the 
cell, care being taken to draw always at the same distance from 
the microscope. ‘The area of the cells was then computed with 
the use of the polar planimeter. The data are tabulated in 
table 2, and will be referred to later in connection with the indi- 
vidual experiments. 


Normal Activity 


Experiment I. Dogs. The animals which served for this experiment 
were two fox terrier puppies from the same litter, three months old. 
A female was used for activity while the male served for the control. 
The latter remained quiet in a cage, where he had been kept for several 
weeks previously. The activity animal was led by a chain on a fast 
walk into the country; the distance covered was fifteen miles in three 
and a half hours. This was a considerable feat for a puppy of this 
size, as the pace meant running all the way for her. At the end of 
three and a half hours, she was so fatigued that she refused to go any 
further, and had to be carried home. She was then killed less than 
one hour after exercise had ceased, the brain and cord at once removed, 
and sections taken from the lumbar and cervical enlargements, from 
the cerebellum, and from the cruciate gyrus, and the sections placed 
in 10 per cent formalin and in Held’s fluid. The control dog was killed 
at the same time, in the same way, and corresponding sections taken 
from the brain and cord, and placed in the same fixing fluid with those 
from the ‘fatigue’ animal. 


MICROSCOPIC STUDY OF THE NERVE CELLS 
Cervical enlargement of the cord 


The cells are uniformly stained, the Nissl bodies standing out 
clear and distinct. In both control and fatigue specimens, there 
is an occasional cell showing slightly clear areas about the nucleus, 
but this is no more marked in either section, and these cells are 
as numerous in the control as in the other. Drawings were made 
with the Leitz camera lucida, using the one-twelfth objective 
and no. 2 ocular. The camera lucida outlines of the cells and 
nuclei were traced with a planimeter with the following result: 

Control cells, 43 measured, average area 0.441 square inches. 


347 


EFFECT OF ACTIVITY ON NERVE CELLS 


‘Opl[s OURS OY} UO Pa}yUNOU Udaq DABY YOIYA\ SUOTIOIS [OIQUOD ONFIYBy oY} WOIy opeUl 
S}UDWIOINSBOUT VsoYy ATWO a1VdUIOD 0} 93BINIDB BIOUI PasOpISUOd SI 41 ‘A]JUOIOYIP poxY 10 ssouyory} JUdANyTp Jo oq yYyB1uT puw uorwser 
ewes oY} AT[ROTYMOpPT YBno1y} 4nd you 919M SUOTJOIS 9Y} BOUIg ‘“sepl[s JUsJOyIP UO epBUI syUNOD oyBavdes yueserdod OSOTLT, , 


02 LT OF Le | OS Po2NSBOU S{[9d JO LaqUIN NY 
"16890 O18TS0"OZ68-0)226-010G2 GieeeB OF Se a eee “PAIOU OTPBIOS JO 
S][99 Uol[sues 101104s0g 
i lk 
eo | 72 | 0F | Og Sh | ZF | $e pe poinsvoul s]jeo0 Jo 1aquinNy 
G66 0/8280) 169° 0) F660 GHG OTE OBST O)ZST'O) , | FP *"* UOT|sUBS [BI 
O6L' 016610 -yoRiq JO s[j9o uolsuery 
oe | 8), 
Og SF LT | 0€& | 62 | 62 | 6¢ re of c% Gr poinsBaut s[[e0 JO aquInNN 
260° 0\€920°0 GTE O/TPE'O) TE O/E9EO}ETS O/S6T'O | Fal'0) ‘er 0 ooce Ofc 0) umy[eq 
92Z'O\S8T'O f -a109 jo sjjeo ofurpaing 
8z | 9% |° °°‘ podnsvoul s[[oo Jo “ON 
l'0 ren olee Ol “paoo 
jo JUeUIESIB Ws IBquIN] 
jo s]j9o uwloy JOWeyuy” 
oni 8 6 | OL FI 6 8 Of €f | °°“ paansBoul sj[[90 Jo “ON 
GOT TEPG6°O)220° 12a 1 88% O/SCE 0 OfF GRP OOP” ee ae ge 
yusulesIB[Ua [BdTAII 
jo sjjeo uloy JOlo}UYy 
rr PN ey Sc ad ee Raa Re a a Ba a Pk fo ae ae u a ea 
spo S1140 AAMAN JO AGNI 
aL sNoapId spoa SLvu bag SMOUUVAS es 
INGWINGdX a 9 UNANINGdxXa ¢ INGWINGdXE f INANINGdXS i ae @ INAWINGdxXa ee sapere 


(pap.09a. jas yova fo woau ay}) sayour apnbs ur passaidxa auinjo A 
@ AIAVL 


348 R. A. KOCHER 


Fatigue cells, 36 measured, average area 0.420 square inches. 
These results are tabulated along with the measurements of 
cells from other regions and experiments in table 2. 


LUMBAR ENLARGEMENT OF THE CORD 


Examination of sections stained according to Nissl’s original 
method, e.g. “‘seifen methylen blue,” after fixation with 95 per 
cent aleohol, and according to Held’s modification of this stain. 
As regards amount and distribution of chromatic material, there 
is no difference between the control and fatigue sections. Several 
sections stained by Held’s method showed grouping of cells near 
the dorsal part of the anterior horns where the chromatolysis 
was slightly more marked in the fatigue specimen than in the 
control. This paucity of chromatic substance was not only 
around the nucleus but also extended to the dendritic trunks. 
This was not constant in all sections nor throughout the same 
section. Cells in the extreme anterior horn show no difference 
from the control. It seems probable that the particular section 
cut through a nucleus of the cord where the character of the cells 
was slightly different, since the similar variation in favor of the 
control cells was observed in some of the subsequent experi- 
ments. The nuclei of the control and fatigue animals show no 
difference whatever in the amount and nature of the staining of 
the chromatic material. 


CRUCIATE GYRUS 


These cells are uniformly stained, the nuclei and nuclear mem- 
branes and chromatic bodies being distinct. No difference in 
the morphology of these cells can be made out with the highest 
power of the microscope. 


CEREBELLUM 


The cells of the cerebellar cortex stained with Held’s method 
are well defined, and show a distinct architecture. Numerous 
cells were examined, but without any discoverable difference in 
the staining reaction between control and fatigue specimens. 


EFFECT OF ACTIVITY ON NERVE CELLS 349 


Sparrows 


Experiment 2. A male sparrow was shot and instantly killed with a 
rifle ball at 6 a.m. The brain, cord, and brachial and dorsal ganglia 
were removed one-half hour later. These were placed in Held’s fluid 
and in 10 per cent formalin. 

Fatigue bird. A male sparrow that had been flying about all day 
was shot, and the brain, cord, brachial and dorsal ganglia removed 
one-half hour later, and placed in Held’s fluid and 10 per cent formalin. 

Microscopic study. Brachial ganglia—one-twelfth objective, no. 2 
ocular, Leitz. The cell architecture is distinct in both the fatigue and 
control specimens. Sections stained with Held’s method show the 
cell bodies stained diffusely pink with distinct Nissl granules stained 
dark blue. There is no difference as regards depth of stain of either 
cell bodies or the amount and size or depth in staining of the Nissl 
bodies. No cells in either section showed any crenation of the nuclear 
membrane such as described by Hodge. Some cells show excentric 
nuclei and nucleoli, but by actual count this occurrence is just as com- 
mon in the control specimen. Examination of the anterior horn cells 
of the spinal cord in the region of the cervical enlargement as well as 
the examination of the Purkinje cells of the cerebellum show no varia- 
tion from the control morning specimens with respect to size of cells 
and nucleu§ or in the morphological markings. 


Pigeons 


Experiment 3. For this experiment pigeons were selected which were 
in the daily habit of making long flights, sometimes remaining on the 
wing for over an hour at a time. One of these pigeons was killed just 
at dusk, and as a control, one approximately of the same age was 
killed from the same flock at six o’clock the following morning. For 
study the entire brain was removed as well as the brachial ganglia and 
sections from the spinal cord in the region of the brachial and lumbar 
enlargements. 

Microscopic study. Brachial ganglia. The evening (fatigue cells) 
show no crenation, no central chromatolysis, and there is no apparent 
difference in the size or distribution of the granules from the controls. 
Measurement of a large number of cells show that the differences in 
size of the cell bodies or nuclei in the fatigue and control specimens fall 
within the limit of “‘variation.”’ 

Anterior horn cells. Brachial cord. No difference in any respect 
could be detected between control and fatigue specimens. 

Cerebellum. A large number of the cells were studied from both the 
evening and morning pigeons, but no constant variation in morpho- 
logical markings could be detected. 


350 R. A. KOCHER 


Rats 


Experiment 4. This and the following experiments differ from the 
preceding in that here the activity was forced to the point of exhaus- 
tion. In the previous experiments exercise was voluntary. Two half 
grown white rats from the same litter served for this experiment. They 
had previously been kept in a cage and well fed. The fatigue rat was 
kept running in a revolving wheel for one-half hour, having become 
tired, he refused to run, and clung to the wall of the wheel. The 
exercise was then changed from running to swimming. The rat was 
placed in a tank of lukewarm water, where he kept up constant swim- 
ming in an attempt to escape. At the end of one hour he was quite 
exhausted, was taken out, and allowed to rest for an hour. He was 
then made to swim a half hour again, followed by a half hour of rest. 
This was continued until the total time of swimming was three hours. 
He was then killed at the same time as the control. The total brain 
and portions from the cervical and lumbar cords were removed. The 
brains were cut sagittally, and placed in 10 per cent formalin; the other 
portions were fixed in Held’s fluid. 

Microscopic examination. In this experiment a large number of sec- 
tions were cut in series, and a thorough search made for constant 
differences in staining reaction, amount of chromatic material, size of 
cells and nuclei, etc. No such constant differences appeared as 
would go beyond the limits of simple variation. In some slides one 
might be quite sure of a preponderance of cells of a certain type; for 
example, showing central chromatolysis; but on actual counting and 
comparison, the number of such cells will be balanced by an equal 
number of the same type in the control. 


Forced activity 


Experiment 5. Dogs. Four young fox terrier dogs of approximately 
the same size and age were used for this experiment; one served for a 
control, the other three were subjected to continuous running for 
periods of one, two and a half, and five hours respectively. They were 
killed immediately after the exercise, and the nerve tissue from the four 
animals given identical treatment as to fixation, imbedding, staining, 
and cutting, the four pieces being mounted side by side in the same 
block of paraffine, and cut with the same stroke of the microtome 
knife. In this way the effect of exercise of various grades of intensity 
could be studied in the cells of the anterior horn of the cord, of the 
posterior ganglia, and of the cerebellum. 

Microscopic examination. Thorough study of all the sections num- 
bering over a hundred, most of which were made in series, was made in 
this experiment. The various types of cells described by Dolley were 
particularly kept in mind, and an attempt made to correlate them with 
various grades of fatigue. Dolley (7) describes thirteen different stages 
of fatigue corresponding to different grades of work and over-work. 


EFFECT OF ACTIVITY ON NERVE CELLS 351 


Representatives of practically all these types of cells were 
found in my specimens, from the resting control animal, as well 
as from those animals exercised for one, two and a half, and 
five hours. In order to determine as accurately as possible the 
relative proportion of these types of cells in the different speci- 
mens, a table was made out listing each of these cell types, and 
then beginning with a section under the microscope, every cell 
showing a complete nuclear membrane was taken in order, and 
checked in the proper column. In this way over three thousand 
cells were counted with the result listed in table 3. 

As will be seen in the table, the number of a particular type 
of cell varies considerably, but this variation is the same for the 
different animals. There are neither progressive changes in the 
morphology of the cells from rest to exhaustion nor are there 
any qualitative or quantitative differences in type of cells from 
resting and fatigued or even exhausted animals. The animals 
- used in Dolley’s experiment exercised at most up to three to four 
hours altogether. Dolley also describes these types of cells as 
persisting in the effort to recuperate for from two weeks to several 
months after exercise no more severe than that of two or three 
hours running in a treadmill. All these types of cells are ad- 
mittedly present at the same time (Dolley, American Journal of 
Physiology, vol. 28, p. 151). Dolley describes thirteen stages 
of fatigue in one animal where the animal exercised one hour in 
a tread mill. Also, the cells selected by him for illustration of 
these stages are taken from a single preparation of three sec- 
tions in the same experiment. Obviously the observations were 
not over a large enough range of sections nor sufficiently con- 
trolled by actual counts of the various types of cells. In the 
Journal of Medical Research, vol. 21, p. 104, Dolley says, 
‘“Measurements were made of five cells of each type in five 
anaemia experiments, one a fatal resuscitation, the other a re- 
peated hemorrhage;’’ a little farther on, ‘‘Measurements were 
made of ten cells in each of three groups.”’ A great many cells 
were skipped (those not entire) in Dolley’s method of counting, 
giving a large leeway for a personal factor in the selection of 
types. Ibid, volume 20, page 291, ‘‘ Measurements were made 


KOCHER 


A. 


R. 


« 


352 


ec 


00€ L 9 al &@ 


00€ L L g OT 
00€ 9 ) al Vv 


él O& 8G 6 96 


IT GG 0Z 0 9€ 
9 ifs GG 0 SI 


‘pezld Used SvY YOIYM 0} sdUEIEJoI oY} ‘AaTIOG 
JO WOTJBOYISSB]D oY} WOrz Uye} oe “OYo “g ‘Z “T 9BVYS JOpUN poJBUTISep dsoY] 8B [[IM SB S]]90 poyuNooUN ay} “gE UUINOD “TTT eqz I, 


I ai IT Or 
85BIg | 951g 


&G 0& 1G ¥G 
6 8 ZL 9 g 
a5RYG | oBBIg | o5RIG | OBRIG | ERIC 


I100 ONILSAY 


LI LG cI 
VG 9G O€ 
O€ 02 0G 
Ji! &@ 0G 
(G0 GG SE 
OL OT VG 
LI GG cE 
ST cE TV 
9T 61 al 
9T €& VG 
61 83 GE 
SI rot FI 
t (5 G 
95819 | 25g | aBRIg 


66 G9 | °* p4o0d Jo yuUeUT 
-951B[U9 [BOTAIID 
Soran . Banos 
6F gg pi00 requinyT 
ze 06 |occctcctt UILIO AA 
0g 90T |°° pao Jo yuour | 
-oB1B[Ud [BOIAIOD || sunoy 
0S Sia ee P2109 1EQuuy 1G 
GI 7 ULIO AA 
0g 88 |°*paxoo Jo yuour 
-99.1G[U9 [BOIAIOD 
see ; IOUT: 
se 09 pioo requin'yT 
0z GOT Ge oe eae ULIO MA. J 
Ze OOM oa p41oo Jo yueur 
ye | eige 
oP SF pekee -m0 
> ) 
[euids jo yuour 
-ds.1B[Ue IBqQuUIN'T Ly 
LT JG. \Egcaee erase ULIO MA J 
ke = zZ ee 
eg aD eShe aastouaxa 
: 4 aWIL 


yaa $7199 00g fo sjunod pouuasaffip Buinoysy 


€ HTaVL 


EFFECT OF ACTIVITY ON NERVE CELLS Goo 


of camera lucida outlines of cells and nuclei. This was done in 
fourexperiments. . . . . Five cells of each type and after 
each fixative seemed to give a fair average.’ Dolley’s material 
in which consecutive stages were studied did not receive the same 
treatment as to fixation, staining, mounting, and cutting. The 
control and fatigue material was handled entirely separately 
(Amer. Jour. of Physiol., vol. 25, p. 155). Slight unavoidable 
variations in the exposure of the tissue to the various agents 
and different thickness of the cut sections would make such 
material worthless for comparative study. As pointed out above, 
this objection cannot be applied to my own experiments, as the 
handling of material was identical in all cases. Dolley’s method 
of measuring size of cells and nuclei is open to serious objection. 
In the Journal of Medical Research, volume 21, he states, ‘‘ While 
not adapted for exact measurement on account of their shape, 
the largest diameters (cells and nuclei) were always taken.” 
In my own experiments the actual relative areas of the nuclei 
and cells was computed. with the use of the polar planimeter, 
which traces the entire outline of the cell and gives very accurate 
results. As seen in the table, I found no difference in the size of 
the cells of the control and exercise animals. 


Electrical stimulation (galvanic) 


Experiment 6. Pigeons. In this experiment a pigeon was fastened 
by the feet in a standing position, while a wire from four 50 amperes 
galvanic batteries was wrapped around the legs. A current was al- 
lowed to pass through the wire once a second by a clock make and break 
arrangement. With the entrance of the stimulus there was a strong 
contraction of the leg and thigh muscles. After four hours the con- 
tractions became feeble, owing to fatigue, and the pigeon was killed. 
Rigor mortis of the leg and thigh muscles set in immediately. A con- 
trol pigeon was killed at the same time. 

Microscopic study revealed absolutely no constant morphological 
differences in the anterior horn cells and in the dorsal ganglion cells 
from the two birds. 


Faradic stimulation 


Experiment 7. Faradic stimulation of the sciatic nerve in a cat was 
applied in this experiment. A cat weighing 2 kilos was anaesthetized 
with ether, and decapitated according to the method of Sherrington by 


354 R. A. KOCHER 


tying off the carotids, ete. After a rest of about three-quarters of an 
hour to give the animal a chance to recover somewhat from the shock, 
a canula was inserted into the carotid artery, and connected with a 
manometer for record. The left sciatic nerve was then exposed and 
stimulated with a faradic current from two dry cells. The secondary 
coil was placed at eight, later at six. Stimulation was applied at inter- 
vals, fifteen seconds stimulation was followed by forty-five seconds 
rest. Stimulation began at twelve o’clock and continued until five 
pM. The heart was still beating strongly at the end of the experiment, 
the blood pressure remained fair, and reflexes were obtained through- 
out by stimulation of the sciatic, both in the right and left leg. The 
contractions on the right were spasmodic, those on the left—the stimu- 
lated side, were tetanic. The latter were feeble, and gave all signs of 
fatigue. The three pairs of dorsal ganglia of the sciatic nerve as well 
as the lumbar and sacral cord, were removed and placed in the fixing 
solution. Microscopic examination of the cells and measurements of 
cells and nuclei from the fatigued and unstimulated side of the cord 
of the same animal failed to disclose any difference in morphology. 

Experiment 8. Faradic stimulation of the sciatic nerve of the frog 
was used in this experiment, the unstimulated dorsal ganglia cells and 
anterior horn cells of the unstimulated side of the same animal as well 
as corresponding material from a second resting frog killed at the same 
time served as control. The two frogs were pithed and placed in a 
moist chamber. One-half minute stimulation of the sciatic nerve was 
followed by one minute rest. The electrodes were applied just above 
the knee at the back. This interrupted stimulation was continued six 
hours. The muscles showed response by tetanic contractions. At the 
end of the period the muscles were still irritable. Three pairs of dorsal 
ganglia corresponding to the sciatic nerve as well as the spina’ cord of 
the same region were taken for study. 

The stimulation produced no changes in the cell morphology that 
could be detected by a measurement of the size of the cells and nuclei 
or by study with various powers of the Leitz compound microscope. 


Drug stimulation (strychnine) 


Experiments 9 to 14... Verworn (13) attributed fatigue in nerve cells 
in part at least to a local asphyxia of the cells due to an accumulation 
of fatigue substances and an insufficient supply of oxygen. By per- 
fusing fatigued frogs with oxygenated salt solution, inserting the canula 
into the aorta, he was able to restore irritability of the nerve cells and 
corresponding response of the muscles in contraction after they had 
ceased to respond from fatigue. He succeeded in keeping the muscles 
and nerve responsive to stimulation for many hours longer than would 
ordinarily be the case. In experiments 9 to 14 this method was applied 
in order to continue stimulation and corresponding nerve exhaustion 
to a degree not possible by the usual methods. Strychnine in doses of 
4 to + of a grain was given in each case by injecting into the subcu- 


= 


EFFECT OF ACTIVITY ON NERVE CELLS 355 


taneous area of the lower back just after beginning perfusion with 
normal salt solution and by adding to the perfusion solution. The 
heart continued beating throughout the experiment. The strychnine 
caused violent tetanic contractions at first, which later gave place to 
continued fibrillary contraction. This served as an indication of the 
condition of irritability of the spinal cord cells. To prevent direct 
action of the strychnine on the nerve endings in the gastrocnemius 
muscle, the femoral artery was tied off. This muscle continued in a 
state of irritability for several hours. The experiments carried out by 
this method were continued for a length of time varying from two to 
ten hours. Controls were used, treated in the same manner, except 
for the strychnine injections. The subsequent treatment of the nerve 
material (sections of the spinal cord and from the thoracic and lumbar 
regions) was identical. There is a striking change in both the control 
and treated cells, which had been perfused for a long period. Many 
cells as well as nuclei are swollen beyond the normal size, owing, doubt- 
less, to a simple turgescence caused by a flooding of the circulatory 
system with the salt solution, and the staining reaction is diffuse. 
There are, however, no differences in morphological characters between 
the cells of the strychnine stimulated and the control cells of the spinal 
cord. 


Shock (anaemia) 


Experiment 15. A four kilogram dog was given morphine and bled 
from the jugular vein; 200 ce. were removed, following which the dog 
remained semi-comatose for five hours. At the end of this time he 
was killed along with a control dog of four and a half kilograms weight. 
Material was taken from the cruciate gyrus and cerebellar cortex, fixed 
in 10 per cent formalin, and stained with polychrome methylene blue. 
Examination with the microscope revealed no difference in the char- 
acter of the nerve cells in the shock animal as compared with the 
control. 


DISCUSSION OF RESULTS 


It has long been known that prolonged activity of mucous 
gland cells results in characteristic histological changes in these 
cells, owing to the disappearance from the cell of certain granules 
(zymogen granules). This granular material is evidently used 
to make organic material of the secretion. It was doubtless on 
the basis of such observations that certain physiologists were 
led to seek for similar alterations in the highly specialized nerve 
cell following functional activity. Many of the investigations of 
these physiologists were confined to one or two experiments, the 
material was often not sufficiently controlled by normal tissue 


356 R. A. KOCHER 


for comparison, and frequently the histological technic was faulty. 
In explanation of the diverse changes described by these workers 
as resulting from fatigue, it has been assumed that certain mate- 
rials present in the nerve cells have been katabolized to form 
energy for the nerve impulse, and that the using up of this 
material during activity can be detected histologically by the 
depletion of the granules (chromatic substance), changes in size 
of cell and nucleus. In a long series of carefully controlled 
expetiments, I could find no evidence of any analogy between 
the effect of activity in certain glands and activity in nerve cells. 
In no experiment did the histological structure of the nerve cell 
following activity show any constant deviation from that of the 
corresponding resting cells of the controls. Some very sweeping 
generalizations have been drawn from the conclusions of pre- 
vious workers; namely, that fatigue, fear, shock and exhaustion 
may lead to permanent damage and even disintegration of nerve 
cells. Crile’s present theory of surgical shock and of certain 
aspects of Graves’ disease, based essentially on these assu np- 
tions, may be cited to show to what extremes these deduct ons 


based on insufficiently controlled experiments of this kind have 
led. 


SUMMARY 


The effect of various grades of activity on nerve cells was 
studied in a series of fifteen separate experiments. The animals 
used were dogs, cats, pigeons, sparrows, frogs, and rats. «very 
experiment was carefully controlled by a resting animal of the 
same species, of the same approximate age and size, and the 
material from both given identical treatment, except for the 
activity. The nerve cells studied were from the cruciate gyrus, 
from the cerebellum, from the anterior horn of the spinal cord, 
and from the dorsal ganglia. In one of the experiments over 
thirty-five hundred nerve cells classified into thirteen types ac- 
cording to histological characters were counted to determine the 
relative frequency of characteristics which might be correlated 
with grades of activity. There was no deviation from the nor- 
mal in even the most advanced fatigue. Over a thousand cells 


EFFECT OF ACTIVITY ON NERVE CELLS SOT 


and nuclei were measured by computing the areas of the pro- 
jected outlines with the planimeter. There was found to be no 
constant difference in size of cells ornuclei resulting from activity. 
Furthermore, no qualitative differences in histological charac- 
ters could be found between fatigue and resting nerve cells. 


BIBLIOGRAPHY 


(1) Hopes, C. F. 1892 Journal of Morphology, vol. 7, p. 95. 1894 vol. 9. 
1887 American Jour. of Psychiatry, vol. I, p. 471. 

(2) Van Durme, Pav. 1901 Nevraxe, p. 115. 

(3) Luxemsura, J. 1899 Neurol. Zentralblatt, vol. 8, p. 629. 

(4) LeGenpre, ReN#, et Pifron, H. 1908 Comptes Rendus de la Soe. de Biol. 
p. 1102. 

(5) Luearo, E. 1895 Archives Italiennes de Biologie, Nr. 24, p. 258. 

(6) Nisst 1895 Neurol. Zentralblatt, Nr. 15, p.39. 1896 Zeitschr. f. Psychi- 
atrie, vol. 48. 

(7) Dottey, D. H. 1909 American Jour. of Physiology, vol. 25, p. 151; Crip, 
1911 Jour. of Med. Research, p. 309; 1909, p. 95. 

(8) Mann, G. 1895 Jour. of Anatomy and Physiology, vol. 29, p. 100. 

(9) Vas, F., Archiv f. mikr. Anat., 1892, Bd. 40, p. 375. 

(10) Lampert, M. M. 1893 Comptes Rendus. 

(11% Maaini, G. 1895 Archives Italiennes de Biologie, Nr. 22. 

(12) Evn, F. C. 

(13) ,Verworn, Max 1900 Archiv f. Anatomie u. Physiologie, supp!. p. 152. 


THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL. 26, NO. 3 


CONTRIBUTIONS FROM THE ZOOLOGICAL LABORATORY OF THE MUSEUM OF COMPARATIVE ZOOLOGY AT 
HARVARD COLLEGE, No. 278, AND FROM THE ANATOMICAL LABORATORY OF THE NORTHWESTERN 
UNIVERSITY MEDICAL SCHOOL, No. 36. 


THE INFLUENCE OF LIGHT AND TEMPERATURE 
UPON THE MIGRATION OF THE RETINAL 
PIGMENT OF PLANORBIS TRIVOLVIS 


LESLIE B. AREY 


NINE FIGURES 


CONTENTS 

SULPGUUCTION BN BIGLOLICAL SUTVOY.«....cceces onus awancswvepENauneseawecccs 359 
Serer ett YAEL NEDRHIONNRS p 2's Wiis viva a vedic 3 WKS aes so o's ope ERM ET ate «iain 362 
Wesersption of the eye of Planorbis. .... 2. 6...0. 500s cewensie see ghia ae sles so 365 
RNP MIATN PORES 5 cane yc ad dine <c,csiceg.« apo > «5.6 cieh eles Regen OA Sie ly 369 
Be Pneteph Of itu BHO CATATIONS.....0dcsesis ei cnscciesas Rab eeNeNe book eae 369 
fPoeeCr Olt ORMIAL GOIUOGIS. . 5 oa... cca he awh Od aE Es 6 sues 370 

Di OCL OT ORDIBGO OV OMe soi ok Sida dio ds ae cach s andein een kha «bs 371 

by Determination of adaption times..::..5.. 6.05. ccs0cendeeemeen venieme 374 
ees Ue OG PAIRING be ci 530 5 0 cam bap on Dap ¥ >< here's wae ee ale owe 376 

em Teu OT POTIiAl GIMNEIG, ... ais > ctcre s'v.ais a 5 <<'9 sala p ete RE eee ees 376 
METER OL OXCISG. CYOH: « s <u a-c sis t's Se oy vin 5's a kts So ee ee 377 
SERGI ASE EIORUIUION yale w wie <div oles aay «04 2b hve e oS RED OEE 378 
eS a ee eee Ae eee eS 379 
I Sot da Fa x Vs Rites REA Le sy 41 Re ARDS UE be ks AU 385 
ve Gl A ee ae Oe a ee Ame 8) ae 386 


INTRODUCTION AND HISTORICAL SURVEY 


The pigment granules of many animal cells, which in general 
are termed melanophores, are capable of undergoing positional 
changes when adequately stimulated. This property has made 
possible the development of an interesting field of research, 
which has had for its goal, the correlation of migratory move- 
ments of the pigment with the presence of definite stimulating 
agents. At first, only the influence of such important environ- 
mental factors, as light and temperature, as well as the effects 
of more artificial stimulating agents like pressure and electricity 
were sought, and but little attention was given to the réle played 
by chemical agents. Recently, however, the possibility of 

359 


THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL. 26, NO. 4, 
AuGust, 1916 


360 LESLIE B. AREY 


chemically stimulating certain types of isolated pigment cells 
has been shown in a striking manner (Spaeth, 713; ’13>). 

All investigations relating to the physiology of pigment 
cells may be divided into two general categories, depending 
on whether body chromatophores or retinal pigment cells serve 
as the material for experimentation. The general tendency, 
however, to keep these fields of research sharply separated is 
wrong, inasmuch as the behavior of the two types of cells has 
much in common, and it is only by summating our knowledge, 
after the comparative method, that a thorough understanding 
of either may be gained. 

Light is the commonest and most potent of the natural stimu- 
lating agents concerned in pigment migration. In general, it 
may be said that light induces an expansion, and the absence 
of light a contraction, of pigment cells. 

Photomechanical changes have been demonstrated in tie 
body chromatophores of crustaceans, cephalopods, fishes, am- 
phibians, and lizards. The earliest experimentation showing 
the effect of light upon the retinal pigment of vertebrates was 
performed by Kiihne (’77) and by Boll (’78), who worked upon 
the frog. Their results have since been extended upon repre- 
sentatives of the remaining vertebrate classes, although strik- 
ing movements of the retinal pigment of reptiles and mammals 
have not as yet been demonstrated. Among invertebrates, 
photomechanical changes have been found in the compound 
eyes of insects (Exner, ’89), crustaceans (Exner, ’91), arachnids 
(Szezawinska, 91), and in the eye of cephalopods (Rawitz, ’91). 

Of the molluses, the cephalopods constitute the only group 
in which a response of the retinal pigment to photic stimula- 
tion has hitherto been demonstrated. Hensen (65), upon 
theoretical grounds, hazarded the guess that the pigment in 
these animals possessed a certain mobility, for Babuchin (764) 
had previously referred to finding the visual rods entirely free 
from pigment in some specimens of Sepia and Octopus. It was 
left for Rawitz (91) to show that in the light the visual rods 
of Sepia officinalis are pigmented along their whole length, with 
an especial accumulation at the lens border, whereas in the 


RETINAL PIGMENT OF PLANORBIS 361 


dark the pigment is limited to the base of the rod. Chun (’03) 
confirmed these general conclusions on deep-sea cephalopods, 
as did Hess (05) by making a comparison of the pigment re- 
sponses in pelagic and littoral forms. 

Temperature is likewise a controlling factor in determining 
the ultimate distribution of pigment in many body melanophores. 
In general (Parker, ’06), low temperature has an influence similar 
to that of light, inasmuch as it favors pigment expansion; high 
temperature, on the contrary, like darkness, induces pigment 
contraction. 

A similar condition holds for the retinal pigment of certain 
animals. Congdon (’07, p. 547) found that: “In both Palae- 
monetes and Cambarus the proximal retinal pigment migrates 
distally when the temperature is lowered and proximally when 
it is raised.”” The writer (Arey, 16*) found an identical ten- 
dency in the retinal pigment of fishes. Working with the 
frog’s retina, Gradenigro (’85) first showed that at a tempera- 
ture of 30°C. the pigment expands, thereby closely simulating 
the distribution characteristic of light. Later Herzog (’05) 
confirmed this discovery, and further stated that maximal 
expansion likewise ensues when the temperature approaches 
the freezing point, the contracted condition typical of darkness 
being obtainable only between the temperatures of 14° and 18°C. 
A reinvestigation of this matter recently made by the writer 
(Arey, 16") has convinced him that Herzog’s results are sub- 
stantially correct and that the behavior of the retinal pigment 
of the frog to temperature must indeed be considered exceptional ; 
the explanation for this is presumably to be found in the exist- 
ence of a superimposed nervous control, resulting in the oblitera- 
tion of the more primitive response. 

Positional changes of the retinal pigment of gasteropods in 
response to definite physiological stimuli have never been re- 
corded. Smith (’06), in his work on the structure of the eyes 
of pulmonate gasteropods, noticed that the retinal pigment 
of Planorbis trivolvis Say showed processes of variable length, 
and accordingly he performed a few experiments to determine 
whether the pigment distribution could be correlated with 


362 LESLIE B. AREY 


corresponding conditions of light and darkness. His results 
concerning a photomechanical influence were entirely negative, 
although he states his belief (p. 255) in the existence of an appre- 
ciable pigment migration caused by certain unknown factors. 

Assuming that the retinal pigment of these animals does 
undergo migratory movements, it is reasonable to expect that 
the behavior of pigment cells, having no demonstrable nervous 
connections (Smith, ’06), is dependent upon definite environ- 
mental stimuli, hence I decided to reinvestigate the influences 
that simple stimulating: agents such as light and temperature 
might have upon the retinal pigment of these snails. 


MATERIAL AND METHODS 


Planorbis was obtained in abundance at certain localities 
on the bank of the Charles River, Cambridge, Mass. During 
warm weather it is easily procurable within arm’s reach of the 
shore, but as the water grows colder a gradual withdrawal of 
animals occurs toward the deeper bed of the river. My experi- 
ence in collecting during the fall of 1914 will illustrate this 
behavior. In the second week of October a great wealth of 
material was discovered at a location where the previous June 
no specimens had been seen; at the time of this collection the 
snails were distributed as far up as the water’s edge. The 
next collecting trip was made on November 21, and, in the 
same spot, where several weeks previously hundreds of animals 
were obtainable, there now remained a few stragglers only, 
and these were withdrawn from the bank almost beyond reach. 
Ten days later scarcely a specimen was left within sight. 

I was interested to discover whether this withdrawal was due 
to an active migration of individuals, or to the effect of the 
low temperature which might cause the snails to become inac- 
tive, whence they would be washed by the current down the 
sloping bank into deeper water. Mr. W. F. Clapp of the Mu- 
seum of Comparative Zodlogy informs me that the former of 
the alternatives is undoubtedly correct. Moreover, low tem- 
perature does not necessarily cause these animals to become 
inactive, since he reports having often broken through the ice 


RETINAL PIGMENT OF PLANORBIS 363 


in mid-winter and dredged Planorbis from the deeper water 
to which they resort; under these conditions he has always found 
them to be in an exceedingly active condition. 

This interesting migratory behavior to what is presumably 
the stimulus of low temperature has a definite ‘protective’ 
value to the snail, since an animal thereby removes itself to 
a depth at which ice is not formed. It is probable that the 
situation is not one in which an annual rhythm, independent 
of temperature, has been impressed upon the species, for when 
Planorbis, during its retreat to deeper water, is removed to 
laboratory aquaria and kept for some time at room temperature 
it neither manifests a tendency to seek the deepest water, nor 
does it exhibit negative geotropism. On the other hand, I 
do not recall that animals introduced into jars of ice water showed 
any marked tendency to seek the bottom of the dish. It is 
possible that there is, in fact, a temperature response which, 
however, comes on but slowly, as evinced by the gradual with- 
drawal under natural conditions. 

Except for minor changes, the following is quoted directly from 
an enlightening statement made by Mr. Clapp during a corre- 
spondence in which I sought his aid in solving the migratory 
behavior of this animal: 


What you write in regard to the migration of Planorbis is my idea 
also. There is little doubt that Planorbis has a rest period comparable 
to that of nearly all other gasteropods—the effect of inactive phases 
in the life cycle is often clearly marked on the shell. In many species 
of land shells there are two annual rest periods, the shorter occurring 
in the summer, the longer in the winter. I have never observed any 
general inactivity or burrowing during the summer on the part of 
our New England species of Planorbis, and therefore have thought it 
probable that their rest period occurs during the winter months. 

The winter rest period of the land shells in this section of the coun- 
try is governed entirely by the temperature. When spring arrives 
early the shells appear early—the hardier species first, the more deli- 
cate species later—but all appear earlier than in other years when 
the warm weather comes on tardily. It seems probable that Plan- 
orbis, Physa, Lymnaea and similar shells should follow the same prin- 
ciple, in other words, become inactive and bury themselves when the 
temperature of the water falls below a definite point. To a certain 
extent this may be true, but if so, the rest period must be much shorter 
than in land shells, and, I think, not compulsory as with the latter 


364 LESLIE: B. AREY 


forms. My reason for believing this is that I have frequently obtained 
Planorbis, Lymnaea and Physa, in mid-winter, by breaking through 
the ice and using a small dredge or net. 

Some believe that the fresh water pulmonates do not hibernate at 
all. That appears to me to be based on negative evidence, since they 
have not been observed under natural conditions, or rather have been 
observed under unnatural conditions (vivaria). 


At the conclusion of all experiments, to be described subse- 
quently, the animals were beheaded, fixation of the head and 
the contained eyes occurring under conditions of light and tem- 
perature identical with those at which each experiment had 
been conducted. Perenyi’s fixing fluid gave very satisfactory 
preservation. Sections of 8 » thickness were cut parallel to 
the long axis of the eye and were stained with Heidenhain’s 
iron-haematoxylin and with orange-G. 

Whenever the experimentation involved the employment 
of light, strongly diffused daylight, such as is obtainable at 
north windows, was used. 

The microscopical preparations were studied and measure- 
ments were made at a magnification of 1400 diameters. In 
expressing the extent of pigment distribution in numerical 
terms, two measurements were used. One, which I shall call 
the ‘zonal measurement,’ represents the breadth of the band 
of dense pigment, exclusive of the finger-like processes extend- 
ing toward the bases of the constituent cells (figs. A, 4 and 5). 
The second, or ‘process measurement,’ consists of the zonal 
measurement plus the average length of the pigmented proc- 
esses just referred to. 

The sense in which a few descriptive terms will be consistently 
employed needs explanation. ‘Peripheral’ and ‘central’ will 
designate those parts of the eye which are respectively farthest 
from and nearest to its center (fig. B). ‘Distal’ and ‘proximal,’ 
used in describing retinal cells, are retained in a sense similar 
to that employed in descriptions of uninvaginated ectodermal 
cells—hence distal indicates the portion nearest the lens or 
the lumen of the optic sac, while proximal refers to that part 
of a cell nearer the connective-tissue sheath at the periphery 
of the eye (figs. B and C). 


RETINAL PIGMENT OF PLANORBIS 365 


DESCRIPTION OF THE EYE OF PLANORBIS 


Planorbis is a representative of the group of pulmonate gas- 
teropods which is characterized by the possession of but one 
pair of non-retractile tentacles, at the bases of which the eyes 
are located. Although the tentacles are not retractile after 
the manner of an introvert, they nevertheless are capable of 
exhibiting a high degree of muscular shortening whereby their 
surface becomes thrown into transverse ridges (fig. A, ta.). 


=e = bes ieee for.mus. 


n.opt. 


Fig. A An axial section of both the eye and the tentacle of Planorbis, show- 
ing the position and mutual relation of these structures (< 60). fbr. mus., mus- 
cle fibers; n. opt., optic nerve; oc., eye; ta., tentacle. 

The eye (fig. A, oc.), which lies just beneath the surface 
epithelium, has the shape of a cone or pear whose base, cor- 
responding to the corneal portion, points outward. The base 
has a diameter of about 150 u, whereas the axial measurement 
is 200 MK. 

An axial section of the eye (fig. B) shows the following parts: 
(1) optic capsule; (2) cornea; (3) retina; (4) optie nerve; (5) 
lens; (6) vitreous humor. 

The optic capsule (fig. 3, cps. opt.) is a thin, connective- 
tissue sheath surrounding the eye, to which the retinal and cor- 
neal elements are attached. It is continuous with the sheath 
of the optic nerve. 


366 LESLIE B. AREY 


The cornea and retina represent differentiated portions of 
an original sac-like epithelial invagination (figs. A and B). 
The wall of this closed ‘optic sac,’ as it is called, never becomes 
more than one cell thick. The pigment-free corneal cells are 
only slightly columnar in shape, hence this portion of the optic 
sac remains as a thin membrane, which is strikingly in con- 


crn.. 


Fig. B- An axial section (semi-diagrammatic) of an entire eye of Planorbis 
(X 275). ern., cornea; hu. vit., vitreous humor; lns., lens; nl. cl. rin., nucleus 
of retinal cell; n. opt., optic nerve; pig., pigment; rin., retina. 


trast with the thick cornea of some other pulmonates. Its 
extent is limited to the broad base of the conical eye (fig. B). 
The retinal portion of the optic sac is easily distinguishable 
from the cornea by the presence, in the former, of pigmented 
cells and by the great elongation of all its elements (figs. B and 
C). In general, the constituent cells may be said to have a 
radial arrangement with respect to the optic sac. Notwith- 
standing the fact that the retina is only one cell thick, it is con- 


RETINAL PIGMENT OF PLANORBIS 367 


venient to follow the suggestion of Smith (’06) and distinguish 
three concentric zones, which, although of an arbitrary nature, 
are on the whole rather clearly defined. The peripheral zone 
(fig. C, rtn. ex.) is in contact with the optic capsule and is quite 
free from pigment; it contains the cell nuclei. The pigment of 
the non-sensory cells is aggregated in a middle or intermediate 
zone (fig. C, rtn. m.); peripherally its limits are ill defined be- 
cause of irregular processes, which extend down into the cells, 
whereas centrally a sharp line of demarcation separates the 
pigment from a third or central zone (rin. i). The position of 
the boundary common to the peripheral and intermediate zones 
is somewhat dependent on the degree of pigment migration 
under various environmental conditions, which will be dis- 
cussed in the main part of this paper. The internal or central 
zone (fig. C, rin. t.) is of constant width and comprises the por- 
tion of the retina between the pigmented zone and the lens. 
It contains the rods (bac.)—the photoreceptive elements. 

Two kinds of cells form the retinal epithelium, the non-sen- 
sory pigmented cells (fig. C, nl. cl. pig.) and the unpigmented 
sensory cells (nl. cl. sns.). The pigmented cells, which are the 
more numerous of the two, are grouped about the sensory cells 
thereby isolating the latter from each other. These pigmented 
cells are of two kinds. One set has its nuclei situated close 
to the optic capsule, while the nuclei of the other set are near 
the center of the cells and slightly distal to those of the sen- 
sory elements. Both types of pigment cells are exceedingly 
slender, although those with more distally placed nuclei gener- 
ally have an enlargement just distal to the nucleus; according 
to Smith (06, p. 255): ““When this part of the cell is free from 
pigment it appears to contain a vacuole.” 

The distribution of pigment is variable under different cir- 
cumstances; the limit of proximal migration in either type of 
cell, however, is conditioned by the position of the nucleus 
(fig. C), for pigment is never found proximal to it. The length 
of the pigment cells, and therefore the thickness of the retina 
as a whole, varies at different levels in the eye. The thickest 
portion of the retina is on the sides at a distance of about 60 u 


368 LESLIE B. AREY 


from the apex of the optic sac. At this thickest region, the 
distance from the optic capsule to the central zone, that-is, the 
length of the pigment cells, is approximately 35 u. 

The unpigmented or sensory cells (fig. C, nl. cl. sns.) are much 
more robust and have larger nuclei than either kind of pigment 
cell just described. These cells are of an elongated spindle 
shape, thickest in the region of the nucleus. The cell body 


ivin.bace. 
ax.bac:----- : 


\rtn.exs 


3 
o 
o 
o 
' 
! 


Fig. C A portion of an axial section of a Planorbis eye, showing the com- 
ponent retinal elements (< 1000). az. bac., axis of rod; cps. opt., optic capsule; 
fobrl’., fibrillae of rod-axis; fbrl’’., fibrillae of rod-mantle; zvlr. bac., mantle (in- 
volucrum) of rod; nl. cl. pig., nucleus of pigment cell; nl. cl. sns., nucleus of sen- 
sory cell; pig., pigment; pre. n’t., neurite-process of sensory cell; rin. ex., periph- 
eral zone of retina; rin. 7., central zone of retina; rin. m., middle or pigmented 
zone of retina. 


continues through the pigment zone and ends, in the so-called 
central zone, in a structure known as the ‘rod’—this being the 
photoreceptive portion of the cell. Each rod consists of two 
parts, centrally a core or ‘axis’ (az. bac.), and peripherally a 
radially striate “mantle’ (ivlr. bac.). Proximal to the nucleus. 
each sensory cell gives off a neurite (pre. 7’t.), which courses 
along the inner face of the capsule and is ultimately gathered 
up with similar neurites to form the optic nerve (figs. A and B, 


RETINAL PIGMENT OF PLANORBIS 369 


n. opt.). The optic nerve is covered by a sheath of connective 
tissue continuous with that of the optic sac. The work of 
Babuchin (’65), Henchman (’97), and particularly that of 
Smith (’06) has proved that the sensory cells of pulmonate 
gasteropods are of fibrillar composition, the terminal brush- 
like neurofibrils of the rod mantle (fig. C, fbrl.’’) being continuous 
with the fibers of the rod axis (fbrl.’) of the cell body, and of the 
neurite. The length of the rod is 6 uw or 7 u. 

The lens (fig. B, lms.) is a large pear-shaped body occupying 
almost the whole interior of the optic sac. It isa non-cellular 
secreted mass, the outer portion of which forms a denser cap- 
sule or rind. 

The vitreous humor (fig. B, hu. vit.) fills whatever spaces inter- 
vene between the central zone of the retina and the lens, as well 
as the spaces between the individual rods. 


EXPERIMENTAL PART 
a. Effect of light and darkness 


The initial experimentation consisted in determining whether 
or not the retinal pigment of Planorbis undergoes positional 
changes, whereby a characteristic distribution is assumed in 
light and in darkness. 

The citation of Smith’s statement (’06, p. 255) relative to the 
tests performed by him on this animal will serve to show both 
his methods and his results: 


In spite of the fact that the rods, lying, as they do, distal to the 
pigment zone, cannot be protected by pigment migration, as can the 
rods in cephalopods, the variable position of the proximal region of 
the pigment suggests the possibility of pigment migration in the eye 
of Planorbis. I therefore attempted to determine by a few experiments 
whether differing light conditions would produce corresponding changes 
in the position of the pigment. Several specimens of Planorbis were 
placed for an hour or more in water in a white, porcelain dish which 
was set in a sunny window. Their heads were then cut off with scissors 
and fixed in Perenyi’s fluid. Sections were made in the ordinary way 
and stained in Heidenhain’s iron-haematoxylin, followed by orange-G. 
Similar preparations were made from specimens which had been kept 
in darkness for an hour or more before killing. Comparisons were 


370 LESLIE B. AREY 


then made between the two kinds of preparations in order to learn 
whether the position of the pigment was different in the two cases. 
More cases in the ‘light’ eyes had the pigment reaching quite to the 
nuclei than in the ‘dark’ eyes; but there was such a lack of uniformity 
in different animals, and even in the same retina, that the evidence 
was not at all conclusive. Repeated experiments did not lead to more 
definite results. I am satisfied that the pigment does travel up and 
down in the pigment cells under the influence of some stimulus, but 
just what is the exciting factor I have not determined. I have not 
been able to get any evidence of pigment migration in the eyes of either 
Helix or Limax. 


In another part of the same paper Smith again returns to a 
consideration of these results and becomes more bold in his 
inferences. Selected excerpts (pp. 268-269) make plain his 
final attitude toward the situation in Planorbis. 


It is probable that pigment movement is a direct response to light- 
stimulation in Planorbis . . . . Not having been able as yet 
to determine the exact conditions under which it occurs, I can only 
suppose from analogy that the migration is a response to light. The 
shape of the cells, the position of the pigment in some cells as com- 
pared with that in others, and the apparent need of pigment migration 
in the eye of Planorbis, all point to a probable responsiveness of its 
pigment cells to light; seit 


A personal experience (Arey, 16) in determining the lengths 
of time needed to complete light and dark adaption of the highly 
mobile retinal pigment of fishes led me to suspect that the 
‘hour or more’ which Smith allowed for the execution of his 
experiments might be wholly insufficient to permit the photo- 
mechanical response to proceed to completion. Hence in 
my experimentation ample time allowances were made for both 
light and dark adaption. ; 

1. Effect on normal animals. Five animals were placed in a 
battery jar containing water and were kept in a dark-chamber 
for 24 hours. In the same manner, an equal number of animals 
were subjected to daylight for 10 hours. A microscopical 
examination of the resulting preparations gave data as in table 1.! 


1 The final measurements recorded for each eye represent the mean of the 
values obtained from several measurements on each side of the retina at adis- 
tance of about 50 u from the entrance of the optic nerve. Experience showed that 


RETINAL PIGMENT OF PLANORBIS 371 


TABLE 1 


Measurements showing the relative distribution of the retinal pigment of Planorbis 
in darkness and in light at room temperature. The values are mean values ex- 
pressed in micra, and indicate: (1) the thickness of the main pigment mass (zonal 
measurement) ; (2) the zonal measurement plus the length of the pigmented processes 
(process measurement); (3) the length of the pigment processes obtained by sub- 
tracting (1) from (2). The percentage change in each mean measurement in 
darkness as compared with that in light is also computed 


PER- 


PER- PER- 
NUMBER | MEAN | CENTAGE] MEAN | CENTAGE Breall peepee 
CONDITION OF RET- | ZONAL | CHANGE | PRocESS | CHANGE | Gp pig. lin LENGTH 
OF ILLUMINATION INAS _| MEASURE-| IN ZONAL | MEASURE-| IN PROC- | “Uo |/NUnNGt 
MEASURED| MENT |MEASURE-| MENT |ESS MEAS- : 
pa UREMENT (PROCESSES| MENT 
sr PROCESSES 
Bb ad 
OS Le Se 10 11.0 64 5 52 5.5 97 
2 2 
Darkness.......... 10 18.0 25.0 7.0 


Since the distal margin of the pigment zone—the boundary 
between central zone (fig. C, rtn. 7.) and middle zone (rin. m.)— 
is fixed in relation to each pigment cell, these values show quite 
decisively that in the light (fig. 2) the pigment migrates distally 
in the cell, that is, toward the source of illumination, whereas 
in the dark (fig. 5) the reverse is true. An appreciable thinning 
out of the pigment in darkness due to its being spread over a 
greater area, is not demonstrable. It may be that the pigment 
involved in migration is only that which lies most proximally 
in the middle zone, or there may occur a general proximal move- 
ment, that is,a readjustment of the relative quantitative amounts, 
throughout the whole mass. Whichever alternative is true, 
it is certain that the edge of the pigment zone which forms the 
boundary between the middle and central zones (fig. C) main- 
tains a constant position in relation to the retinal cells. 

2. Effect on excised eyes. Since light exercises a photome- - 
chanical influence on the retinal pigment, the interesting query 
arises as to whether this effect is direct, upon the cell itself, 


restricting the measurements to this most favorable region afforded the fairest 
numerical representation of the pigment distribution. In these, as well as all 
experiments to be described subsequently, individuals of Planorbis were selected 
of a more or less uniform size—this precaution will, I think, serve to obviate 
criticism, if not actual error. 


372 LESLIE B. AREY 


or whether it is accomplished through some nervous mechanism 
related to the cerebral ganglion. It is true that the probability 
is against the existence of nervous connections with the pigment 
cells, since all previous researches have failed to demonstrate 
structures of this kind, yet it is entirely conceivable that such 
neuro-fibrils may exist and have hitherto escaped detection; 
indeed, the extreme difficulty with which the neuro-fibrils of 
the sensory cells are demonstrated, makes this possibility all 
the more real. Even if a direct influence of light can be proved, 
the existence of such neuro-fibrils is not precluded, nor yet 
their codperation in effecting the migration of pigment; but it 
does render this event less likely, and definitely shows, more- 
over, that whatever may happen in the normal animal, a direct 
action of light on the pigment cell is a demonstrable phenomenon. 
If the direct action of light can not be shown to induce migra- 
tory movements of the pigment in excised eyes, three possibilities 
exist: (1) fibers in connection with the pigment cell perform 
double conduction, transferring afferent impulses to the cere- 
bral ganglion and efferent impulses back to the retina; (2) affer- 
ent impulses travel in the optic nerve to the cerebral ganglion 
while efferent impulses return by hypothetical nerves to the 
pigment cells, or (3) the optic nerve is of a mixed nature, possess- 
ing both afferent and efferent components. Since, however, 
the existence of double conduction in nerves is a mere postula- 
tion, the first of the three possibilities may be safely eliminated. 

There is, nevertheless, still another way of viewing the situa- 
tion, and one that demands serious consideration. The metabo- 
lism of an organ isolated from the body to which it belongs is, 
of necessity, fundamentally altered. Not only is its supply 
of nutriment cut off, but what is more serious in a short experi- 
ment, the elimination of waste is not adequately provided for. ~ 
If isolated pigment cells do not respond to the direct action 
of light, the reason may easily be ascribable to an autoanaes- 
thetization caused by the accumulation of the cell’s own catabolic 
products. The probability of this course of events was suggested 
in some experiments upon the eyes of fishes (Arey, 16°; 16>). 
Moreover, it will be shown in another part of this paper (p. 378) 


RETINAL PIGMENT OF PLANORBIS 373 


that anaesthetics are probably capable of exerting an inhibition 
on the migration of pigment, even in normal snails. 

Spaeth (713) found the melanophores on the isolated scale 
of Fundulus to be responsive to ultra-violet rays but not to 
those of the visible spectrum. Parker (’97), working upon 
Palaemonetes, obtained responses from all three types of pig- 
ment cells when excised eyes were brought from light into dark- 
ness, or the reverse. This result was observed equally well 
when stalks containing ganglia were used or when merely retinas, 
exclusive of ganglia were employed. Hamburger (’89) reported 
that the pigment of enucleated frog’s eyes exhibits migratory 
movements both in darkness and in light. In another paper 
(Arey, ’16*) I have described the occurrence of a pigment migra- 
tion when the previously dark-adapted excised eye of the com- 
mon horned pout, Ameiurus, is brought into the light; in the 
reverse exposure (light to dark), however, no response ensues; 
furthermore, in several other fishes which were studied, even 
the direct influence of light could not be shown. 

Experiments on Planorbis were made upon animals that had 
been adapted to light or to darkness for 6 and 24 hours, respective- 
ly. Such animals were beheaded and small pieces of tissue, 
bearing a tentacle, were then placed in a watch glass containing 
an abundance of Ringer’s solution and subjected to light or 
to darkness according to desire. The exposure to light lasted 
4 hours, whereas in the dark, eyes were left for 5 hours. Table 
2 summarizes the results of these determinations. 

The results of both these sets agree very closely with the 
values given previously for normal light-adapted eyes (11.0 u 
and 16.5 uw for zonal and process measurement respectively). 
The obvious conclusion to be drawn from these experiments is 
that light has a direct influence on a previously dark-adapted 
pigment cell, but that light-adapted retinas do not change in 
darkness. The direct action of light on the retinal pigment 
of this animal, therefore, is identical with that of Ameiurus. 

I am inclined to interpret these results in the following way. 
The conditions under which an isolated eye is placed undoubtedly 
favor the accumulation of catabolic products. This unremoved 


374 LESLIE B. AREY 


TABLE 2 


Measurements showing the relative distribution of pigment in the excised eyes of 
Planorbis in darkness and in light. The values are mean values expressed in 
micra and indicate: (1) the thickness of the main pigment mass (zonal measure- 
ment); (2) the zonal measurement plus the mean length of the pigmented pro- 
cesses (process measurement); (3) the length of the pigmented processes obtained 
by subtracting (1) from (2). The percentage change in the mean zonal measure- 
ment in darkness as compared with that in light is also computed 


PERCENTAGE 


MEAN MEAN 
NUMBER MEAN ZONAL| CHANGE IN 
CONDITION OF RETINAS MEASURE- ZONAL PROCESS LENGTH OF 
OF ILLUMINATION MEASURED MENT MBEASURE- MEASURE- PIGMENT 
MENT MENT PROCESSES. 
MB M M 
Dark-adapted eyes sub- 
jected to light.......... 6 11.0 14 14 3.0 
Light-adapted eyes sub- > 
jected to darkness...... 6 12:5 14 1.5 


waste exerts an inhibitory or anaesthetic influence upon the 
normal activity of the protoplasm of the pigment cells. Since 
in the case of the retina it is even probable that light favors 
catabolism, it follows that the movement of the retinal pigment 
in the light only, must be due to the greater efficiency of light 
as a stimulating agent. The strong stimulatory effect of light, 
therefore, is able to break through the protoplasmic inhibition 
caused by accumulated wastes, whereas darkness is ineffective 
in this respect. If this reasoning is tenable, one may infer 
that movements of the pigment would also be realized in dark- 
ness, provided artificial circulation could be maintained. That 
light actually is a more efficient stimulus than darkness, is 
further suggested from the determinations of adaption times 
(vide infra). 


(b) Determination of adaption times 


It is evident from the foregoing observations that photo- 
mechanical changes do occur in the pigment of the Planorbis 
retina, and furthermore, one is led to suspect that the indecisive 
results obtained by Smith (06) are assignable to the short 
duration of his experiments, which, in his own words, lasted 
but ‘“‘an hour or more.”’ Hence the pertinent inquiry as to 


RETINAL PIGMENT OF PLANORBIS ay Ais 


the exact periods of time necessary to complete light- and dark- 
adaption becomes the next object of consideration. 

Snails which had previously been thoroughly adapted to 
light (fig. 2) or to darkness (fig. 5), were subjected to opposite 
conditions of illumination. Individuals were removed and 
killed at half-hour intervals until the experiments had con- 
tinued for 5 hours. Microscopical examination of the retinas 
gave the following results: 


Dark-adapted Planorbis subjected to light: 
Incomplete light-adaption occurred in 3 hours 
Complete light-adaption occurred in 4 hours 

Light-adapted Planorbis subjected to darkness: 
Incomplete dark-adaption occurred in 4 hours 
Complete dark-adaption occurred in 5 hours 


Our suspicion as to the cause of Smith’s failure to procure 
satisfactory and consistent results is corroborated, since the 
time alloted for his experimentation was quite inadequate. 

From the comparative standpoint, it is interesting to see how 
the rapidity of pigment migration in Planorbis agrees with 
corresponding movements in the eyes of other animals. 

In the compound eyes of the prawn, Palaemonetes, Parker 
(97) gives the following data for the time consumed in the 
adaption of the proximal, or retinal, pigment: 


WIATENENS UOMO ts sche a ce ciekic co oe aoe ne s Saint seal ane 30 to 45 minutes 
ATAU CUOP OME RTIOSS tec © Sy, e's whatt ofe o.c:alaco eave eas eee ere 45 to 60 minutes 


The writer (Arey, 16°), working upon several fishes, has shown 
that the complete adaption of the retinal pigment requires 
longer periods of time than have generally been supposed. 
These determinations may be combined in the following sum- 
marization : 


IDI SS: GO nT Gib hie as ci fine tee sic on aye a cis ee ee 45 to 60 minutes 
WIIG TOSUATENERSin wc earese nce na os ke tec oe eee ee 30 to 60 minutes 


An interesting comparison with the cephalopod type of eye 
is made possible through the values determined by Hess (’05), 
from whose work it appears that many hours (48 or more) are 


THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL. 26, NO. 4 


376 LESLIE B. AREY 


necessary for dark-adaption, whereas a much shorter time 
(approximately 13 hours) suffices for light-adaption. Hess’s 
results upon Octopus defilippi (one of the several species studied 
by him) are especially astonishing, since he states that dark- 
adaption occurs in 48 hours, whereas light-adaption may be 
practically completed in 15 minutes. 


(c) Effect of temperature 


Next to light, temperature is the commonest environmental 
stimulating agent that controls the movement of retinal pig- 
ment. Accordingly, a series of tests were made, both on nor- 
mal and excised eyes, to discover whether a correlative tempera- 
ture response is demonstrable. 


TABLE 3 


Measurements showing the relative distribution of pigment in the eyes of Planorbis at 
8° and 32°C. both in darkness and in light. The values are mean values expressed 
in micra and indicate both the thickness of the main pigment mass (zonal meas- 
urement), and the zonal measurement plus the length of the pigmented processes 
(process measurement). Corresponding values at room temperature, copied 
from Table 1, are given in parenthesis 


NUMBER MEAN ZONAL MEAN 

CONDITION OF ILLUMINATION me ete TEMPERATURE MEASUREMENT dana 
deg. C. Me iL 

1 bilfea OU Hates oreo MVNO ERD PLACA cytes 6 3 10.0 14.0 
(11.0) (16.5) 

1 Gifts IVF 58g he ia ea ee ae 6 30 25 19.0 

Darkivess puter. cyte sees 4 3 13.0 19.0 
(18.0) (25.0) 

Danknessey ac soe. fee 6 32 14.0 24.0 


1. Effect on normal animals. Fully light- and dark-adapted 
snails, retained in jars of water, were each subjected to tem- 
peratures of approximately 3° and 32° C. At the end of 4 hours 
the eyes were fixed at temperatures identical with those at 
which the respective experiments had been conducted. The 
data from these determinations are presented in table 3. 

These data show that temperature is extremely efficacious 
in evoking positional changes in the retinal pigment. Since, 


RETINAL PIGMENT OF PLANORBIS ah 


both in the light and the dark, the pigment shows greater distal 
migration at low (figs. 1 and 4) than at high temperatures (figs. 
3 and 6), it follows that low temperature and light on the one 
hand, and high temperature and darkness on the other, tend 
to influence pigment migration similarly. As the data show 
(this being a logical corollary of the proposition just stated), 
extreme distal migration is favored by the codperation of light 
and low temperature (fig. 1), whereas an extreme proximal 
migration is best obtained at high temperatures in the dark 
(fig. 6). It is pertinent to mention in this connection that the 
pigmented processes at the higher temperatures in the dark 
(figs. 5 and 6) were very thick as well as long. 

It will be noticed that in the light the values given in paren- 
thesis (representing the results at room temperature copied 
from table 1) are intermediate between the measurements at 
the extreme temperatures; in the dark, on the contrary, this 
progressive relation does not hold. 


TABLE 4 


Measurements showing the relative distribution of pigment in the excised eyes of 
Planorbis at 2° and 30°C. in the light. The values are mean values expressed 
in micra and indicate both the thickness of the main pigment mass (zonal meas- 
urement), and the zonal measurement plus the length of the pigmented processes 
(process measurement) 


NUMBER Maas : MEAN 
CONDITION OF ILLUMINATION OF RETINAS | TEMPERATURE beepers PROCESS 
MEASURED nae, oS ae | MEASUREMENT 
deg. C. u uM 
| Ly (210 ee ace as Wy, See Gn ana 3 2 9.6 10.2 
MOUs he oh cr vere oa, Giants 3 30 10.3 Leo 


2. Effect on excised eyes. A few experiments were performed 
to determine whether the temperature response is obtainable 
in light-adapted eyes which are isolated from the body. In all 
respects the method of experimentation was identical with 
that employed when normal animals were used. 

Table 4 gives the results from the small number of retinas 
measured, an accident having destroyed the remainder of the 
material. 


378 LESLIE B. AREY 


The results thus tabulated are so similar to those just discussed 
in the case of normal animals that comment is scarcely neces- 
sary. The specimens used were rather undersized and _ this 
probably accounts for the low values obtained in each set. 

Unfortunately no material was available when temperature 
experiments upon dark-adapted excised eyes might have been 
performed. Had an influence of temperature been proven 
under these conditions it would have been of considerable in- 
terest, inasmuch as darkness was found (p. 371) to be ineffective 
upon excised eyes which had previously been subjected to light. 
This result was indeed realized in my work (Arey, ’16*) upon 
the isolated eyes of fishes, thereby suggesting that temperature 
is a more efficient stimulating agent than is light. 


(d) Effect of anaesthetics 


In a few instances only, have clear cut observations been 
recorded as to the anaesthetic action of definite substances upon 
retinal pigment cells. Ovio (95) and Lodato (’95) agreed that co- 
caine arrested pigment migration in the frog. The writer (Arey, 
16) has found that carbon dioxide or ether, without perma- 
nently injuring the retinal pigment cells of fishes, is capable 
of arresting completely the movements of the pigmeut in light 
as well as in darkness. Chioretone and urethane, on the con- 
trary, although of sufficient strength to kill the animals as or- 
ganisms, did not prohibit pigment migration. Since 5 per cent 
solutions? of carbon dioxide effectively controlled these migra- 
tions, 16 was suggested that this, the commonest of catabolic 
products, might well be the effective agent in preventing mi- 
eratory movements of the pigment in excised eyes.? It is pos- 
sible that the results, already presented, in which it was found 
that the pigment of the isolated eyes of Planorbis migrated in 
light but not in darkness, may be interpreted in a similar manner. 


2 No attempt was made to ascertain the minimal concentrations which would 
accomplish this end. 

3 Of four fishes used, Ameiurus, Abramis, Carassius, and Fundulus, the retinal 
pigment of the excised eyes of Ameiurus changed its position in the light, where- 
as in none of these species did a change occur in darkness. 


RETINAL PIGMENT OF PLANORBIS 379 


A few experiments seemed to indicate that both ether and 
carbon dioxide did completely check the pigment movements 
when Planorbis was introduced from darsness into light, yet 
the snails were very susceptible to the anaesthetics and in no 
case survived. By careful experimentation it might be possible 
to find concentrations at which movements of the pigment would 
be prevented, yet the pigment cells and the animals (as deter- 
mined by controls) survive. After several trials, in all of which 
the snails did not outlive the 3 to 4 hours subjection to the 
anaesthetics necessary for light adaption, the work was dis- 
continued. 

It is evident that the chief value of the few determinations 
made lies in their suggestiveness and not in the actual results 
gained, which, because of the absence of appropriate controls, 
are properly open to criticism. Hence one may be allowed 
merely to suggest that the pigment in the excised eyes of Planor- 
bis, similarly to that in Ameiurus, is able to overcome the prob- 
able anaesthetic effect of accumulated wastes in the light only, 
whereas the weaker response in the dark fails to appear at all 
because of the presence of the same catabolic products. 


DISCUSSION 


In 1906 Parker analyzed the results of many workers con- 
cerning the influence of light and temperature upon melanophores 
and stated his conclusions in the following generalization (p. 
413): “It is probable that in all melanophores in which there is 
a migration of pigment, light or low temperature will induce 
a migration toward the source of illumination and the absence 
of light or a high temperature a migration in the reverse direc- 
tion.”’ Certain cases, however, may be cited which do not 
conform to Parker’s dictum. Such reversed behavior to light 
is exhibited by the melanophores of the frog (Harless, ’54), the 
eel (Steinach, ’91) and Triton (Hertel, 07). A lack of conform- 
ity is likewise shown in the temperature responses of the frog’s 
retinal pigment (Herzog, ’05; Arey, ’16*). It is not improbable 
that, in these animals, a more or less active nervous control 


380 LESLIE B. AREY 


has been superimposed upon the more primitive direct melano- 
phore response, thereby making the behavior appear anomalous. 

It is evident, however, that the results enumerated in this 
paper relative to the influence of light and temperature upon 
the retinal pigment of Planorbis, are conformable with the 
previously quoted generalization of Parker. The existence 
of a similar agreement in the behavior of the retinal pigment to 
temperature was found by Congdon (’07) upon the prawn, 
Palaemonetes, and by myself (Arey, 716°) upon several fishes. 

The role played by body chromatophores in the economy of 
an animal is supposedly adaptational with respect to its en- 
vironmental coloration, and perhaps the regulation of its body 
temperature as well. The significanace of movements of the 
retinal pigment, on the contrary, are by no means patent, and 
the agreement of its responses to light and temperature with 
those of melanophores in general is of considerable speculative 
interest. Is the reason for this unaminity of response a phylo- 
genetic one, retinal pigment cells retaining a primitive behavior? 
Or is it merely the similar but discontinuous expression of com- 
mon physiological needs? Or is the agreement purely fortuitous, 
not involving the fulfilment of amy common need? Or is the 
influence of light and temperature upon the pigment-containing 
protoplasm necessarily similar wherever mobile pigment. cells 
are found, the pigment migration under any condition (when 
not controlled by the nervous system), therefore, being always 
predictable? By allowing one’s fancy free rein, numerous 
other possibilities may be conceived. 

Since we know many more instances of immobile than of 
mobile pigment cells in the tissues of animals, it is reasonable 
to suppose, and, moreover, evolutionary doctrines compel us to 
assume, that the existence and perpetuation of the latter type 
of cells is not without significance. Thus, ignoring all ques- 
tions as to the reason for this or that behavior of iat cells, 
we at least can safely assume with Parker (’06, p. 411):“" .. 
it might be supposed that if a case arose in which a revere 
migration of pigment would be of service to the organism, such 
form of migration would be evolved and a set of pigment cells 


RETINAL PIGMENT OF PLANORBIS 381 


in which the pigment granules under illumination would migrate 
away from the source of light instead of toward it would be 
produced.” 

It has already been pointed out that the exceptional behav- 
ior of the body chromatophores of the frog, eel, and Triton 
probably is the result of a nervous control. As far as direct 
responses are concerned, Parker’s further assertion (’06, pp. 
411-412), for all we know to the contrary, is true: ‘Hence it 
seems probable that the melanophores, retinal pigment cells, 
and other like structures in which dark pigment granules ex- 
hibit migratory movements, are restricted as to these possibili- 
ties, and that in light they always transport their pigment 
toward the source and never in the reverse direction.” 

Since the pigmented cells of the gasteropod eye have no 
demonstrable nervous connections, this condition, if true, ren- 
ders these cells wholly indifferent in the process of light per- 
ception. It follows, therefore, that such cells are not comparable 
to the pigment-bearing retinular cells of certain arthropods 
in which the pigment granules, contained within the sensory 
cells themselves, change position in light and in darkness. The 
situation in the gasteropod, however, is quite similar to that 
in the vertebrate eye in this respect. 

In several crustaceans (e.g., Parker, 99, upon the amphipod, 
Gammarus) it has been shown that in darkness the pigment 
of the retinular cells moves proximally, thereby leaving part 
of the rhabdome devoid of pigment, whereas in light the rhab- 
dome again receives a pigment sheath. These responses were 
interpreted as having the following significance. In the light, 
the rhabdome, surrounded by pigment, is protected from over- 
stimulation by light reflected internally from the white pigment; 
in dim light, on the contrary, the efficiency of the visual apparatus 
is increased by the withdrawal of pigment, whereby the reflect- 
ing mechanism enables the eye to make the best use of the 
available diffuse light. Theories involving the principles of 
over-stimulation as well as of optical isolation have also found 
many supporters among those who attempt to explain the 
phenomenon of pigment migration in the vertebrate eye (Arey, 
715). 


382 LESLIE B. AREY 


The interpretation of migratory movements in the retinal 
pigment of gasteropods, or to go further, the interpretation of 
the presence of such pigment at all, from the standpoint of 
preventing overstimulation or of securing optical isolation, 
involves greater difficulties than is the case in arthropods. In 
the snail, the brush-like rods, which have been described as the 
photo-receptive portions of the sensory cells, are entirely distal 
to the pigment zone, hence these elements are at all times ex- 
posed to the full strength of light. It has been pointed out, 
however, that the pigment may serve as a background to pre- 
vent reflection and thus it indirectly produces a limited ‘opti- 
cal isolation.’ 

The accumulation of pigment far from the cell bases, and the 
slight extent to which it can be induced to move proximally, 
suggests that its chief utility may exist in a relationship with 
that portion of the sensory cell which it immediately surrounds. 
In this connection, a statement by Smith (06, p. 270) is 
suggestive: 

Aside from preventing internal reflections within the rod zone, 
it is possible that the pigment is directly protective to that part of 
the visual cell which is surrounded by it. We do not know that the 
middle part of the sensory cell is not sensitive to light. Neither do 
we know how or where light vibrations are transformed into nervous 


impulse. If the transformation takes place in the middle zone, the 
pigment may serve some purpose there. 


Having thus seen that the presence of retinal pigment of 
gasteropods is only to be interpreted with difficulty, how much 
greater is the task of assigning explanations to the feeble move- 
ments exhibited by this pigment. It may be said that explana- 
tions of the presence of retinal pigment and of its movements 
are self-inclusive, an explanation of one necessarily involving the 
other. That presence and mobility represent two more or less 
distinct factors, follows, however, from the fact that in some, 
and perhaps in most gasteropods (e.g., in Helix and Limax, 
Smith, ’06) the retinal pigment is non-motile. 

Theorists who have viewed the migratory movements of 
retinal pigment in those eyes where striking changes occur as 
indicative of the prevention of overstimulation or the procural 


RETINAL PIGMENT OF PLANORBIS 383 


of optical isolation, have likewise associated a retreat of the pig- 
ment in dim light with the presence of a more diffuse and weaker 
photic stimulation, but, withal, a stimulation which, under the 
circumstances, thereby rises to its highest efficiency. An assump- 
tion, however, is involved in this reasoning, for it must be shown 
that the pigment really does retreat in dim light, as, in truth, 
it does in darkness. This essential point, however, has too 
often been ignored. 

Furthermore, if the distribution of pigment in twilight were 
essentially similar to that in bright light, it follows that ex- 
planations which attempt to solve the meaning of the position 
assumed by the pigment in light, only answer half of the ques- 
tions involved, for it may well be asked—Why should there 
be an extensive movement in total darkness, or why, since a 
permanently expanded condition would seem to be all that is 
required, any movement at all? These queries merely show that 
easily devised explanations, as for example those based upon 
the theory of optical isolation, may not touch the primary reason 
at all—such obvious relations may have a secondary importance 
or may even be purely fortuitous. 

Returning to the case under consideration, the movements 
of the retinal pigment of Planorbis, through the influence of light, 
are very limited in comparison with those in many animals. 
The dispersion in the compacted pigment mass, and the forma- 
tion of sparse granular processes, do not visibly alter the den- 
sity of the main pigmented zone. It is difficult to see (assuming 
that a proximal migration does occur in dim light as well as in 
darkness) how these changes would be of any great value to 
the animal in the ways in which the pigment commonly has been 
supposed to act. Even the vague assumptions of a nutritional 
relation of the pigment to adjacent sensory cells is discrepant 
with certain phases of the photo- and thermo-mechanical changes. 

In the body chromatophores of certain animals, e.g., lizards, 
the response of the pigment to light and temperature may be 
of use in regulating the body temperature of the animal. Thus 
(quoting from Parker, ’06, p. 411): ‘The dark color of the 
lizard’s skin in moderate illumination at a moderate temperature 


384 LESLIE B. AREY 


insures, possibly, among other things, a certain degree of warmth 
which would be superfluous, if not dangerous, at a higher tem- 
perature, and in consequence the skin becomes light-colored 
in hot sunlight.” 

The significance of temperature responses of the retinal pig- 
ment of Planorbis is even more obscure than the meaning of 
photic influences. It is easy to construct several groundless 
hypotheses to explain why low temperature, acting similar to 
light, and high temperature, acting similar to darkness, could 
be of advantage to the snail in the regulation of its retinal econ- 
omy. Such interpretations, however, possess even less value 
than the speculations concerning the role played by light. 

To devise an explanation from the adaptional standpoint 
that will account for the temperature responses occurring 1n 
total darkness would seem a hard task. It may be, however, 
that the cell thus responds to temperature stimulation because 
the change is of use in the light-adapted eye only, whereas in 
darkness the similar response to similar stimulation is gone 
through with perfunctorily, as it were, even though it is of no 
use to the organism. 

The cautiousness of the following position, with the formula- 
tion of which this discussion will be closed, may be its only com- 
mendation; yet when our absolute ignorance of fundamental 
facts is considered, caution may well qualify as a cardinal virtue. 
An analysis of the conditions of pigment migration in the various 
vertebrate classes has led the writer (Arey, 715; 716) to an iden- 
tical conclusion for these animals as well. 

Recognizing, therefore, that the movements of the retinal 
pigment of Planorbis, as well as of other animals, to light and 
to temperature may have an adaptive significance, and, further- 
more, realizing that (with the possible exception of arthropods) - 
we at present are quite unaware of the meaning of these move- 
ments, it would seem that we are permitted to indulge only 
in interpretations formulated in terms of protoplasmic re- 
sponses to definite stimulating agents and that the questions of 
utility thereby involved must await the establishment of a 
more thorough knowledge of the co-existing factors. 


RETINAL PIGMENT OF PLANORBIS 385 


SUMMARY 


1. Light causes the retinal pigment of normal Planorbis to 
migrate distally (toward the source of illumination). Dark- 
ness causes a proximal migration. 

2. About 4 hours is necessary for the pigment to assume the 
distribution characteristic of light-adaption, whereas about 5 
hours is demanded in accomplishing the reverse process of 
dark-adaption. 

3. The pigment of dark-adapted, excised eyes exhibits mi- 
gratory movements when exposed to light, and is, therefore, 
capable of direct stimulation. In the converse treatment (light 
to dark) of excised eyes, on the contrary, no change in the posi- 
tion of the pigment occurs. It is possible that the absence of 
positional change in darkness is indicative of an anaesthetic 
effect, produced by the unremoved products of catabolism, the 
more vigorous influence of light, however, being able to over- 
come this inhibitory tendency. 

4. High temperature (30°C.) induces proximal migration of 
the normal retinal pigment both in darkness and in light. Low 
temperature (3°C.) induces distal migration both in darkness 
and in light. 

5. Upon light-adapted excised eyes, low temperature likewise 
favors proximal migration and high temperature favors distal 
pigment migration. 

6. Planorbis affords the first case among gasteropods in which 
the occurrence of positional changes in the retinal pigment 
has been correlated with the presence and absence of light. 
For the first time among molluscs, a thermo-mechanical in- 
fluence upon the retinal pigment has been demonstrated. 

7. It is probable that low concentrations of anaesthetics, such 
as carbon dioxide or ether, are capable of completely arresting 
migratory movements of the Planorbis retinal pigment. This 
suggests that carbon dioxide, as a product of catabolism, may 
be the inhibiting factor that prevents migratory movements 
of the pigment when excised eyes are brought from light into 
darkness. 


386 LESLIE B. AREY 


8. The adaptional significance of the existence of mobile 
pigment in the retina of Planorbis is at present quite obscure. 
These movements can only be correlated with the presence of 
known environmental conditions, and interpreted in terms of 
protoplasmic responses to definite stimulating agents. 


Anatomical Laboratory, Northwestern 
University Medical School 
December 6, 1915 


BIBLIOGRAPHY 


(Papers marked with an asterisk have not been accessible in the original 

Arey, L. B. 1915 The occurrence and the significance of photomechanical 
changes in the vertebrate retina—an historical survey. Jour. Comp. 
Neur., vol. 25, no. 6, pp. 535-554. 
19162 The movements in the visual cells and retinal pigment of the 
lower vertebrates. Jour. Comp. Neur., vol. 26, no. 2, pp. 121-201. 
1916 The function of the efferent fibers in the optic nerve of fishes. 
Jour. Comp. Neur., vol. 26, no. 3, pp. 213-245. 

Basucuin, A. 1864 Vergleichend histologische Studien. Wirzburger Natur- 
wiss. Zeits., Bd. 5, pp. 125-143. 
1865 Ueber den Bau der Netzhaut einiger Lungenschnecken. Sitz- 
ungsb. Akad. Wiss. zu Wien., math.-naturw. K1., Bd. 52, Abth. 1, pp. 
16-27. 

Bou, F. 1878 Zusatz zur Mitteil. vom 11. Januar, mitgeteilt in der Sitzung 
vom 19. Februar. Monatsber. preuss. Akad. Wissensch. zu Berlin 
aus dem Jahre 1877, pp. 72-74. 

Cuun, C. 1903 Ueber Leuchtorgane und Augen von Tiefsee-Cephalopoden. 
Verhandl. deutsch. Zool. Gesell., Bd. 13, pp. 67-91. 

Conepon, BE. D. 1907 The effect of temperature on the migration of retinal 
pigment in decapod crustaceans. Jour. Exp. Zodl., vol. 4, no. 4, 
pp. 5389-548. 

Exner, 8S. 1889 Durch Licht bedingte Verschiebungen des Pigmentes im In- 
sectenauge und deren physiologische Bedeutung. Sitzungsb. Akad. 
Wiss. zu Wien., math.-naturw., Kl., Bd. 98, Abth, 3, pp. 143-151. 
1891 Die Physiologie der facettirten Augen von Krebsen und In- 
secten. Deuticke, Leipzig u. Wien., vi+ 206 pp. 

GRADENIGRO, G. 1885 Ueber den Einfluss des Lichtes und der Warme auf die 
Retina des Frosches. Allg. Wiener med. Zeitung, Bd. 30, No. 29 u. 
30, pp. 348-344, u. 353. 

*HamBurGER, D. J. 1889 De Doorsnijding van den nervus opticus bij Kik- 
vorschen, in verband met de Beweging van Pigment en Kegels in het 
Netolies, onder den Invloed van Licht en Duister. Onderzoekingen 
d. Utrechtsche Hoogeschol, Reeks 3, Deel 11, pp. 58-67. 

Harurss, E. 1854 Ueber die Chromatophoren des Frosches. Zeitsch. f. 
wissensch. Zool., Bd. 5, Heft 4, pp. 372-879. 


RETINAL PIGMENT OF PLANORBIS 387 


Hencuman, A. P. 1897 The eyes of Limax maximus. Science, n.s., vol. 5, 
no. 115, pp. 428-429. 

Hensen, V. 1865 Ueber das Auge einiger Cephalopoden. Zeitsch. f. wiss. 
Zool., Bd. 15, Heft 2, pp. 155-242. 

Hertet, E. 1907 Einiges iiber Bedeutung des Pigmentes fiir die physiologische 
Wirkung der Licht-strahlen. Zeitsch. f. allg. Physiol., Bd. 6, Heft 1, 
pp. 44-70. 

Herzoc, H. 1905 Experimentelle Untersuchung zur Physiologie der Beweg- 
ungsvorginge in der Netzhaut. Arch. f. Anat. u. Physiol., Physiol. 
Abth., Jahrg. 1905, Heft 5 u. 6, pp. 413-464. 

Hess, C. 1905 Beitrige zur Physiologie und Anatomie des Cephalopodauges. 
Arch. f. ges. Physiol., Bd. 109, Heft 9 u. 10, pp. 393-439. 

Kitune, W. 1877 Ueber den Sehpurpur. Untersuch. Physiol. Inst. Univ. 
Heidelberg, Bd. 1, pp. 15-104. 

*Lopato, G. 1895 Ricerche sulla fisiologia dello strato neuroepitheliale della 
retina. Arch. di Ottalmologia, vol. 3, pp. 141-148. 

Ovio,G. 1895 Diun speciale azione della cocain sulla funzione visiva. Annali 
di Ottalmologia, Anno 24, Suppl. al Fasc. 4, p. 23. 

Parker, G. H. 1897 Photomechanical changes in the retinal pigment cells 
of Palaemonetes, and their relation to the central nervous system. 
Bull. Mus. Comp. Zoél., Harvard Coll., vol. 30, no. 6, pp. 273-300. 
1899 The photomechanical changes in the retinal pigment of Gam- 
marus. Bull. Mus. Comp. Zodl., Harvard Coll., vol. 35, no. 6, pp. 
141-148. 

1906 The influence of light and heat on the movement of the melan- 
ophore pigment, especially in lizards. Jour. Exp. Zoél., vol. 3, no. 
3, pp. 401-414. 

Rawitz, B. 1891 Zur Physiologie der Cephalopodenretina. Arch. f. Anat. 
u. Physiol., Physiol. Abt., Jahrgang 1891, Heft 5 u. 6, pp. 367-372. 

Smitn, G. 1906 The eyes of certain pulmonate gasteropods, with special 
reference to the neurofibrillae of Limax maximus. Bull. Mus. Comp. 
Zool., Harvard Coll., vol. 48, no. 3, pp. 233-283. 

Spartu, R. A. 1913° The mechanism of the contraction in the melanophores 
of fishes. Anat. Anz., Bd. 44, No. 20-21, pp. 520-524. 

1913 The physiology of the chromotaphores of fishes. Jour. Exp. 
Zool., vol. 15, no. 4, pp. 527-585. 

Sremnacu, E. 1891 Ueber den Farbenwechsel bei niederen Wirbelthieren, 
bedingt durch direkte Wirkung des Lichtes auf die Pigmentzellen. 
Centralbl. f. Physiol., Bd. 5, No. 12, pp. 326-330. 

Szczawinska, W. 1891 Contribution A l'étude des yeux de quelques Crustacés 
et recherches expérimentales sur les mouvements du pigment granu- 
leux et des cellules pigmentaires sous l’influence de la lumiére et l’ob- 
scurite dans les yeux des Crustacés et des Arachnides. Arch. de 
Biol., Tome 10, pp. 523-566. 


38S LESLIE B. AREY 


PLATE 1 
EXPLANATION OF FIGURES 
eps.opt., optic capsule rtn.ex., peripheral zone of retina 
pig., pigment rtn.i., central zone of retina 


The figures of this plate, taken from axial sections of the Planorbis eye, are 
photographs at a uniform magnification of 540 diameters and show the distribu- 
tion of retinal pigment at various temperatures in light and in darkness. 


1 At 3°C. in the light. 

2 At room temperature (21°C. +=) in the light. 
3 At 30°C. in the light. 

4 At 30°C. in the dark. 

At room temperature (21°C. +) in the dark. 
At 32°C. in the dark. 


mS or 


PLATE 1 


RETINAL PIGMENT OF PLANORBIS 
LESLIE B. AREY 


Light Dark 
10) gears 
pig. 
-pig. , 
cps.opt. 
rtn.ex.----- 
3°C. 
| 
4 
pig. pig.- 
rtn ex 
rin.ex. 
cps.opt 
ca2rC 


pig 
pig 
rtn.ex rin.ex 
cps.opt 
ca.3I°C. 


L. B. A. Photo 


389 


ABSENCE OF CHROMATOLYTIC CHANGE IN THE 
CENTRAL NERVOUS SYSTEM OF THE WOODCHUCK 
(MARMOTA MONAX) DURING HIBERNATION 


A. T. RASMUSSEN AND J. A. MYERS 
From the Physiological Laboratory, Medical College, Cornell University, Ithaca, 
New York 


SIX FIGURES (TWO PLATES) 
INTRODUCTION 


Since the discovery by Nissl in 1885 that the chromophilous 
granules in the cytoplasm of nerve cells—granules which had 
already been described by Flemming (’82) and which von Len- 
hossék later termed tigroides, though now more generally spoken 
of as Nissl granules—are intensely stained by basic aniline dyes, 
especially by methylene blue, the disposition of these granules, 
under numerous physiological and pathological conditions, has 
been the subject of much study. From the extensive functional 
alterations that numerous authors (Valentin, Dubois, Merz- 
bacher, ete.) have reported to occur in the nervous system 
during hibernation, and from the morphological variations 
which are said to take place in the nerve cells in certain other 
conditions, such as sleep and starvation, one would naturally 
expect to find marked changes, especially in the Nissl granules, 
during hibernation. This state, as is well known, is attended 
in some animals by almost continuous sleep and profound torpor 
for four months and even longer, during which time no food 
whatever may have been eaten and the body temperature has 
been reduced to but a few degrees above the freezing point. 


HISTORICAL 


An examination of the literature revealed the fact that some 
observers have reported marked structural changes in the nerve 
cells during hibernation. Levi (’98) in the toad (Bufo vulgaris) 

391 


THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL. 26, NO. 4 


392 A. T. RASMUSSEN AND J. A. MYERS 


found that the Nissl bodies had greatly diminished in size and 
the acidophil granules had become basophilic during winter- 
sleep; but in hibernating mammals no changes were observed. 
A more detailed study by Legge (99), however, led this author 
to conclude that the cells of the cerebro-spinal axis of hibernating 
bats do undergo visible changes. In the cells of the cerebral 
cortex he found the Nissl granules to have elongated, being more 
fusiform in shape, and to be displaced towards the periphery, 
forming a sort of envelope to the cytoplasm. Baronecini and 
Beretta (00) also reported morphological changes in the spinal 
cord and especially in the cerebral cortex of mammals (mus- 
cardin and bat) during winter-sleep. They found that the 
Nissl granules had greatly decreased and stained more diffusely. 
The chromatic substance of the nucleus was also more diffuse 
than in active specimens and the nucleolus seemed to have 
disappeared in many cases. Essentially the same changes in 
the Niss! granules in the cells of the cerebro-spinal axis and in 
the Purkinje cells of the cerebellum of hibernating hedgehogs, 
were reported by Marinesco (’05) who found a distinct decrease 
in these granules during the torpid state. Those that remained 
were reduced to fine granules or were diffused in the cytoplasm. 
A more recent study by Zalla (10) agrees with the older results 
obtained by Levi in regard to mammals. Zalla found no ap- 
preciable morphological difference in the Nissl substance in the 
dormouse (Myoxus glis) during hibernation as compared with 
the active state. His results in amphibia were not constant, 
but in reptiles he found a distinct decrease in this substance in 
the motor cells of the cord and pons during winter-sleep. 

The question of changes in the chromophilous substance of 
the nerve cells during hibernation in mammals thus seems to 
be unsettled. Legge, Baroncini and Beretta, and Marinesco 
apparently have observed marked changes, while Levi and Zalla 
could not establish any such chromatolysis. However, these 
authors did not all work on the same species of mammal. In 
amphibia there is also some disagreement. Levi found a dis- 
tinct decrease in the Nissl granules during hibernation, while 
Zalla, whose results were not very constant, believes that there 


CHROMATOLYTIC CHANGES DURING HIBERNATION 393 


are no changes. In reptiles, however, Zalla did find a marked 
decrease during torpor. 

Some other morphological changes in the nerve cells during 
hibernation may be briefly mentioned. Querton (’98) by means 
of the Golgi method on cerebral neurones found that the pro- 
toplasmic prolongations retract and assume a moniliform ap- 
pearance during the dormant period. This, however, could not 
be substantiated by Baroncini and Beretta (’00).. Some work 
has been done on the neurofibrillae. Tello (’03) reported that 
in the motor cells of the spinal cord of lizards the neurofibrillae 
are much thicker than usually found. Tello and Cajal (’04) 
found that this was true only during the cold season when the 
animal is dormant, and that when the animal is warmed up and 
made active the fibrillae become more numerous and much 
finer. Cajal thus believes that the hypertrophy of the neuro- 
fibrillae is due to the cold and resulting diminished spinal reflexes, 
because exposure of the animal to a low temperature brought 
about this gigntism of the neurofibrillae while warming it up 
caused a return to the normal, and because this change is not 
seen in the telencephalon and mesencephalon whose cells retain 
their activity to a much greater extent during the lethargy. 
Marineseco (’05) has repeated these experiments on young cats 
and dogs and Dustin (’06) has done the same on young rabbits 
with essentially similar results. A temperature below 10°C., 
however, was found to be less effective than 10°C. in bringing 
about these changes in the neurofibrillae, according to Marinesco. 
This may be related to the fact that a temperature too low excites 
hibernating mammals and finally wakes them up since the body 
temperature rises as a result of increased activity. The latter 
author found no such modification of the fibrillae in hibernating 
hedgehogs, which fact he interprets as indicating that the activity 
of the nervous system of hibernating mammals is not reduced to 
the extent that it is in the lizard and other cold blooded animals. 
Zalla (10), on the other hand, found that the neurofibrillae were 
fewer and farther apart in the dormouse during hibernation, 


394 A. T. RASMUSSEN AND J. A. MYERS 
PRESENT INVESTIGATION 


In view of the conflicting reports as just reviewed and the 
more recent observations by Crile on changes in the brain cells 
under various emotional and other conditions, much less strik- 
ing than the phenomenon of hibernation, the work we have done 
on the woodechuck in this regard, seems worthy of a brief note. 
Woodchucks, or ground hogs, which represent the American 
marmots, are some of the best examples of hibernating mammals 
in this country. All species remain dormant for four to six 
months each year, and hence constitute good material for a 
study of hibernation. This work was commenced early in 
January 1913 by J. A. Myers, who fixed and imbedded the 
central nervous system of six woodchucks, four of which were 
killed at various intervals while hibernating (January 18, Febru- 
ary 6, March 15 and April). One was killed shortly after waking 
up (Mareh 15) and another, during the following summer. 

In addition to studying the above series by means of the 
Nissl stain, the other co-author prepared another series consist- 
ing of the brain and spinal cord of fifteen woodchucks killed 
during the autumn, winter and spring of 1913-1914. This series 
includes one animal killed about a month before hibernation 
(October 25), one just before hibernation (November 22), five 
during hibernation (February and March)—one of these was 
partly awake when killed on February 16, but was sluggish 
and had a rectal temperature of 19°C.—one within two days 
after waking up (March 16) and seven others which had been 
awake from three days to more than a month. Three of this 
last group had been fed for one, two and three weeks respectively. 
These animals were kept in the artificial burrows which were 
designed by Professor Simpson of this laboratory and which 
have already been described elsewhere.! The rectal temper- 
ature of the dormant animals varied from 8°C. to 12°C., whereas, 
the temperature of the active animals ranged from 32°C. to 
30 ©. 


1 Rasmussen, A. T., Amer, Jour. Physiol., 1915, vol. 39, p. 20. 


CHROMATOLYTIC CHANGES DURING HIBERNATION 395 


All the animals of both series were killed quickly by trans- 
fixing the heart through the chest wall. No anaesthetic was 
used, except in five cases where only sufficient ether was given to 
keep the animal quiet. These five animals were all killed after 
hibernation while awake and active. The amount of ether 
given did not seem to have any noticeable effect on the Nissl 
granules. If, however, these five cases are excluded from con- 
sideration because of the introduction of this additional factor, 
the two series involve as strictly comparable cases two before 
hibernation, nine during hibernation and five after hibernation. 
The blood was washed out immediately after death with normal 
saline solution by injection through the aorta. The saline was 
followed by a saturated aqueous solution of bichloride of mer- 
cury to which had been added 10 per cent of formalin. Thus 
the central nervous system was fixed very quickly in situ. The 
whole brain and cord were then removed and cut into transverse 
sections a few millimeters thick. The desired levels were further 
fixed in a saturated aqueous solution of bichloride of mercury for 
48 hours, washed in running water 36 hours, and dehydrated in 
graded alcohols containing iodine in the usual manner. The 
tissue was cleared in xylol and imbedded in paraffin melting at 
54°C. The levels thus imbedded were: olfactory bulb, motor 
cortex, mid-thalamus, midbrain at the level of the superior 
colliculus, mid-cerebellum and pons, medulla oblongata at ex- 
treme inferior border of fourth ventricle, first cervical, sixth 
cervical, sixth thoracic, second lumbar and lower lumbar seg- 
ments of the spinal cord. 

Sections were cut five microns in thickness, except in the case 
of the spinal cord where the sections were six microns thick, 
and stained on the slide for ten minutes in a large quantity of hot 
(70°C.) solution of 1 per cent methylene blue in water saturated 
with aniline oil. The excess stain was washed off rapidly in 
water and decolorization carried on by transferring the slides 
directly to 95 per cent alcohol for two to ten minutes. Dehy- 
dration was completed in absolute alcohol and clearing in caju- 
put oil followed by xylol. The sections were mounted perma- 
nently in Canada balsam dissolved in xylol. Where the size 


396 A. T. RASMUSSEN AND J. A. MYERS 


permitted, corresponding sections from all animals of a series 
were fixed on the same slide to insure equal staining. In other 
vases several sections from the same block were placed on one 
slide and all the corresponding slides stained together by carry- 
ing them through the reagents by means of a basket, or rack, 
which would contain the entire lot. 


RESULTS 


The nerve cells of the woodchuck are essentially typical, 
containing the usual large round nucleus, with one nucleolus. 
The chromophilous substance in the cytoplasm has the usual 
appearance and arrangement so well known that no description 
will be necessary here. In spite of all precautions there are notice- 
able variations in the size, distinctness and arrangement of the 
Nissl bodies in homologous cells of animals in the same state and 
even in the cells of the same group in a particular section. These 
variations are found in both dormant and active animals. We 
can detect no modification in the Nissl granules characteristic 
of the hibernating as compared with the non-hibernating state. 
Certainly in these woodchucks there is not the difference indi- 
cated by the figures given by Marinesco in the case of the hedge- 
hog. The chromophilous substance is present during hibernation 
in at least as great a quantity as at other times and presents the 
usual appearance when stained with methylene blue. The 
arrangement of the granules varies somewhat even in cells of the 
same group, being more abundant in the periphery of the cells in 
some cases and in others being grouped more densely around 
the nucleus. The size and shape of the bodies vary from fine 
irregular granules to larger elongated ones; but when a large 
number of cells are examined the extreme variations may be 
found in animals in the same state and often in the same section. 
A predominance of a particular variation in either state can not 
be established. The accompanying figures and explanations will 
suffice to indicate the general cell picture before, during and 
after hibernation. The larger types of nerve cells were selected 
as illustrations because in them the Nissl bodies are more distinct 
and make better photographs. 


CHROMATOLYTIC CHANGES DURING HIBERNATION 397 


We wish to thank Prof. Sutherland Simpson for his helpful 
guidance in this research. We are also indebted to Prof. B. F. 
Kingsbury for assistance in taking the photomicrographs.* 


LITERATURE CITED 


BaRoncini, L., AND Beretra, A. 1900 Ricerche istologiche sulle modif. 
degli organi nei mammiferi ibernanti. Riforma med., vol. 16, p. 206. 

CasaAL, RamM6N y 1904 Variaciones morphologicas, normales et pathologicas 
del reticulo neurofibrillar. Trabajos del lab. de invest. biolog. de la 
Univ. de Madrid, vol. 3, p. 9. Compt. Rend. Soe. Biol., vol. 56, p. 372. 

Dustin, A. P. 1906 Contribution 4 l’étude de l’influence de l’Age et de |’ activ- 
ité fonctionnelle sur le neurone. Ann. Soe. Roy. d. Se. méd. et 
natur. de Bruxelles, vol. 15. 

Lecce, F. 1899 Sulle variazioni della fina struttura che presantano durante 
V’hibernazione le cellule cerebrali dei pipistrelli. Monit. zool., vol. 10, 
p. 152. 

Levi, G. 1898 Sulle modificazioni morfologiche delle cellule nervose di animali 
a sangue freddo durante l’hibernazione. Riv. d. patol. nervos. e. 
med., p. 443. 

Marinesco, G. 1905 La sensibilité de la cellule nerveuse aux variations de 
température. Revue Neurolog., No. 14 (July 30, 1905), p. 784. Re- 
cherches sur les changements de structure que les variations de tem- 
p¢érature impriment 4 lacellulenerveuse. Revista Stiintelor Medicale, 
No. 3, Bucarest. 

1906 Recherches sur les changements des neurofibrilles consécutifs 
aux différents troubles de nutrition. Le Névraxe, vol. 8, p. 149. 

QuERTON 1898 Lesommeil hibernal et les modifications des neurones cérébraux. 
Ann. Roy. d. Se. med. et natur. de Bruxelles, vol. 7, p. 147. 

TrELLo 1903 Sombre la existencia de neurofibrillas gigantes en la medula espinal 
de los reptiles. Trabajos del lab. de invest. biolog. de la Univ. de 
Madrid, vol. 2, p. 223. 

Zauua, M. 1910 Recherches expérimentales sur les modifications morpholo- 
giques des cellules nerveuses chez les animaux hibernants. (Résumé). 
Arch. ital. de biol., vol. 54, p. 116. Riv. d. patol. nervos. e. ment., ann. 
15, p. 211. 


2In regard to the general problem of the correlation of structural changes to 
activity in nerve cells, attention should be called to the experiments carried 
out on six different species of animals by R. A. Kocher, reported in this journal 
(vol. 26, No. 3, p. 341) since this article was submitted for publication. This 
author could find no correspondence between the size or structural character- 
istics of nerve cells and various grades of activity. 


PLATE 1 
EXPLANATION OF FIGURES 


The three photographs in this plate were taken from the nucleus hypoglossus. 
The sections from the various animals were mounted on the same slide and hence 
stained together. > 320. 

1 Before hibernation. Woodchuck killed October 25, 1918, without any 
anaesthetic. Animal active. Rectal temperature 37.6°C. 

2 During hibernation. Woodchuck killed March 7, 1914, without any anaes- 
thetic. Animal very dormant and had been so nearly all the time for at least 
three months. Rectal temperature 9°C. 

3 After hibernation. Woodchuck killed April 11, 1914, without any anaes- 
thetic. Animal active. Had been fed for two weeks. Rectal temperature 37°C. 


398 


CHROMATOLYTIC CHANGES DURING HIBERNATION PLATE 1 
A. T. RASMUSSEN AND J. A. MYERS 


se 
7 | yp 
, 
s 
a 
> > 
ell a . 2 ss 
... 
> ’ 
* 
‘ : 
ps . 
1 CP. 
e . - => 
< 
’ 
sd i 


' M v. 
i i 
og \ 

at ’ Y ; 
ae 2 
\ : ti 


399 


RGAE, 22 


EXPLANATION OF FIGURES 


Photographs of motor cells from the antero-lateral group in the sixth cervical 
segment of the spinal cord. The sections were mounted on the same slide and 
hence stained together. > 320. 

4 Before hibernation. Same animal as in figure 1. 

5 During hibernation. Same animal as in figure 2. 

6 After hibernation. Same animal as in figure 3. 


400 


CHROMATOLYTIC CHANGES DURING HIBERNATION PLATE 2 
A. T. RASMUSSEN AND J. A. MYERS 


401 


A STUDY OF A PLAINS INDIAN BRAIN 


J. J. KEEGAN 


From the Anatomical Department, University of Nebraska, Omaha, Nebraska 
EIGHT FIGURES 
INTRODUCTION 


The primary object of this paper is to furnish an account of 
the morphology of the cerebral hemispheres of a North American 
Indian brain, thereby filling a gap in the literature of racial 
cerebral anatomy. Although a large amount of work has been 
done upon the Indian tribes from an anthropological stand- 
point, no description of an Indian brain has, to my knowledge, 
ever been given. The study of cerebral morphology has not 
yet furnished characters distinctive of race, but the studies of 
Elliot Smith (04), Duekworth (’07), and others have revived 
an interest in this work and advanced greatly the interpretation 
of sulei and gyri. 

The secondary but more important aim of this study,from 
which the conclusions are largely deduced, is the application of 
the principles of fissure relation to cortical areas, both in the in- 
terpretation of sulci and in the comparison of cortical areas and 
sulci of the two hemispheres and of other cerebra. This is 
especially valuable in this case on account of the completeness 
of the history and the exceptional mental abilities evidenced. 

The study of cortical localization has been conducted by two 
methods, cortical stimulation and histological differentiation. 
The most important names connected with the former method 
are those of Sherrington and Griinbaum (’01) in their experi- 
mental work upon the ape brain, supplemented by a few similar 
observations upon the human cortex and numerous clinical ob- 
servations proved by operation or autopsy. This work neces- 
sarily gave results in only a few regions of the cortex, mainly 

403 


404 J. J. KEEGAN 


the somatic motor and sensory areas of the central region. The 
histological field was first entered by Flechsig (’96) who de- 
limited cortical areas by the order of the myelination of their 
fibers. Campbell (05) and Brodmann (’07) made purely his- 
tological surveys of the cortex, but were not able by this tedious 
method to examine enough material to establish the relation of 
many cortical areas to fissures. KE. Smith (07) was the first to 
make such a topographical survey of the human cortex and, 
while in general the areas plotted agreed with the areas of 
Campbell and Brodmann, the added value of their interpretation 
in relation to fissures facilitated greatly the study of the fissures 
of the cerebrum. 

The value of E. Smith’s interpretation is evidenced by its 
general acceptance for the occipital region by most of the recent 
writers, as Duckworth (07), Appleton (10), Cole (11), Schuster 
(08), and many others. The application of this method to other 
regions, however, has been neglected, chiefly on account of the 
difficulty in obtaining fresh material and the inexpertness in 
distinguishing the lamination in macroscopic sections. The 
boundaries of the area striata are so easily identified that this 
area can be plotted even in imperfectly preserved material, 
which accounts for the general application in this region. 

In this study an attempt has been made to apply by compari- 
son the cortical area plan of E. Smith (14) to a valuable cerebrum 
in which the preservation prevented a knife-section analysis of 
the cortical areas and their relation to sulci. The accuracy of 
this method may well be questioned for undoubtedly many 
small errors are present, but it was found that such a plotting 
aided greatly in the interpretation of fissures and perhaps made 
possible a more intelligent comparison of similar cortical areas 
of the two hemispheres and of other cerebra. It at least furnishes 
the most expressive figures of the sulci and gyri that present 
knowledge makes possible and might be applied profitably for 
comparison to the figures accompanying studies of other brains, 


AN AMERICAN INDIAN BRAIN 405 


HISTORY 


The brain which is the subject of this account was obtained at 
autopsy from the body of a female Indian of the Omaha tribe, 
and thanks are due Dr. A. A. Johnson of the Department of 
Pathology for the opportunity of the use of the specimen for 
study. 

The individual was fifty years of age, about 125 pounds in 
weight and possessed rather typical Indian features although 
not extreme. There was united in her person the blood of the 
Indian, the French, and the English settlers of Nebraska, her 
grandfathers carrying the half French and the half English 
admixture and the grandmothers of pure Indian descent. It is 
thus seen that the Indian blood was carried down entirely on 
the female side of the family, which might give rise to specula- 
tion upon the formerly much discussed question of the relation 
of sex to cerebral morphology. 

The intellectual status was far above the average as super- 
ficially judged of Indian people, and above the average of the 
white race as judged from the high quality of her attainments. 
The educational training consisted of mission school and gov- 
ernment Indian school as a child, five years of general education 
in preparatory schools, and a three year medical course which 
she completed in two years, graduating at the head of her class. 
A year’s hospital training completed the medical work. Her 
life was spent in medical service with her own people. She was 
generally recognized by the profession and laymen alike as ¢ 
woman of very superior personality whose intellectual endow- 
ments and qualities of character placed her’ quite above the 
ordinary level of humanity. 


DESCRIPTION 


The brain, after standing several weeks in a formaldehyde 
solution, weighed 1353 grams with the dura mater removed. A 
comparison of the weights of the two hemispheres showed the 
right considerably heavier than the left, 605 grams and 575 grams, 


406 J. J. KEEGAN 


respectively, but these figures are unreliable on account of the 
pathological condition of the left hemisphere, which had de- 
stroyed a part of the temporal lobe. This hemisphere also was 
distorted in preservation, due to the softened condition of the 
cerebral tissue, consequently no measurements of any value of 
lengths or indices could be made. Such measurements were 
established by Cunningham (’92) but have not proven of com- 
parative value even in perfectly preserved brains. 

The general type of fissuration presented nothing very unusual, 
giving the impression of a fairly well and evenly fissured cere- 
brum. There was no undue prominence of any region, perhaps 
a greater tendency to a vertical course of the sulci of the central 
region of the right hemisphere and a tendency to irregular 
fissuration in the posterior parietal region of the left hemisphere. 
There was no indication of an exposure of the insula. 

The following brief account of the sulci and gyri is not intended 
as a description of all points observed, for many of these of no 
recognized significance or variation can be determined equally 
as well by an examination of the accompanying figures. These 
illustrations were made by tracing from photographs and subse- 
quently inking in the fissures by careful examination of course, 
depth and bridging gyri. The depth of the more important 
fissures is entered in the drawing. In general the heavier lines 
indicate the deeper fissures. A bridging gyrus is indicated by 
an interruption of the line and an intervening dot. The cortical 
areas are filled in with different symbols, the plan being to place 
no two of similar type adjoining. 

These areas were determined by placing those first in which 
there could be little doubt of their boundaries. The area striata 
was delimited by macroscopic examination of knife sections. The 
remaining areas generally could be filled in by a process of elimi- 
nation, but in some cases they were determined by arbitrary judg- 
ment. This latter is very evident in cases where the boundary 
does not correspond to a fissure, but it is surprising how little is 
left to the judgment and how easily the area conformation can 
be made to agree with the plan of E. Smith without violating 
the rule of the deeper fissures being in the main bounding fis- 


AN AMERICAN INDIAN BRAIN 407 


sures. With this scheme it is hardly necessary to label the dif- 
ferent fissures, for a comparison with E. Smith’s figures in Cun- 
ningham’s Text-book of Anatomy will show the interpretation. 

The fissura cerebri lateralis is sharply upturned and _ bifur- 
cated in the left hemisphere at its posterior extremity, but hori- 
zontal and single in the right. Since this bifurcation is caused 
by an overgrowth of the posterior inferior parietal region, as has 
been determined in the Negro brain by Poynter and Keegan 
(15), the two hemispheres are peculiarly contrasted in their 
centers of greatest growth in the inferior parietal lobule, the left 
being in the posterior region and the right in the anterior region. 
This is further evidenced in the conformation of other sulci in 
these areas, and will be discussed at a later point. In both 
hemispheres there is a distinct operculation by the superior lip 
in the region of the area subcentralis. The anterior rami of 
the fissure are completely separate on both sides, enclosing a 
prominent pars triangularis. 

The sulcus centralis is unusual in its upper third in the left 
hemisphere. It lacks several millimeters of reaching the mesial 
border and in the superior third is displaced anteriorly in an 
arcuate form, apparently due to an overgrowth of the gyrus 
centralis posterior opposite the motor area of the lower limb. 
According to Cole (’10), interruption of the suleus by a sub- 
merged or bridging gyrus occurs at the posterior arch of the 
superior genu in poorly convoluted brains and just above in 
well developed brains. This brain shows the gyrus exactly at 
the apex of the superior genu in both hemispheres, narrower and 
more prominent in the left. An examination of a large number 
of Negro and Caucasian brains failed to corroborate Cole’s 
conclusion. The gyrus is always at the same point, varying only 
in depth and width. The maximum depth of the sulcus cen- 
tralis is about the same as the sulcus interparietalis. The 
greater depth of the latter is stated by Appleton (710) to be a 
distinctive feature of Australian and simian brains. It is a 
question whether this would not simply indicate an overgrowth 
of the parietal lobe in a human brain and not an undergrowth 
of any region. 


THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL. 26, NO. 4 


408 J. J. KEEGAN 


The sulei precentrales are typical of Cunningham’s (’92) de- 
scription in the right hemisphere, but in the left hemisphere the 
superficial interrupting gyrus occurs lower, apparently between 
the two rami of the sulcus precentralis inferior. The sulcus 
diagonalis is incorporated in the ramus verticalis in both hemi- 
spheres, thus reducing the posterior area of the gyrus frontalis 
inferior to a small area in the floor of the sulcus precentralis 
inferior. The factor which would account for this might be an 
overgrowth of the motor area of the mouth region in the gyrus 
precentralis inferior or of the motor speech center in the gyrus 
frontalis inferior. Cole (10) has attempted to relate the sub- 
mergence of this area or gyrus to a simian condition and quotes 
Kohlbriigge and Bolk as supporting his contention that the 
sulcus diagonalis represents a detached portion of the simian 
sulcus precentralis inferior. 

The sulcus frontalis superior is well defined in both hemispheres 
and superficially continuous from the s. precentralis superior to 
the s. fronto-marginalis, maintaining about an even distance 
from the mesial border throughout. A prominent bridging gyrus 
near the posterior extremity permits the interpretation of the 
continuity of the posterior middle frontal region with the ante- 
rior superior frontal region as illustrated in E. Smith’s chart. 
Appleton (10) considers the tendency towards extra segmenta- 
tion of the s. frontalis superior as a lowly character. 

The sulcus fronto-marginalis is better defined in the left hemi- 
sphere than in the right but is too irregular in both to permit of 
much comparison. 

The sulcus frontalis inferior is a high arcuate fissure on both 
sides in full communication with the s. precentralis inferior. Its 
length and its distance from the fissura cerebri lateralis are each 
about a centimeter greater in the right hemisphere. This gives 
rise to a larger gyrus frontalis inferior on the right side which 
would be contrary to expectation if the speech center had any 
relation to the size of this gyrus. 

The sulcus frontalis medius is a very subordinate system of 
shallow sulci, according to the interpretation of the s. frontalis 
superior. The greater width of the gyrus frontalis medius of 


AN AMERICAN INDIAN BRAIN 409 


the left hemisphere does not seem to increase the regularity of 
these elements. The greater tendency is towards a transverse 
direction. 

The sulcus frontalis mesialis is seen as an irregular system of 
alternating transverse and longitudinal elements extending to the 
frontal pole. The more posterior of these were chosen arbi- 
trarily as a guide between the superior frontal area and the 
anterior frontal area. 

The orbital surface has the usual type of fissuration, the only 
unusual feature being the extension of the mesial limb of the 
s. orbitalis transversus into the fossa Sylvii of both hemispheres. 

The sulcus cinguli corresponding to E. Smith’s chart is very 
poorly developed in its anterior two-thirds, the more prominent 
sulcus being the s. paracinguli. This might indicate a disap- 
pearance of the s. cinguli due to an increased growth of the mesial 
region, but was called to attention by Cole (10) in a micro- 
cephalic brain in which he suggested that the supracingulate 
suleus was the older. 

The sulcus postcentralis is a very prominent fissure in both 
hemispheres, extending from the mesial border to within a few 
millimeters of the fissura cerebri lateralis. In the right hemi- 
sphere there is no communication with the s. interparietalis, 
thus giving rise to a fissure very similar to the s. centralis. In 
the left hemisphere the posterior reflection of the two extremi- 
ties forms an arcuate sulcus and a wide gyrus centralis posterior 
in the corresponding regions. 

The sulcus interparietalis is contrasted on the two sides. In 
the right hemisphere it appears as a fourth vertical sulcus in series 
with the precentral, central and postcentral sulci. It communi- 
cates across an almost superficial bridging gyrus with the s. 
paroccipitalis. The tendency in the left hemisphere is towards 
a sagittal and more posterior arrangement. The sulcus is repre- 
sented by two horizontal elements in intercommunication with 
the s. postcentralis and the body of the s. paroccipitalis across 
bridging gyri. 

The suleus paroccipitalis is more lateral and more prominent 
in the right hemisphere. The independence of this sulcus is 


410 J. J. KEEGAN 


well represented in both hemispheres. Appleton (’10) has classi- 
fied this as a fetal tendency. The narrow antero-posterior 
width of the gyrus arcuatus posterior is a noticeable feature and 
might be indicative of an unusual growth in the superior parietal 
area. 

The suleus parietalis superior is poorly represented, perhaps 
due to its partial incorporation in surrounding sulci. 

The lobulus parietalis inferior gives evidence of a prominent 
growth. This is more noticeable in the left hemisphere where 
the prominent sulcus interparietalis extends into the anterior 
region as an accessory fissure. The s. angularis, which to some 
extent is an index to the growth of the inferior part of this lobule, 
is not independent from the s. temporalis superior as found in 
the majority of Negro brains. It appears as the sharply upturned 
extremity of this fissure, more anterior in the left hemisphere. 

The sulcus temporalis superior of the right hemisphere was 
injured by the tumor growth. The absence of the anterior trans- 
verse element may be associated with an increased growth of the 
acoustic area of the superior temporal gyrus. This was further 
evidenced by very prominent transverse temporal gyri of Heschl 
and the extension to the lateral surface of the fissure separating 
the two larger of these gyri. 

The sulcus temporalis medius and the sulcus temporalis 
inferior have never been interpreted well enough to permit mor- 
phological comparison. The ascending ramus of the former, 
suleus occipitalis anterior of some authors, is typical in the right 
hemisphere of an arcuate communication with the s. occipitalis 
inferior, classified by Appleton (10) as a simian character. 

The fissura rhinalis is present in both hemispheres as a shallow 
groove connecting the Sylvian fossa with the s. collateralis. 
The interpretation of these two fissures, applied as in the Negro 
brain, indicates an earlier or less developed condition in the 
right hemisphere but in neither is the simian or fetal type ap- 
proached, regardless of the greater depth than usual. 

The sulcus collateralis is superficially continuous to the occipi- 
tal region in both hemispheres. The posterior portion lies con- 
siderably more lateral in the right, which, as a limiting sulcus 
for the area peristriata, would indicate a greater extent of this 


AN AMERICAN INDIAN BRAIN All 


area on this side. The rhinal element of the sulcus is separated 
in both hemispheres by a bridging gyrus near the anterior 
extremity. 

The sulcus lunatus can be very plainly identified in the left 
hemisphere. It is 35 millimeters in length and lies nearer the 
lateral border about 3 centimeters from the occipital pole. It 
has an arcuate or rather angular form and a slightly operculated 
posterior lip. The area striata, delimited by knife sections, 
_ lacked about 2 millimeters of reaching this lip but followed the 
course of the fissure quite regularly. In the right hemisphere the 
tendency is towards a longitudinal disposition of the fissures in 
this region, to which the boundary of the area striata does not 
bear any definite relation. The most prominent sulcus is inter- 
preted as a modified s. prelunatus or s. occipitalis lateralis of 
some authors. Comparison of the extent of-the area striata in 
the two hemispheres shows very plainly a greater lateral extent 
in the left hemisphere. 

The sulcus occipitalis paramesialis lies more upon the lateral 
surface in the left hemisphere but is very prominent in both 
hemispheres. This lateral position indicates that it bears a 
relation to the lateral extension of the area striata. 

The sulcus occipitalis inferior courses on the tentorial surface 
along the lateral border. It is very similar on the two sides, a 
deep, slightly operculated fissure about 40 millimeters in length, 
communicating at the two extremities with the sulci of the 
corresponding regions. 

The sulcus calearinus in both hemispheres is separated from 
the fossa parieto-occipitalis by a small bridging gyrus cuneus 
and from the sulcus retrocalearinus by a larger bridging gyrus. 

The sulcus retrocalearinus terminates in both hemispheres in a 
deep operculated polar element but separated by a prominent 
bridging gyrus. Landau (’15) interprets this polar element (s. 
extremus, s. occipitalis triradiatus, s. calearinus externus of E. 
Smith, s. occipitalis polaris), as the true posterior bifurcation as 
described by Cunningham (’92), separated by the gyrus cuneo- 
lingualis posterior and homologous to the sulcus triradiatus of 
ape brains. The prominence of this suleus was noted by the 
writer (’15) in the Negro brain, in which it was described as an 


412 J. J. KEEGAN 


independent suleus occipitalis polaris. If the interpretation of 
Landau is correct, its prominence would be an indication of a 
greater proportionate development of the area striata in the 
same manner as the sulcus lunatus. 

The fossa parieto-occipitalis is quite typical of E. Smith’s 
description in the left hemisphere. The gyrus intercuneatus is 
about 5 mm. in height and separates widely the s. paracalearinus 
and the s. limitans praecunel. It is traversed by an independent 
sulcus incisura which appears superficially on the lateral surface, . 
separate from the termination of the sulcus limitans praecunei 
in the gyrus arcuatus posterior. The s. paracalearinus under- 
mines the posterior wall of the fossa, becoming superficial along 
the border of the hemisphere posterior to the gyrus arcuatus. 
The fossa of the right hemisphere is less typical. The gyrus 
intercuneatus is faintly indicated, the deepest part of the incisura 
being formed by the sulcus paracalearinus. The s. limitans 
praecunei incises the lobulus praecuenus. 

The: sulci limitantes areae striatae are easily identified by the 
delimitation of the area striata. 

The insula, as far as could be determined, presented nothing 
unusual in its fissuration. The operculation was complete in 
both hemispheres. 


SUMMARY 


The detailed study of the fissures and convolutions brings a 
number of points to attention. First is noted the striking differ- 
ence between the two hemispheres in almost every fissure or 
region examined. This variation is so great that a comparison 
of hemispheres is of very little value in the interpretation of 
sulci, which method proved so valuable in the interpretation in 
the Negro brain. A similar fact of asymmetry is stated by 
Appleton (10) to represent. an agreement with European cere- 
bra in general and a contrast with supposedly lower types of 
cerebra. Many writers have called attention to the asymmetry 
of the cranium, brain, and intracranial venous sinuses sepa- 
rately, but comparatively little work has yet been done to study 
the co-relation of these asymmetrical conditions the one to: the 
other or to explain their origin. E. Smith (’07), in a note upon 


AN AMERICAN INDIAN BRAIN 413 


the asymmetry of the brain and skull, contrasts the asymmetry 
of the cranium of the higher races of man with the symmetry of 
the apes and to a less degree of the black races. This lack of 
symmetry of the cranium in man is attributed to the unequal 
development of homologous parts of the two cerebral hemi- 
spheres, especially the great parietal and frontal association areas. 
The greater size of the right parietal association area is given 
as an explanation of the greater prominence of the correspond- 
ing parietal bone and a relative shifting backward of the right 
parietal boss, of the usually greater extent of the lateral part 
of the left visual cortex and the retention of a more pithecoid 
form in the left hemisphere. 

This asymmetry is suggestive of the predominance of functional 
areas upon one side or the other of the brain. This has been 
generally accepted in the predominance of the speech center 
upon the left side in the inferior frontal gyrus, Broca’s convolu- 
tion, and of the visual cortex in the more extensive area striata 
and more prominent sulcus lunatus in the left hemisphere of the 
majority of brains. Cunningham (’02) attributed right-handed- 
ness to a functional pre-eminence of the left hemisphere, al- 
though not supported by any constant observation of greater 
weight or greater prominence of the so-called motor arm center, 
of appreciable histological difference or of any regular plan of 
asymmetry. The asymmetry, although noticeable also in the 
lower animals, ‘‘never attains the same degree as in man.” <A 
deficient growth of the ascending parietal convolution (gyrus 
post-centralis) has been found associated with the congenital 
absence of the arm of the opposite side (Moorhead, ’02). The 
fact that this is not in the so-called motor center of the arm is 
interesting in that it shows that the functional center does not 
necessarily correspond to the motor center. 

The importance of this aspect of cerebral morphology would 
tend to complicate the comparison of sulei and gyri, not only 
between the two hemispheres of the one cerebrum but also be- 
tween hemispheres of different cerebra, of individuals or of 
races. It would be necessary, before a rational comparison could 
be made, to establish as many points as possible of functional 
similarity. At the present time there are few such points for 


414 J. J. KEEGAN 


comparison, but it is not unreasonable to expect that a closer 
clinical or functional analysis may disclose more features that 
can be translated as indications of the predominance of functions 
upon one side or the other of the brain. An interesting observa- 
tion would be to determine if a predominant vision with the 
right eye is related to a more prominent area striata and sulcus 
lunatus of the left hemisphere. 

The most striking differences between the two hemispheres of 
this brain are found in the occipital, parietal and central regions. 
The suleus lunatus can be identified only in the left hemisphere 
where it is very prominent and associated with a more lateral 
extent of the area striata, a larger, more lateral and more inde- 
pendent sulcus occipitalis triradiatus, a sulcus occipitalis para- 
mesialis more upon the lateral surface, a sulcus occipitalis in- 
ferior nearer the lateral surface and in arcuate communication 
with the sulcus occipitalis anterior. 

The inferior parietal area is more extensive in the right hemi- 
sphere and the increased growth appears to be in the anterior 
portion. This is evidenced by the extension of the anterior ex- 
tremity of the sulcus interparietalis into this region and the 
more posteriorly situated sulcus angularis. The conformation 
of the sulcus interparietalis of the left hemisphere would seem 
to indicate a predominance of growth in the superior parietal 
lobule. This has resulted in a low position of the sulcus inter- 
parietalis, a posterior deflection of the sulcus post-centralis and 
a striking anterior arching of the sulcus centralis above its su- 
perior genu. The lower part of the gyrus post-centralis is also 
widened by the posterior deflection of the inferior extremity of 
the sulcus post-centralis. The entire gyrus is noticeably wider 
than in the right hemisphere. 

The value of these observations of the dissimilarity of the two 
hemispheres is difficult to judge on account of the lack of knowl- 
edge of the extent of unilateral functional predominance and of 
clinical tests to corroborate the morphological findings. The 
study is interesting from an anthropological standpoint, dem- 
onstrating that in a race of inferior status all of the elements 
necessary for a higher individual development are present. 
The same condition was concluded in the Negro brain (Poynter 


AN AMERICAN INDIAN BRAIN 415 


and Keegan, 715). No single morphological point could be 
selected which would represent inferiority. While in the Negro 
the mental characteristics may be in part explained by the 
great predominance of the parietal lobe over the frontal lobe, 
in this brain the characteristic feature is not a disproportionate 
growth of any large area, nor any striking complexity or sim- 
plicity of fissuration, but a marked asymmetry of the fissures 
and convolutions. 
CONCLUSIONS 


1. This Indian brain represents in practically all features a 
high type of cerebrum, the only possible exceptions being the 
shallow communication of the incisura rhinalis with the sulcus 
collateralis and the presence of a typical sulcus lunatus in the 
left hemisphere. 

2. The great asymmetry of the two hemispheres in fissuration 
is the most convincing evidence of a highly specialized cerebrum. 
Future combination of physiological observation with cerebral 
morphology will undoubtedly lead to a better interpretation of 
such variations. 

3. The comparison of cerebra of different individuals should 
take into consideration this relation of asymmetry to functional 
localization and should be confined to hemispheres of the same 
side in individuals of as near similar mental traits as possible. 

4. The method of plotting the cortical areas established by 
E. Smith, although inaccurate in many details, aids greatly in 
the interpretation of fissures and in the comparison of the de- 
velopment of different areas. It serves as the most expressive 
manner in which the morphology of cerebra can be presented 
for comparison. 

LITERATURE CITED 

AppLETON, A. B. 1911 Descriptions of two brains of natives of India. Jour. 
Anat. and Physiol., vol. 45. 

_Bropmann, K. 1907 Beitrige zur histologischen Lokalization der Grosshirn- 
rinde. Journ. fiir Psychol. und Neurol., Bd. 6. 

CaMPBELL, A. W. 1905 Histological studies on the localization of cerebral 
function. Cambridge. 

Cote, 8S. J. 1910 On some morphological aspects of microcephalic idiocy. 
Jour. Anat. and Physiol., vol. 44. 


1911 Remarks on some points in the fissuration of the cerebrum. 
Jour. Anat. and Physiol., vol. 46. 


416 J. 


CUNNINGHAM, D. J. 1892 
hemispheres. 
Dublin. 
1902 


vol. 32. 


Duckworth, W. L. H. 1907 


Right-handedness and left-brainedness. 


J. KEEGAN 


Contribution to the surface anatomy of the cerebral 
Cunningham Memoirs of the Roy. Irish Acad., no. 7, 


Jour. Anthropol. Inst., 


The brains of aboriginal natives of Australia. 


Jour. Anat. and Physiol., vol. 42. 


Fiecusi@¢, P. 
Lanpavu, E. 1915 

Folia Neuro-Biol., Bd. 9. 
Mooruweap, T. G. 1902 


1896 Gehirn und Seele. 


Leipzig. 
Zur vergleichenden Anatomie des Hinterhauptlappens. 


A study of the cerebral cortex in a case of congenital 


absence of the left upper limb. Jour. Anat. and Physiol., vol. 37. 


Poynter, C. W. M. anp KeEEa@an, J. J. 
Jour. Comp. Neur., vol. 25. 
Descriptions of three Chinese brains. 


brain. 
Scuuster, E. H. J. 1908 
and Physiol., vol. 42. 


SHERRINGTON, C. S. AND GriNBAuM, O. 


1915 <A study of the American Negro 
Jour. Anat. 


1901 Observations on the physiology 


of the cerebral cortex in some of the higher apes. Proc. Roy. Soc., 


vo. 69. 
SuitH, G. E. 
ence to the Egyptians. 
vol. 2. : 


1907 Asymmetry of brain and skull. 


1904 Studies in the morphology of the brain, with special refer- 
Recs. Egyptian Gov’t. School of Med., Cairo, 


Jour. Anat. and Physiol., vol. 41. 


1907 A new topographical survey of the human cerebral cortex. 
Jour. Anat. and Physiol., vol. 41. 
1914 The nervous system. Cunningham’s Text-book of Anatomy, 


fourth edition. 


ABBREVIATIONS 


A, Sulcus angularis 

C, Sulcus centralis 

CA, Sulcus calearinus 

CL, Fissura cerebri lateralis 

CO, Sulcus collateralis 

D, Suleus diagonalis 

FT, Sulcus frontalis inferior 

FM, Sulcus frontalis medius 

FMA, Sulcus fronto-marginalis 

FS, Suleus frontalis superior 

IP, Sulcus interparietalis 

L, Sulcus lunatus 

OA, Sulcus occipitalis anterior 

OI, Sulcus occipitalis inferior 

OL, Sulcus olfactorius 

OP, Sulcus occipitalis paramesialis 

OR, Sulcus orbitalis transversus 

P, Suleus occipitalis polaris (retro- 
calearine bifurcation ) 

PC, Sulcus postcentralis 

PCT, Sulcus paracinguli 


PL, Sulcus prelunatus 

PO, Fossa parieto-oce pita 1s 

POC Sulcus paroccipitalis 

PRI, Sulcus precentralis inferior 

PRS, Sulcus precentralis superior 

R, Suleus radiatus 

RA, Ramus ascendens of fissura cerebri 
lateralis 

RC, Suleus retrocalearinus 

RH, Ramus _horizontalis 
cerebri lateralis 

RI, Fissura rhinalis 

RO, Sulcus rostralis 

SSA, Sulcus subcentralis anterior 

SC, Sulcus cinguli 

SP, Sulcus subcentralis posterior 

TA, Sulcus temporalis anterior 

TI, Suleus temporalis inferior 

TM, Sulcus temporalis medius 

TS, Sulcus temporalis superior 


of fissura 


BRAIN 417 


AN AMERICAN INDIA 


+ 
if. HETtHE + 
Oot > * 


Fig. 1 Right hemisphere, lateral view 
Fig. 2 Left hemisphere, lateral view 


EGAN 


J. KE 


J. 


418 


oo 
So 
o=o 


° 
C) 
X69, 
o%o 
2264 
Chreby 
95° 
bis ° 
a 20 
99 fe) 
©\9 5 y Og 
(-) 0195/09 
@° 6° eke) 
NG on 
WP OS 
00 Wie) 
aod 
So\2 
2090 |” 
09 
Ey 
0% 
6 09 
19 99 


° 


iew 
1 view 


© 


mesial v 


ht hemisphere, 
here 


1g 


3 


Fig. 


1a 


mes 


d 


isp 


Fig. 4 Left hem 


AN AMERICAN INDIAN BRAIN 419 


Fig. 5 Both hemispheres, basal view 
Fig. 6 Both hemispheres, superior view 


SS 985 SS 
$tPerO? 90.99 
+t aeeser ec? : 


ete 


GAN 


4 


- 
4 


KEI 


J. 


J. 


420 


ital view 


, occip 


heres 
Smith’s (’14) figure 


isp 
of E 


hem 


7 Both 
8 Copy 


1g. 
ig. 


F 
F 


AN EXPERIMENTAL STUDY OF THE VAGUS 
NERVE 


MARTIN R. CHASE 


From the Anatomical Laboratory of the Northwestern University Medical School' 


FOUR FIGURES 


In a previous paper a study was made of the structure of the 
roots, trunk and branches of the vagus nerve (Chase and Ran- 
son 714). 

It was shown, in conformity with the work of Molhant (’10) 
and Van Gehuchten and Molhant (’11) that the vagus nerve in 
the dog contains myelinated fibers which can be classified as 
large, medium and small. In addition there were found enor- 
mous.numbers of unmyelinated fibers. These are present in 
certain of the vagus rootlets, and throughout the trunk of the 
vagus. In the cervical and thoracic trunk they constantly 
increase in proportion to the myelinated fibers, so that at the 
level of the diaphragm the vagus is almost a pure unmyelinated 
nerve. 

The most of the large myelinated fibers, and many of medium 
size, are given off in the cervical branches of the vagus. The 
bronchial and esophageal rami receive nearly all of the remain- 
ing myelinated fibers. Relatively few unmyelinated fibers are 
found in the cervical branches. The increased proportion of 
unmyelinated fibers in the lower vagus is due in part at least 
to the withdrawal from the trunk of the myelinated fibers by 
the upper branches. Essentially the same histological picture 
was seen in sections from the vagus of man (Ranson ’14), the 
rat and the rabbit. 

Gaskell ’86 observed unmyelinated fibers in the vagus nerve, 
and found the lower thoracic vagus to contain few myelinated 
fibers. He interpreted the unmyelinated fibers in the vagus 


1Contribution No. 37. 
421 


422 MARTIN R. CHASE 


as being post-ganglionic fibers arising from sympathetic cells 
in the vagus ganglia. 

Molhant 710 working with the Cajal reduced silver method, 
saw masses of unmyelinated fibers in the vagus nerve and thought 
they: were of sympathetic origin, being presumably post-gan- 
glionic fibers arising from cells in the sympathetic ganglia. Nei- 
ther Gaskell nor Molhant demonstrated unmyelinated fibers in 
the roots of the vagus. 

Langley has suggested the possibility of preganglionic vis- 
ceral efferent fibers losing their myelin sheaths during their 
course down the vagus. 

Ranson (’15) has recently studied the structure of the vagus 
nerve in the turtle, and has made observations of importance 
in clearing up the origin of the unmyelinated fibers in the vagus 
nerve. The vagus nerve in the turtle divides high in the neck 
into a cervical and a thoraco-abdominal ramus. The cervical 
ramus is composed almost entirely of myelinated fibers and the 
cells of the cervical ganglion of the vagus are associated: only 
with the fibers of this ramus. 

The thoraco-abdominal ramus presents the same structure 
as the thoracic vagus in mammals. It has a ganglion in the 
upper part of the thoraco-abdominal cavity. It consists largely 
of unmyelinated fibers, with scattered myelinated fibers, and 
maintains the same structure from its origin to the origin of 
the recurrent nerve. Ranson shows that the unmyelinated 
fibers are not post-ganglionic fibers arising from cells in the cer- 
vical ganglion, since the thoraco-abdominal trunk is not asso- 
ciated with this ganglion. Neither do preganglionic visceral 
efferent fibers lose their myelin sheaths, since the structure of 
the thoraco-abdominal ramus does not change. There is no 
association with sympathetic ganglia, and the proportion of 
unmyelinated fibers does not change below the thoraco-abdomi- 
nal ganglion, so there are probably no sympathetic elements in 
this ganglion. 

In the dog there is a very close association of the vagus and 
sympathetic trunks in the neck, and it seemed possible that some 
of the unmyelinated fibers of the thoracic vagus in this animal 


THE VAGUS NERVE 423 


might be of sympathetic origin. The present paper presents the 
results of some experiments performed on dogs to determine 
by degeneration methods to what extent sympathetic fibers 
entered into the composition of the vagus. 

Figure 1 is a diagram of the right vagus nerve in the dog, 
showing the relations with the sympathetic, and the origin of 
the branches. Note that the two trunks are closely united 
from high in the neck to the upper thoracic region. Note also 
that just below the bronchial rami, 7.b., each vagus gives off a 
branch which goes to unite with the nerve of the opposite side. 
In addition each nerve is also joined by smaller twigs from the 
contra-lateral vagus. Sections of either vagus below the level 
of the bronchial rami may include fibers which originate in 
the opposite vagus, a fact of importance in interpreting the 
histology of the degenerated nerves. 

It would be possible for post-ganglionic fibers to enter the vagus 
from communications with the superior cervical ganglion of 
the sympathetic. In the preparation of the former paper (Chase 
and Ranson) we were unable, in a study of serial sections, to 
trace any considerable number of such fibers into the vagus. 
Throughout their course in the neck, although the two nerves 
are contained in a common sheath, no interchange of fibers 
eould be traced. At the level of the inferior cervieal (@.c.7.) 
ganglion of the sympathetic, however, the association of the 
vagus and sympathetic is very close, and numerous large bun- 
dles of sympathetic unmyelinated fibers can be seen in serial 
sections entering and leaving the vagus trunk. If, then, any 
considerable proportion of the unmyelinated fibers in the thoracic 
vagus of the dog have their origin in the sympathetic trunk, 
they must arise from cells in the inferior cervical ganglion g.c.7., 
or ganglion stellatum, g.s. 


METHODS 


The right vagus nerve was severed in a number of dogs. In 
some cases it was separated from the sympathetic trunk high 
in the neck and about an inch of the vagus nerve was removed 
(fig. 1, A). There would then be no interference with fibers 


THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL. 26, NO. 4 


MARTIN R. CHASE 


Stherine Hill, 


THE VAGUS NERVE 425 


from the superior cervical ganglion, the sympathetic trunk or 
inferior cervical ganglion. 

In other cases about an inch of combined vagus and sym- 
pathetic trunk was removed in the midcervical region (fig. 1, B). 
Such a procedure would cut off any fibers to the vagus from the 
superior cervical ganglion, but would in no way interfere with 
fibers from the inferior ganglion. 

The dogs were autopsied after 2 to 8 weeks and portions of 
the nerves studied after preparation by the pyridine silver 
method (Ranson 711, 712) and the osmie acid method. 

CHANGES AFTER SECTION OF THE HIGH IN 
THE NECK 


VAGUS ALONE 


Sections from the right vagus of a dog autopsied five weeks 
after removal of a portion of the upper cervical vagus, the sym- 
pathetic being left intact, show definite degenerative changes. 
A eross section stained by the pyridine silver method, taken 


Fig. 1 Diagram of the right vagus nerve in the dog 
Fig. 2. From a section of right vagus nerve at level indicated (fig. 1, 2, 3), 


following removal of a portion of the vagus and sympathetic trunks at level B, 


figure 1. Pyridine silver. X 1140. 

Fig. 3 From a section of the same nerve at same level as figure 2. Osmic 
acid. xX 1140. 

Fig. 4a From a section of a normal vagus nerve just below pulmonary plex- 
uses (fig. 1, 4a). Pyridine silver. X 1140. 

Fig. 4b From a section of the same nerve as figure 2, below the pulmonary 
plexuses (fig. 1, 4b). Pyridine silver. X 1140. 


a, vagus rootlets 

A,B., indicate levels at which portions 
of the vagus trunk were removed 

art.s., arteria subclavia 

a.s., ansa subclavia 

b, bulbar rootlets of the accessory nerve 

c, spinal root of the accessory nerve 

g.c.v., ganglion cervicale inferior 

g.c.s., ganglion cervicale superius 

g.j-, ganglion jugulare 

g-n., ganglion nodosum 

g.s., ganglion stellatum 

l.v., left vagus 


n.c.i., hervus caroticus internus 

n.l.s., nervus laryngeus superior 

n.r., hervus recurrens 

n.v., hervus vagus 

n.v.t.s., nervus vagus and truncus sym- 
pathicus 

p.o., plexus esophageus 

r.b., rami bronchiales 

r.e., ramus externus n. accessoril 

r.p., ramus pharyngeus 

i.s., trunkus sympathicus 

2, 3, 4a, 4b, indicate levels at which the 
corresponding figures were taken 


426 MARTIN R. CHASE 


through the thoracic vagus just above the bronchial rami, 
shows no normal myelinated fibers and only a few scattered 
unmyelinated axones. There are present, however, a few bun- 
dles of unmyelinated fibers which are clearly of sympathetic 
origin. They cling in groups and stain in a characteristic man- 
ner. Similar groups of unmyelinated fibers, present at this level 
in the normal vagus nerve, were readily identified as of sym- 
pathetic origin by their contrast with the vagus fibers. They 
occupy only a small part of the total cross section area. 

Sections taken below the pulmonary plexuses show the right 
vagus to have retained its degenerated character. It is joined 
early by twigs from the normal left vagus, whose deeply stained 
fibers contrast sharply with the degenerated areas. It is im- 
possible to locate in these sections the sympathetic fibers seen 
at a higher level. They are not present in bundles, and it 
would be impossible to differentiate individual sympathetic 
fibers from normal fibers from the left vagus. 


CHANGES PRODUCED BY REMOVAL OF A PORTION OF THE 
CERVICAL VAGO-SYMPATHETIC TRUNK 


In the animal from which the illustrations are taken, a por- 
tion of the right sympathetic vagus trunk was removed as in- 
dicated (fig. 1, B) and the dog was allowed to live four weeks. 

Figure 2 is a drawing from a cross section of the right vagus 
trunk just proximal to the bronchial rami (fig. 1, 2, 3) stained 
by the pyridine silver method. There are seen no normal myel- 
inated or unmyelinated nerve fibers. The aggregations of 
granules, surrounded by definite zones, are nuclei of neurilemma 
cells, as is proven by longitudinal section. The entire cross 
section shows the same structure, there being only an occasional 
nakedaxone. Thereareno bundles of unmyelinated sympathetic 
fibers as was seen at the same level in the specimen previously 
described. 

Figure 3 is a drawing of a small part of a cross section of the 
same nerve at about the same level, stained with osmic acid. 
Note the presence of a number of deeply stained globules, which 
are the remains of degenerated myelin sheaths. There is pres- 


THE VAGUS NERVE 427 


ent in the entire cross section only an occasional myelin ring, 
and most of these clearly represent degenerating fibers. 

Figure 4 is a composite drawing. Figure 4a, the upper half 
of the figure, is taken from a cross section of a normal right 
vagus nerve just below the origin of the bronchial rami (fig. 1, 
4a). There are present in this field nine small myelinated axones. 
The remainder of the field is packed with unmyelinated axones. 
Figure 4a may be taken to represent the condition in the normal 
thoracic vagus, it being remembered that above the bronchial 
rami there are more myelinated axones, while near the dia- 
phragm there are even fewer than in the figure. Figure 4b 
is a drawing from a cross section of the same nerve as figures 
2 and 3, taken at a level just proximal to the union with the 
branch from the left vagus nerve (fig. 1, 4b). There are in the 
field no normal unmyelinated or myelinated axones, and the 
entire cross section shows the same structure, there being only 
an occasional naked axone. 

In both pyridine silver and osmic acid preparations of the 
nerve below the level described sections, both of the right nerve 
after its union with the branch from the left vagus, and of the 
left nerve after union with the branch from the right, show a 
mingling of the degenerated nerve with the normal, but the 
degeneration can be clearly traced to the diaphragm. 

In the second case, where both the vagus and sympathetic 
were cut in the mid-cervical region, no bundles of fibers of sym- 
pathetic origin could be found, the entire nerve being degenerated. 

We conclude, then, that following section of the cervical 
vagus nerve in the dog, with or without section of the sympathetic 
trunk, all fibers of the vagus, myelinated and unmyelinated 
alike, undergo complete degeneration. There are to be found 
in the degenerated thoracic trunk, above the level of the pul- 
monary plexuses, in some individuals bundles of unmyelinated 
nerve fibers having their origin in the cells of the sympathetic, 
presumably in the inferior cervical ganglion. In some individuals 
no bundles of this sort are found. In no ease does the total 
number of fibers of sympathetic origin present amount to more 
than an insignificant fraction of the total number of unmyelinated 


428 MARTIN R. CHASE 


fibers normally present. In the material studied it was possible 
to demonstrate more groups of unmyelinated fibers of sympa- 
thetic origin in the thoracic vagus in the animals in which only 
the vagus was severed high in the neck, than in those operated 
by division of both vagus and sympathetic trunks. It is thought 
that this finding is the result of individual variation in the ani- 
mals, but it is possible that the difference in operative procedure 
has something to do with it. 

The unmyelinated character of the thoracic vagus nerve in 
the dog is not due to the presence of fibers derived from the 
sympathetic trunk, during the close association of the vagus 
and sympathetic nerves in the neck. 


BIBLIOGRAPHY 


CHASE AND Ranson 1914 The structure of the roots, trunk and branches of 
the vagus nerve. Jour. Comp. Neur., vol., 24, no. 1, pp. 31. 

GaAsKELL, W. H. 1886 On the structure, distribution and function of the nerves 
which innervate the visceral and vascular systems. Jour. Phys., 
London, vol. 7, p. 19. 
1889 On the relation between the structure, function and origin of 
the cranial nerves. Jour. Phys., London, vol. 10, p. 153. 

Lanecuery, J. N. 1900 The sympathetic and other related systems of nerves. 
Schifer’s text-book of Physiology, vol. 2, p. 665. 
1903 Die Kranialen autonomen Nerven und ihre Ganglien. Ergeb. 
der Phy she Bde pone: 

Motuant, M. 1910 Le nerf vague. Le Névraxe, vol. 11, p. 187. 

Ranson, 8. W. 1911 Non-medullated fibers in the spinal nerves. Am. Jour. 
ANN, WO, WAY jo), Me 
1912 The structure of the spinal ganglia and of the spinal nerves. 
Jour. Comp. Neur., vol. 22, p. 159. 
1914 The structure of the vagus nerve of man as demonstrated by a 
differential axon stain. Anat. Anz., Bd. 46, S. 522. 
1915 The vagus nerve of the snapping turtle (Chelydra serpentina). 
Jour. Comp. Neur., vol. 25, p. 301. 


CHANGES IN THE ROD-VISUAL CELLS OF THE FROG 
DUE TO THE ACTION OF LIGHT 


LESLIE B. AREY 


From the Anatomical Laboratory of the Northwes ern University Medical School! 


TWO FIGURES 


CONTENTS 

ee EU STANA LSE Wis 5 gil get ak aia oe VF o's) Gk a Coe. <e © tags +a aR ee 429 
Pe PR eA POLITOT ELDON. y-5i5-p so 3/5 5's Wiss = She Aid's im vw ble w ajave’'p lw Bee ke RG ha ie an 434 
A. Influence of light on the position of the red rod nucleus........... 435 

B. Influence of light on the length of the rod myoid. Relative changes 
St center. ANd peripbery of retina, . .....\.<.02 0s .pcememeeek eee es bs 436 
BLE EUG OCS gf eo oeniecd 5 bu-a 91 5:40y o siGie's al wok Sisko 8516-7 pisiy a SRE eC 436 
Ey NNO oy si kas on +558, d' v's. dsy wee © ts ee 439 
C. Influence of light on the diameter of the red rod................... 440 
Cre at eas Oe, ee nr eRe hes oo WN ae 441 
Pm NET EMUGEM E hercisce Wateis norik ocala ano 8 cvetle tnie.a'aceieis nséie'e-b ate hic Fo Neo a Ce Sea 442 


PRELIMINARY 


Whenever the results of experimentation upon any particular 
animal, or group of animals, are at variance with a well estab- 
lished canon, it is desirable that such results be subjected to the 
closest scrutiny possible before the existence of a deviation 
from the generally accepted norm is admitted. The present 
paper, which attempts to answer the following query, is concerned 
chiefly with the reéxamination of a case of this kind—Query: 
Are the photomechanical responses of the visual rods of the frog 
similar to those occurring in the retinas of other investigated verte- 
brates? j 

Both the rods and the cones of many vertebrates undergo 
positional changes when subjected to photic or thermal stimu- 
lation (Arey 715, 716), 


1 Contribution No. 388, April 20, 1916. 
29 


430 LESLIE B. AREY 


Wherever movements of the cones have been detected, light 
causes a shortening, and darkness an elongation, of the contrac- 
tile portion of the inner member known as the myoid (figs. 1 
and 2, my.con.). 

Stort (’86), working on the crow, first showed that the rod myoid 
of birds exhibits photomechanical changes which are the exact 
reverse of those found in the cone—that is, elongation in the 
light and retraction in the dark. The following year (’87) 
he extended his observations upon certain fishes. Although 
this striking response in fishes can be demonstrated with com- 
parative ease (especially in fishes possessing large rods; Arey, 715, 
16), it was not until long after that these forgotten observations 
were corroborated and their correctness admitted. 

With respect to the frog’s rod,? however, Angelucci (84, ’90) 
and Gradenigro (’85) came to a different conclusion. These 
workers recorded that the myoid of the inner member (figs. 1 
and 2, my.bac.rb.) of the frog’s rod shortens when exposed to 
light and elongates in darkness. Gradenigro (p. 343) states 
this conviction most emphatically: ‘“Zuerst habe ich mit posi- 
tiver Bestimmtheit ersehen dass an der Froschretina unter der 
Einwirkung des Lichtes die Stabcheninnenglieder kiirzer und 
dicker werden, in dhnlicher, wenn auch nicht so ausgedehnter 
und rascher Weise, wie Englemann bei den Zapfeninnengliedern 
die Beobachtung gemacht hat.”’ 

Angelucci (’90, p. 245) makes a similar straightforward as- 
sertion, which applies to both the red and green visual rod: 


2'There are two kinds of visual rods in the frog’s retina. The more numer- 
ous form (figs. 1 and 2) has a short inner member (my.bac.rb. + ell.bac.rb.) 
and a long outer member (prs.dst.bac.rb.). Since the first imperfect observa- 
tion by Miller (’51), it has been shown repeatedly that, in the fresh retina of 
an animal which has previously been retained in darkness, the outer member 
of the rod appears reddish due to the presence of unbleached visual purple (rho- 
dopsin). Schwalbe (’74) saw a second type of rod (figs. 1 and 2) characterized 
by an elongated inner member (my.bac vr. + ell.bac vr.) and an abbreviated 
outer member (prs.dst.bac.vr.); Boll (’77) also described this element and in- 
terpreted correctly its green appearance, in fresh dark-adapted retinas, as due 
to a specific photo-sensitive material, which, from its color has been called visual 
green (chloanopsin). Hence red and green rods differ both anatomically and 
as to the nature of their photo-sensitive contents. Unless otherwise stated 
reference to the red visual rod is usually understood. 


CHANGES IN THE VISUAL CELLS OF THE FROG 431 


“Die Stiébchen sind dick und sowohl in ihrem 4usseren Glied 
als in ihrem myoiden Theil zusammengezogen. . . . Auch 
die griinen Schwalbe’schen Stibchen sind sowohl in ihrem dus- 
seren als in ihrem inneren Gliede zusammengezogen.”’ 

Arcoleo (’90), using pithed toads, likewise reported similar 
responses. 

More recently, Lederer (’08), in a brief communication, has 
challenged the results of the previously named workers on the 
frog’s rod. From the study of fixed material, which had been 
subjected to teasing, he concludes (p. 764): ‘‘ Die Hellstabchen 
waren im allgemeinen linger, schlanker und hatten gleich breites 
Innen- und Aussenglied, wiihrend bei den Dunkelstibchen, die 
kiirzer und plumper erschienen, das Innenglied dort, wo es an 
das Aussenglied grenzt dicker wird.” It should be remarked 
in passing that Lederer’s two schematic figures of isolated frog’s 
rods have presumably been interchanged—at least, it is obvious 
that they illustrate conditions exactly the reverse of those which 
his text descriptions maintain. His experience with stained cel- 
loidin sections is summarized in the following statement (p. 764): 


. man nach den Zupfpriparaten am gefiirbten Schnitte 
ebenfalls dhnliche Verhiiltnisse hiitte erwarten sollen. Indessen waren 
hier die Veriinderungen der Stibchenschicht sehr wenig markant. 
In ungefihr der Hialfte der geschnittenen Licht-Bulbi zeigten die 
Stibchen liingere, gestrechtere Form, grésseren Abstand ihres Aussen- 
gliedes von der Membrana limitans externa, die Dunkelstibecben 
gedrungenere gestalt, kleinere Distanz von der ausseren Grenzmem- 
bran. In der anderen Halfte der Schnitte aber war eine Verinderung 
der Stellung der Stibchen sehr wenig ausgesprochen, und die Hell- 
und Dukelstiibchen das Aussenglied ungefihr gleich weit von der 
Membrana limitans externa entfernt. 


The rods of certain amphibians exhibit photomechanical 
movements which should be clearly distinguished from those 
produced by the contractilty of the myoid. Thus Stort (’87) 
first asserted, that, in the dark, the nuclei of the rods in Triton 
(figs. 1 and 2, st.nl.ex.) migrate partially through the external 
limiting membrane, thereby causing the whole rod to become 
extended, whereas, in the light, these nuclei lie wholly within 
the outer nuclear layer. The contractility of that portion of the 


432 LESLIE B. AREY 


rod-visual cell between the rod nucleus and the external reticular 
layer was believed to cause these changes. Angelucei (90) 
made similar observations on the salamander, as did Garten 
(07) on Triton. 

Changes in the cylindrical outer member (figs. 1 and 2, prs.dst. 
bac.rb.) have also been reported. Ewald und Kithne (’78) first 
observed a swelling of the outer member of the frog’s rod as the 
result of strong illumination. That Lederer (’08) obtained 


> 


Dy 


prs. dst. bac. vr. _ 


. | ----prs, dst. bac. rb, 
prs. dst. con,----f 


mb, lim. exc” (fo: SEF: 


Fig. 1 From a dark-adapted retina of the frog, Rana pipiens, showing the 
positions assumed by the rods and cones. X 1130. ell.bac.rb., ellipsoid of red 
rod; ell.bac.vr., ellipsoid of green rod; ell.con., cone ellipsoid; gtt.ol., oil globule; 
mb.lim.ex., externa limiting membrane; my.bac.rb., myoid of red rod; my.bac.or., 
myoid of green rod; my.con., cone myoid; prs.dst.bac.rb., outer member of red 
rod; prs.dst.bac.vr., outer member of green rod; prs.dst.con., outer member of 
cone; st.nl.ex. external nuclear layer. 


results, the exact opposite of those reported by Ewald and 
Kihne, is assumed from the context of the previously cited 
quotation (p. 431), and from his illustrations. 

Angeluecci (84) measured the length of the outer member of 
the frog’s rod and found it shortens in the light; later (90) he 
confirmed this result by measurements of the large rods of the 


CHANGES IN THE VISUAL CELLS OF THE FROG 433 


salamander where the differences in length were more striking. 
Arcoleo (90) and Garten (’07) reported similar conditions for 
the toad and frog respectively. 

In most accounts of the changes occurring in the anuran rod, 
due to the action of light, reference is not made to the number 
of individuals experimented upon, and in no case are definite 
measurements of length given, judgment of the eye apparently 
being the only criterion adopted. 


= | Be = ae. prs. dst. bac. vr. 


rs. dst. bac. rb: 
prs. dst. bac. r ----ell. bac. vr. 


j 
+---my. bac. vr. 


& ig dst. con. 
f 


mb. lim. exc 


Fig. 2. From a light-adapted retina of the frog, Rana pipiens, showing the 
positions assumed by the rods and cones. X 1130. 


In the experimentation which forms the basis of this paper, 
an attempt has been made to correlate an influence of light with 
the following possible changes in the frog’s visual rod: 1) changes 
in the position of the red rod nucleus with respect to the external 
limiting membrane; 2) changes in the length of the red and green 
rod myoids; 3) relative changes in the length of the red rod myoid 
at the center and at the periphery of the retina; 4) changes in the 
diameter of the various component parts of the red rod. 


434 LESLIE B. AREY 
EXPERIMENTATION 


Individuals of Rana pipiens, approximately uniform in size, 
were exposed to bright diffuse daylight for eight hours, or to 
total darkness for a minimum length of twenty-eight hours. 
At the expiration of these periods of light and dark-adaption, 
the cranium of each animal was first split sagittally, and then 
cut transversely just caudad to the eyes. The resulting moieties 
of the cranium, with the contained eyes, were dropped into 
Perenyi’s chromo-nitric fixative; in this solution they underwent 
fixation, the condition of illumination being identical with that 
to which they had been experimentally subjected. The oper- 
ation on dark-adapted animals was accomplished by the light 
from a photographic red lamp and demanded but a few seconds 
time. 

After dehydration and the removal of the lens, the eyes were 
imbedded in paraffine, sectioned 8 thick, and stained with 
Ehrlich-Biondi’s acid fuchsin-orange G-methyl green mixture. 
In light-adapted retinas the expanded pigment masked the 
visual cells to a greater or less degree, hence, nascent oxygen 
was used as a bleaching agent in these preparations. The 
results obtained with the Ehrlich-Biondi stain were very satis- 
factory, yet it is of interest to note that the staining reaction of 
several definite structures was often variable. For example, the 
outer member of the red rod in most cases stained red with the 
acid fuchsin, yet in some retinas from the same series they were 
colored an intense orange from the orange G, although the treat- 
ment in both cases had been, so far as is known, identical, the 
series having been carried along simultaneously step by step. 
Moreover, in one preparation, at least, it was observed that the 
outer members in one half of the retina were stained orange, 
whereas in the other half they were stained red. ‘This selective 
stainability is doubtless indicative of a variable physiological 
cytoplasmic state. Preparations in which the outer members 
of red rods selected the orange G, showed the green rods stained 
red with the acid fuchsin. In retinas that had been bleached 
(potassium chlorate and hydrochloric acid being the reagents 


CHANGES IN THE VISUAL CELLS OF THE FROG 435 


used), the rod outer member varied in color from blue-violet 
to red-violet. Cases involving a somewhat similar variability 
might be cited with respect to several other structures in the 
retina. 

Experimentation was made upon sixty retinas, from which 
twenty-three light-adapted and twenty-three dark-adapted prep- 
arations were selected for measurement. Preparations, in 
which the external limiting membrane was not apparent, or in 
which wrinkles caused oblique sections, were rejected. All 
Measurements were made with a Leitz 1/12 homogeneous im- 
mersion objective and a Zeiss No. 2 micrometer eyepiece. 


A. Influence of light on the position of the red rod nucleus 


The object of this series of measurements was to determine the 
effect of light and darkness upon the position of the nuclei of the 
red rods with respect to the external limiting membrane. 

The nuclei of the rod- and cone-visual cells comprise the external 
nuclear layer (figs. 1 and 2, st.nl.ex.) of the retina. According 
to Greeff (00, p. 133), these two kinds of nuclei in amphibians 
can be distinguished, morphologically, only with great difficulty; 
hence it is essential to inquire whether any criterion exists where- 
by the rod nuclei can be identified with a tolerable degree of 
certainty. The illustrations of the frog’s retina given by Greeff 
(00, pp. 96, 102) represent the nuclei of the red rod-visual cells 
as lying considerably nearer the external limiting membrane 
than the nuclei of either the green rod-visual cells or of the cone- 
visual cells. If, therefore, in any preparation attention be di- 
rected to the nuclei which protrude farthest beyond the external 
limiting membrane, one may be reasonably sure that the nuclei 
of the red rods only are under consideration. Ten light-adapted 
and ten dark-adapted retinas were selected at random for measure- 
ment. Not only was considerable regional variability in the 
position of nuclei found in individual retinas, but also the varia- 
bility in any limited area was extensive—ranging from nuclei 
whose edges were at the same level as the external limiting 
membrane to nuclei which were 3.0 « above that membrane. 
Partly for this reason, and partly to be certain that only the 


436 LESLIE B. AREY 


nuclei of red rods were under observation, attention was directed 
solely to the mean distance to which the maximally extended 
nuelei protruded beyond the external limiting membrane. 

The following values (grand means) were obtained: in hght 
(10 retinas), 2.94; in darkness (10 retinas), 2.2n. 

On account of the degree of variability observed, I do not re- 
gard this slight difference, which, incidentally, is not in agree- 
ment with the results of Stort (87) and Garten (07) on Triton, 
or of Angelueci (90) on the salamander, as indicative of a photic 
influence: Since the condition recorded in Triton and the 
salamander is not clearly demonstrable in my preparations of 
the frog, it follows that whatever changes are found in the rod 
myoid will be due chiefly to the activity of the myoid itself. 


B. Influence of light on the length of the rod myoid. Relative 
changes at center and periphery of retina 


a. Red Rods. In every retina, at least ten measurements 
were made, on each side of, and not far from, the optic nerve. 
The averages of these twenty or more values are given in tables 1 
and 2 as central measurements. Similarly the means of at least 
twenty measurements (ten on each side), made well toward the 
periphery of the retina, are recorded as peripheral measurements. 
In order to avoid unconscious selection all measurements were 
made on consecutively-placed rods. The myoid length in the 
tables constitutes the distance from the external limiting mem- 
brane to the nearer edge of the rod ellipsoid (figs. 1 and 2, my. 
bac.rb.). 

The tables show that there is considerable variation in indi- 
vidual retinas, yet the rod myoid unmistakably elongates in the 
light and shortens in darkness (figs. 1 and 2). Although the 
mean values, at the center and periphery of individual retinas, 
may also vary greatly in either set, the grand means are practi- 
cally identical. 

The measurements of individual light- and dark-adapted rods 
overlap in many instances; furthermore, the mean for certain 
groups of ten deviates considerably from the average for both 
groups of ten, which constitutes the central or peripheral values 


CHANGES IN THE VISUAL CELLS OF THE FROG 437 


of the tables, as the case may be. In such instances the process 
of averaging two groups of ten serves to mask this condition, 
hence the tables, for the sake of compactness, are faulty in this 
respect. 

The conclusion follows, therefore, that the photomechanical 
responses of the frog’s red visual-rod myoid are in agreement 
with those of other vertebrates in which changes have been 
demonstrated. 

The older workers (Angelucci ’84; ’90; Gradenigro, °85; and 
Arcoleo ’90) perhaps erred, either in not making actual measure- 


TABLE 1 

Measurements from twenty-three dark-adapted retinas of Rana pipiens. The 
values are in micra and represent measurements taken along axes coinciding 
with radii of the eyeball. Each value for the length of the red rod myoid is the 
mean obtained from twenty consecutively-placed elements 


| 
| NERVE FIBER CHORIOID TO LENGTH OF LENGTH OF 


NUMBER OF LAYER TO ee ine woh RED ROD MYOID RED ROD MYOID 
en a ere a oe 

1 | 107 64 5.0 4.9 

2 136 70 4.0 5.7 

3 114 72 4.6 3.0 

1 114 64 4.4 4.9 

5 122 69 6.6 7.0 

6 110 69 5.3 6.7 

7 100 64 4.0 6.4 

8 106 64 6.7 6.6 

9 129 67 6.4 | 6.4 
10 100 69 9.3 7.6 
11 114 69 6.9 6.9 
12 129 69 5.9 5.7 
13 114 73 6.2 5.7 
14 97 62 3.6 4.3 
15 97 63 4.7 5.0 
16 111 69 4.6 9.1 
17 107 63 5.4 6.0 
18 107 64 6.4 6.2 
19 114 63 6.6 6.4 
20 103 60 7.2 6.9 
21 89 57 (Or 6.3 
22 93 62 5.7 (hers 
23 93 60 6.4 6.2 
1) TN 7 109 66 5.8 6.0 


438 LESLIE B. AREY 


ments, or were influenced by their observations on the more 
strikingly mobile cone (figs. 1 and 2; my.con.), the movements 
of which are the reverse of those exhibited (such is my belief) 
by mobile rods in general. 

In recent papers (715; 716) the writer favored the anomalous 
photomechanical responses of the frog’s rod, reported by the 
several older workers, as being more trustworthy than the some- 
what confusing account of Lederer? (08). The writer (15), 


TABLE 2 


Measurements from twenty-three light-adapted retinas of Rana pipiens. The 
values are in micra and represent measurements taken along axes coinciding 
with radii of the eyeball. Each value for the length of the rod rod myoid is the 
mean obtained from twenty consecutively-placed elements 


NUMBER OF ; Bia = oe ‘i E resend e a NG eee ne bee D R a Sone D 
ANIMAL EXT pee ae NG MEMBRANE AT ee OF et 

1 100 67 9.9 10.2 

2 111 73 13.6 14.9 

3 93 74 11.9 12.0 

4 129 62 ml 7 12.6 

5 ILL? CZ 11.9 12.6 

6 114 72 12.6 11.9 

a 114 64 13.33 12.2) 

8 127 64 9.2 eZ 

9 127 69 8.6 10.3 
10 119 We, 16.9 16.2 
11 114 72 13.9 14.7 
12 120 72 143 AG 
13 92 62 9.9 9.2 
14 102 69 13)4 13.0 
15 107 60 13.0 11.9 
16 125 63 9.6 9.2 
17 100 62 11.4 12.4 
18 106 62 9.9 9.6 
19 93 61 8.6 7.9 
20 96 69 127 9.6 
21 107 64 7.9 ee, 
22 97 64 8.6 9.7 
23 129 69 11.9 12.9 
Miéeantic erie 100 67 1S | 11.6 


3 Lederer s figures of isolated rods do not necessarily show how long he rod 
myoid really was, either in darkness or in light. Presumably, the myoid, as 
the result of teasing, broke approximately at the level of the external limiting 


CHANGES IN THE VISUAL CELLS OF THE FROG 439 


furthermore, made use of these data in an argument against the 
feasibility of attempting to advance (in the light of our present 
knowledge) a single rational explanation for the diverse photo- 
mechanical movements of the visual rods. It is evident, how- 
ever, that the conclusion reached in the present investigation 
renders this particular objection invalid. 

It is reasonable to expect, although material is not available 
at present to put the matter to an experimental test, that the 
photomechanical behavior of the toad’s visual rod will be found 
to vary in no essential detail from that herein described for the 
frog. 

b. Green rods. From the sixty retinas experimented upon, 
the ten light-adapted and ten dark-adapted preparations which 
showed the most perfect histological preservation were selected. 
for measurement. In each retina, measurements were made of 
the myoid length (figs. 1 and 2, my.bac.vr.) of ten consecutively- 
placed green rods; the results are recorded in tables 3 and 4. 


TABLE 3 


Measurements from ten dark-adapted retinas of Rana pipiens. The values are in 
micra and represent measurements taken along axes coinciding with radii of the 
eyeball. Each value for the length of the green rod myoid is the mean obtained 
from ten consecutively-placed elements 


wowsrn oF | qoexrmawat uiwriva | exremvatuimimina | gggtt"iTH OF 

1 95 66 21.4 

2 94 65 23.0 

3 98 75 28.6 

4 90 69 26.2 

5 112 68 22.5 

6 135 $2 27.7 

¥ 102 63 24.4 

8 109 70 24.0 

9 105 70 24.1 

10 115 72 22.5 
Meant. 2:3 2%’. 106 70 24.4 


membrane, although nothing to this effect is stated. Furthermore, the varicose 
and atypical appearing rods naturally increases one’s caution in accepting his 
conclusions. Although Lederer gave but little emphasis to his observations 
on sectioned material (p. 431), I believe that those observations constitute the 
strongest evidence in support of his thesis. 


THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL. 26, No. 4 


440 LESLIE B. AREY 


TABLE 4 

Measurements from ten light-adapled retinas of Rana pipiens. The values are in 

micra and represent measurements taken along axes coinciding with radii of the 

eyeball. Each value for the length of the green rod myoid is the mean obtained 
from ten consecutively-placed elements 


NERVE FIBER LAYER CHORIOID TO 


pistes TO ayant cea pr ieg ne Re Ge Piette iors pit ss 

1 105 68 25.0 

2, 102 P 7A). 1 

3 127 5 28.2 

4 120 69 28.2 

5 97 68 31.0 

6 112 63 28 .2 

7 Pe 87 PAS 

8 120 75 Dea 

9 127 72 32.4 

10 | 123 66 : 24.9 
Mean sacs. | 116 2, Pil 7 


The tables show that although the difference in the length 
of the rod myoid, in darkness and in light, is small, the length 
of the light-adapted element is quite consistently the greater 
(figs. 1 and 2); hence it seems probable that this difference is 
significant and the photomechanical responses. of the red and 
green rods are analogous. 

This conclusion is not in accord with the results of Angelucci 
(94). Angelucci believed that the red and green rod myoids 
responded similarly to photic stimulation, yet in both cases a 
shortening was said to take place. 


C. Influence of light on the diameter of the red rod 


From my preparations many rods were measured, yet no con- 
stant or significant differences were found in the diameters of the 
outer members (figs. 1 and 2, prs.dst.bac.rb.), or of the ellipsoids 
(figs. 1 and 2, ell.bac.rb.) of the inner members, which could be 
attributed to the influence of light and darkness. The myoid 
of the inner member (figs. 1 and 2, my.bac.rb.) naturally becomes 
tenuous when elongated by the action of light (p. 436). 

One is inclined to question the morphological normality of the 
teased rods figured by Lederer (’08), especially since the appear- 


CHANGES IN THE VISUAL CELLS OF THE FROG 441 


ance of these elements in his own sectioned material was dis- 
similar. In the figures corresponding to the text descriptions 
of dark-adapted rods (the figures and their appended descriptions 
are apparently interchanged in his paper), the varicose outer 
members are twice, and the ellipsoids three to four times, as 
broad as the corresponding parts of the light-adapted rods. 

My observations, therefore, are opposed both to those of 
‘Lederer (’08) who believed that darkness causes the rod to 
swell, and to those of Ewald and Kiihne (’78), who recorded 
that light acts in this manner. 


SUMMARY 


1. Distinct movements of the nuclei of the red rod-visual cells, 
due to photic stimulation, are not demonstrable. Hence move- 
ments of the rods are not produced indirectly in this way. 

2. The myoid of the rod-visual cell elongates in light and 
shortens in darkness. Therefore, contrary to the conclusions of 
the older workers, the photomechanical response of the frog’s rod 
myoid is found to be similar to that occurring in the retinas of 
all other investigated vertebrates. 

The mean length of approximately 1000 myoids from 23 
light-adapted retinas is 11.6 u. 

The mean length of approximately 1000 myoids from 23 dark- 
adapted retinas is 5.9 uy. 

3. A significant regional difference in the length of the red- 
visual rod myoid is not apparent from comparisons of approxi- 
mately 2000 measurements made at the center and the periphery 
of light- and dark-adapted retinas. 

4. The myoid of the green visual rod probably elongates 
slightly in the light and shortens in darkness. 

The mean length of 100 myoids from 10 light-adapted retinas 
ea ag M. 

The mean length of 100 myoids from 10 dark-adapted retinas 
is 24.4 M. 

5. Definite changes in the diameter of the outer member, or 
of the ellipsoid, of the red visual-rod can not be correlated with 
photic influences. The rod myoid, however, does become 
tenuous in the light. . 


442 LESLIE B. AREY 


REFERENCES 


Papers marked with an asterisk have not been accessible in the original. 

Ance.uccr, A. *1884a Una nuova teoria sulla visione. Communic. preven- 
tiva presentata all’ Accad. med. di Roma, 14 Luglio. 

1884 b Una nuova teoria sulla visione. Gazetta Medica di Roma, 
1884, pp. 205-210; 217-223. 

1890 Untersuchungen iiber die Sehthitigkeit der Netzhaut und der 
Gehirns. Untersuch. zur Naturlehre d. Menschen u. d. Thiere (Mole- 
schott), Bd. 14, Heft 3, pp. 231-357. 

ArcoLeo, E. 1890 Osservazioni sperimentali sugli elementi contrattili della 
retina negli animali a sangue freddo. Annali d’Ottalmologia, Anno 
19, Fasc. 3 e 4, pp. 253-262. 

Argy, L. B. 1915 The occurrence and the significance of photomechanical 
changes in the vertebrate retina—An historical survey. Jour. Comp. 
Neur., vol. 25, no. 6, pp. 535-554. . 

1916 The movements in the visual cells and retinal pigment of the 
lower vertebrates. Jour. Comp. Neur., vol. 26, no. 2, pp. 121-202. 

Bout, F. 1877 Zur Anatomie und Physiologie der Retina. Arch. f. Anat. 
u. Physiol., Physiol. Abth., pp. 4-86. 

Ewaup, A., unp KiiHne, W. 1878 Untersuchungen iiber den Sehpurpur. 
Teil 3. Verinderungen des Sehpurpurs und der Retina im Leben. 
Untersuchung. aus d. Physiol. Inst. d. Univ. Heidelberg, Bd. 1, 
Heft 4, pp. 370-422. 

GarRTEN, S. 1907 Die Verinderungen der Netzhaut dureh Licht. Graefe- 
Saemisch, Handbuch der gesammten Augenheilkunde. Leipzig, Aufl. 
2, Bd. 3, Kap. 12, Anhang; 130 pp. 

GRADENIGRO, G. 1885 Uber den Einfluss des Lichtes und der Wirme auf die 
Retina des Frosches. Allz. Wiener med. Zeitung., Bd. 30, No. 29 
u. 30, pp. 343-344, u. 353. 

GREEFF, R. 1900 Die mikroskopische Anatomie des Sehnerven und der Netz- 
haut. Graefe-Saemisch, Handbuch der gesam. Augenheilkunde. 
Leipzig, Aufl. 2, Bd. 1, Kap. 5, 212 pp. 

LEDERER, R. 1908 Verinderungen an den Stiibchen der Froschnetzhaut unter 
Kinwirkung von Licht und Dunkelheit. Centralbl. f. Physiol., Bd. 
22, No. 24, pp. 762-764. 

Miituer, H. 1856 Anatomisch-physiologische Untersuchungen iiber die Retina 
bei Menschen und Wirbelthieren. Zeitschr. f. wissensch. Zool., Bd. 
8, Heft 1, pp. 1-122. 

ScuowaLBeE, J. 1874 Microscopische Anatomie des Sehnerven, der Netzhaut 
und der Glasskérpers. Graefe-Saemisch, Handbuch der gesam. Augen- 
heilkunde, Leipzig, Aufl. 1, Bd. 1, Teil 1, pp. 321-456. 

Srort, A. G. H. van GenpeREN 1886 Uber Form und Ortsveriinderungen 
der Elemente in der Sehzellenschicht nach Beleuchtung. Bericht 
uber d. 18. Versamm. d. Ophthal. Gesell. zu Heidelberg., pp. 43-49. 
1887 Mouvements des éléments de la rétine sous l’influence de la 
lumiére. Arch. néerlandaises des Sciences exact et naturelles, publ. 
p. l. Soc. holland. des Sciences, Tom. 21, Livr. 4, pp. 316-386. 


A PRELIMINARY DETERMINATION OF THE PART 
PLAYED BY MYELIN IN REDUCING THE WATER 
CONTENT OF THE MAMMALIAN NERVOUS SYSTEM 
(ALBINO RAT) 


H. H. DONALDSON 
From The Wistar Institute of Anatomy and Biology 
ONE CHART 
It is a familiar fact that there is a progressive loss of water in 


the brain and spinal cord with advancing age. This is illus- 
trated for the albino rat in Chart 1. 


SSSGSSEsseeesssceesssesssseeerizs 


wae at 


CEE ECE EEELE La 
SESSSER ARF outa 
CTT ere Te 


SB? 2_maee 


200 250 300 360 
° . 


Chart 1 Showing the percentage of water on age in the central nervous 
system of the albino rat. The upper graph gives the values for the water in the 
brain as determined by the formulas (Hatai, in ‘The Rat,’ Donaldson, ’15). The 
lower graph gives the corresponding values for the spinal cord, determined in the 
same way. 

The small black dots indicate the observed values for the several age groups 
for the brain, and form the data for the formulas. The small black triangles 
have a like value in relation to the spinal cord. : 


443 


444 H. H. DONALDSON 


It is also well known that in some mammals all the axons in 
the central nervous system are unmyelinated at birth, while in 
other species a greater or smaller number of them may already 
have their sheaths. In all eases, however, an active formation 
of myelin occurs during the period of rapid growth (Koch and 
Koch 713). 

In a previous study on the percentage of water in the albino 
rat I was misled by certain graphs (Donaldson 710) to the con- 
clusion that probably both the axons and their myelin sheaths 
changed in their water content so as to produce the well known 
reduction of water which is observed, but further study has 
shown that this is an incorrect view, and it is the object of this 
paper to present the evidence of my revised conclusion. 

Since it has already been shown that the loss of water in the 
human brain follows the same course as in the brain of the albino 
rat—and has similar limits (Donaldson ’10)—it is permissible to 
use in the argument certain observations on the human brain. 

From the data for the human brain already in the literature 
I have selected those published by de Regibus (84), because 
this author evidently examined only the outer layers of the cor- 
tex when making his determinations for the water in the gray 
substance and because he was able also to obtain remarkably 
uniform results for all of his determinations. 

De Regibus tested four male brains, 25 to 76 years of age and 
three female brains, 30 to 60 years of age. 

In Table 1 there appear also determinations of the water 
content in the human cortex and in the fibers at birth. These 
are based on the records of Weisbach (’68). 

This enables us to contrast in Table 1 the conditions at birth 
with those at maturity. 

TABLE 1 


Percentage of water in the gray and white substance of the human brain at birth and 
at maturity 


CALLOSUM 


| CORTEX (GRAY) (WHITE) 


per cent per cent 


AC ibintha (Weisbach) ie .cuae. > cite alee cee epee eeaereaetae 88 88 
At Me puniby Ge: Ee2TDUS) \.4:05.5., so gmlsutetait at eee ae coed 86 70.4 


WATER CONTENT—MAMMALIAN NERVOUS SYSTEM 445 


According to this table the (gray) cortex has lost 2 points and 
the (white) callosum 17.6 points in the process of maturing. 

It is never possible at maturity to obtain the cortex or any 
other gray mass without some admixture of myelinated fibers 
and I have therefore, provisionally, credited one point of the 
joss, noted by de Regibus in the water content of the cortex, to 
the presence of myelin. This implies but a small proportion of 
myelin since if we assume that myelin has 48 per cent of water 
the reduction of 1 per cent would mean that about one thirty- 
ninth of the mass was represented by myelin. According to 
this assumption the mature gray substance (cortex)—when the 
myelin is excluded—contains 87 per cent of water, and in the 
computations which follow the neurons without myelin are 
assumed to have 87 per cent of water, except at ten days, when 
they are credited with 88 per cent. 

The fact that the fibers without myelin have at birth a high 
percentage of water (88°) while at maturity, after myelination, 
they have lost 17.6 points is the sort of evidence which fur- 
nishes the basis for the current, but unsupported view, that the 
loss of water is to be associated with the formation of the myelin. 
It is our object to present more precise information bearing on 
this point. 

To obtain a notion of the approximate distribution of the water 
between the myelin and neurons proper, it is necessary to have 
data on the relative abundance of these two constituents of the 
brain. 

In 1913 W. Koch and M. L. Koch made a study of the chemical 
composition of the brain of the albino rat at six ages between 
birth and maturity, and of the spinal cord at one age. The 
data thus obtained are those which will be utilized here. The 
authors determined seven fractions: Proteins, organic extractives 
and inorganic constituents, which three taken together, we shall 
designate protein (or non-lipoid), and phosphatides, cerebro- 
sides, sulphatides and cholesterol, which four taken together, we 
shall designate lipoid. 

These data give us at each age, therefore, the protein and the 
lipoid present in the brain, or to be a little more exact, we should 


446 H. H. DONALDSON 


say the lipoid and the non-lipoid fractions. The lipoid (in 
part) represents the myelin sheaths, while the protein, with the 
remainder of the lipoid, represents the cell bodies and their 
unsheathed axons. 
With the exception of the one day group, the ages for which 
analyses were made, are given in table 2. 
TABLE 2 


To show, for the brain of the albino rat at five ages and for the spinal cord at one 
age the percentage of water in the myelin as computed according to the method 
described. The protein and lipoid are given in percentages of the total dry sub- 
stance. (Based on table 2. Koch and Koch ’13) 


Brain 
(1) (2) (3) (4) (5) (6) (7) 
PERCENTAGE OF 
Ga ikee ron Weree WATER. 
AGE IN CORRECTED CORRECTED DIFFERENT IN MYELIN = 
DAYS PROTEIN LIPOID AGES. ; In Neurons} Liporps (C) 
LIPOID aT 20 Pere = protein (COMPUTED) 
DAYS = 1 b a a (C) (See p. 447) 
EOE (assumed) 
per cent per cent 
10 93.80 6.2 86.5 88 63.8} 
20 88.88 il 1.0 82.5 87 46.5? 
40 82.86 IA te! 145) 79.4 87 42.7 
120 (aye Ml 24 84 Pe Pe 78.4 87 52.4 
210 76.36 23:08 Paya \t Toil 87 47.4 
ANenage;oih20 to 2h ais aercs eiaahelnc atetvale Gait enna store eee 47.8 
Spinal cord 
120 | 52.92 | 47.08 | 4.2 | 70.4 | 87 | AL) 


1 First traces of myelin. 
? Myelin well shown. 


At one day—or practically birth—it is found that the lipoid 
is present to the extent of 0.31 or nearly one-third of the weight 
of the protein. There is, however, no visible myelin at this age, 
so it is concluded that this proportion of the total lipoid is nor- 
mally associated with the protein and is not to be included in the 
lipoid which forms the myelin sheaths. We have treated the 
data for the later age groups in accordance with this relation, 
and in each case have taken from the total lipoid found an 


WATER CONTENT—MAMMALIAN NERVOUS SYSTEM 447 


amount equal to 0.31 of the protein found. The remaining 
amount of lipoid is assumed to be that used for the sheaths. 

In table 2, the column (2) headed Corrected Protein gives the 

observed protein (non-lipoid) plus 0.31 of itself and the column 
(3) headed Corrected Lipoid gives the observed 5 less the 
amount of lipoid added to the protein. 
_ In table 2 the data are given in five age groups for the bos 
and in one age group for the spinal cord. It is to be noted that 
the 10 day brain group—which stands just at the beginning of 
the myelin formation—is for the moment excluded from the 
discussion and we begin the comparisons which are to be made, 
with the 20 day brain group. 

In the brain series (with one exception) the corrected protein 
diminishes and the corrected lipoid increases with advancing 
age. Between 20 and 210 days the proportion of the lipoid 
doubles—column (4). We have in column (5) the observed 
percentage of water in the brain as a whole. It is assumed, as 
previously noted, that the corrected protein (neurons in the 
strict sense =both cell bodies and axons) have 87 per cent of 
water. From these several data we can compute the percentage 
of water to be assigned to the corrected lipoid, which represents 
the myelin.! 

The method of computation may be illustrated by the data for 
the 20 day group. Reference to table 2 shows that at this age 
there is 1 part of lipoid (11.12°%) to 8 parts of protein (88.88%). 
This gives 9 parts, representing the entire brain and having 82.5 
per cent of water. The product, 9 X 82.5 = 742.5. We assume 
that the 8 parts of protein have 87 per cent of water. The 
product, 8 X 87 = 696. The 1 part of lipoid, representing the 
myelin, will then have a percentage of water equal to the dif- 
ference of these products = 742.5 — 696 = 46.5 per cent. 

It is hardly necessary to point out that a division of the brain into myelin 
on the one hand and neurons on the other fails to enumerate several structural 
elements which are also present though representing a relatively small mass. 
There are, in addition to the neurons, glia and ependymal cells; blood and lymph 
vessels; blood and lymph and a slight amount of connective tissue. Over against 
the neurons plus this group of elements we put the myelin, but for the purposes 


of the present discussion it is convenient to speak of the rest—the non-myelin 
portion—of the brain as if represented by the neurons alone. 


448 H. H. DONALDSON 


As table 2 shows, similar computations give percentages of 
water for the myelin in the brain, which range from 42.7 per cent 
to 52.4 per cent and which yield a mean value of 47.8 per cent. 
The computation for the spinal cord gives 51 per cent. The 
significance of these results lies not in the particular percentage 
of water here determined for the myelin—as that depends some- 
what on the percentage of water assumed for the protein—but 
on the similarity of the values found in all the five cases examined. 

However, it is found on trial that one cannot depart far from 
the value of 87 per cent for the proteins without obtaining rather 
improbable percentages for the myelin, so that this value is 
probably nearly right. 


DISCUSSION 


Those familiar with the published data for the percentage 
of water in the cortex will at once perceive that the value given 
by de Regibus (86%) seems high. The various water records 
for the cortex run down as low as 83.5 per cent. The differences 
between the various determinations are, however, almost cer- 
tainly due to the varying amounts of white substance included 
in the sample, and as has already been stated the value chosen 
is probably close to the true value. 

In connection with the computations there are, however, two 
conditions which have been assumed to be constant but which 
in all probability, are subject to variation. I refer to the density 
of the myelin and to the fraction of the lipoid to be assigned to 
the protein. As to this last condition, it would be plausible 
to think of a larger fraction of lipoid in the axons than in the 
cell bodies. If this were true it would be necessary to increase 
this fraction in the case of the spinal cord or the callosum. 

It is also possible that aside from this method of distribution 
the fraction of lipoid in the neuron may increase with age. The 
slight loss of water during the first ten days of (rat) life is possi- 
bly due to such an increase. Finally, the density of the myelin 
may change with age, as its chemical composition certainly does, 
and it is conceivable that it has a higher water content when 
first formed, as is suggested by the 63.8 per cent given for the 
ten day record in table 2. 


WATER CONTENT—MAMMALIAN NERVOUS SYSTEM 449 


If we compare the drawings of Watson (’03), showing the in- 
crease of the visible myelin in the cerebral hemispheres and in 
the spinal cord of the rat, with the chemical results here used, 
we see that the histological pictures show a more gradual ap- 
pearance of myelin than the chemical results, or the water 
determinations, would suggest. This probably depends on the 
fact that it is only a fraction of the lipoids forming the sheaths, 
which takes the haematoxylin stain, and-this stainable fraction 
forms at first a smaller, but later a larger portion of the entire 
sheath (Koch and Koch 713; Smith and Mair ’08). 

There is still one more modification in the formation of myelin. 
Tribot (05) has contrasted in terms of dry substance the relative 
amounts of albuminoides and the fats in the nerve tissue of the 
guinea-pig at different ages. The percentage value of the fats 
increases from 11 days (his first observation) up to 120 days— 
after which it begins to fall. The fats of Tribot are the lipoids 
of Koch’s analysis and it is of interest to note that the 120 day 
record for the brain in table 2, column (4), also shows the highest 
proportion of corrected lipoids. The observations of Dunn (’12) 
which show in the myelinated fibers of the second cervical nerve 
of the rat the highest relative areas for the myelin sheaths at 
75 days and 132 days—seem to fit with these other observations 
and to suggest that the formation of lipoids with advancing age 
fluctuates in such a way as to show a maximum about the end 
of the active growing period of the central nervous system. 

This discussion of possible factors modifying these determi- 
nations has been introduced to clear the way for further work on 
the main question, but so far as one can foresee the effect of 
taking them into consideration, it would tend to make more 
uniform the values thus far obtained. 


CONCLUSIONS 


We conclude from these results that there is no evidence that 
the cell bodies.and their unsheathed axons suffer more than a 
slight loss of water between birth and maturity, and that the 
progressive diminution in the water content of the entire brain 
and spinal cord is mainly due to the accumulation of myelin, 


450 H. H. DONALDSON 


with a water content of about 48 per cent. Moreover, the myelin 
must be regarded as a more or less extraneous substance, having 
but little significance for the characteristic activities of the 
neurons. 

If we compare the loss of water in the case of the nervous 
system with that in the muscular system, which also contains 
a large proportion of fat (Tribot ’05), we find that while the two 
systems lose about the same percentage of water between birth 
and maturity (Lowrey 713), yet in the case of the nervous system 
alone is this lipoid (or non-protein substance) accumulated out- 
side of the cell. From this it is seen that the neuron is peculiarly 
able to maintain its early water-solids composition and that it 
accomplishes this by throwing out the material, which in the 
muscles is retained within the cells. 

As the diminution in the percentage of water in the central 
nervous system is preéminently a function of age, and as it 
appears to be due almost entirely to the formation of the myelin, 
it follows that the myelin formation is also a function of age 
(Donaldson 711). A glance at the graphs in Chart 1 shows that 
the most active production of myelin, as indicated by the rapid 
loss in the percentage of water, occurs early, i.e., during the first 
forty days of rat life, in the brain, and during the first hundred 
days, in the spinal cord. 


WATER CONTENT—MAMMALIAN NERVOUS SYSTEM 451 


LITERATURE CITED 


Donautpson, H. H. 1910 On the percentage of water in the brain and in the 

spinal cord of the albino rat. Jour. Comp. Neur., vol. 20, no. 2, pp. 
119-144. 
1911 The effect of underfeeding on the percentage of water, on the 
ether-alcohol extract, and on medullation in the central nervous 
system of the albinorat. Jour. Comp. Neur., vol. 21, no. 2, pp. 139-145. 
1915 The Rat. Memoirs of the Wistar Institute of Anatomy and 
Biology, no. 6, pp. 1-278. 

Dunn, E. H. 1912 The influence of age, sex, weight and relationship upon the 
number of medullated nerve fibers and on the size of the largest fibers 
in the ventral root of the second cervical nerve of the albino rat. 
Jour. Comp. Neur., vol. 22, no. 2, pp. 131-157. 

Kocu, W. anp Kocu, M. L. 1913 Contributions to the chemical differentiation 
of the central nervous system. III. The chemical differentiation of 
the brain of the albino rat during growth. J. of Biol. Chemistry, vol. 
15, no. 3, pp. 423-48. 

Lowrey, L. G. 1913 The growth of dry substance in the albino rat. Anat. 
Rec., vol. 7, pp. 143-168. 

DE Rearisus, C. 1884 Determinazione dell’ acqua contenuata nelle sostanze 
grigia e bianca del cervello umano. Atti. d. reale Accad. di Med. 
di Torino, vol. 6, pp. 323-328. 

Smit, J. L. anp Marr, W. 1908 An investigation of the principles underlying 
Weigert’s method of staining medullated nerve. J. of Pathology and 
Bacteriology, vol. 13, pp. 14-27. 

Trisot, J. 1905 Sur les chaleurs de combustion et composition chimique des 
tissus nerveux et musculaire chez le cobaye, en fonction de l’age. 
C. R. Acad. Se. Paris, T. 140, pp. 1565-1566. 

Watson, J. B. 1903 Animal education. Contributions from the Psychol. 
Lab. of the University of Chicago, vol. 4, no. 2. Univ. of Chicago 
Press, Chicago, Ill. 

Werssacu, A. 1868 Der Wassergehalt des Gehirns nach Alter, Geschlecht und 
Krankheiten. Med. Jahrbiicher, vol. 16, nos. 4 and 5, pp. 1-76. 


x ry" 4 sie oy f' ht a bust ve Drea 


i, A wi was 
Yuk: ‘3 +a sam Wey ay vow Tah i 
eau) fo Mame ac prey fie, jy aha 


* 


: it gta Meee, hs, ti 
j / io mer <. y HE te rt 
“ge ys ie 5% Wier tal aa rh SN oe 
a ~ j if; ae ‘4 tage sd MY) pes Sb at aeay von 
ee of ay in yh 
i he f 4 
y i? ches y A “te ‘ y 
oe zi 4 PPA IW ery Stan + | F 
ani x Wie. 
mt a, | ' ¥ Vf j . shear Be q 4 
Bis Shh Bod \ i ”, 
Pe Ce aan 
ay 7 f ae eat 4)" 4 nO! { u g4A 
ties , . 
. nh ih f 7 C ‘> 
j i, Ps a a y h - 


Bb | ee | Abeta inte A Rae ) ‘ina 
' . Ly : WAS a . ; Je 

My ’ RED gh h lbp! a ‘ Lz oe Oe ea er Vane) wn hd Teo aly r 
\ ’ a ain : Man ; P ir oL ae! 
Ke i. ey A iF fo , ‘ Pah bone Oe bt 13 4 Ve arent é oe id : 
a tS L } ‘ + a’ 7 
Hal mp | rie ene a Ty 
re meee ete GOA be. A yi Aa na 


a 
' * 
, a 4 . 
‘| g Se i 
' 
a i y * +h) { ( 
ee aha: | to 
a te ' DAF * . 4 ; id 
ton ae 
. oe vey 
ae ri 
B \ * ; 
er M i} 
~ yy P , . A s 
: a Pe dad 0 ie 
wf y 
i , ea ik ca i 
[ cy 
_ 
‘ 
“ 
x ‘ 
= 7 
’ 
wv 
, ‘ 
ah ‘ 
/ 
> 
a . 
ui 
uf 
cS : ci ‘ 
Be 5 é 
; ‘ 
x rn j 


THE TASTE OF ACIDS! 
W. J. CROZIER 


TWO FIGURES 


I. The problem of irritability is essentially a matter con- 
cerning sense organs; in these structures we find cells ‘‘as nearly 
as possible unifunctional”’ (Lucas, ’09, p. 328) with respect to 
the capacity for being stimulated. Hence the analysis of irri- 
tability should properly be founded upon the study of receptor 
physiology in higher animals, rather than upon an examination 
of the properties of protoplasm in deceptively simple protozoans. 

Two contrasted methods have in the past been used in the study 
of stimulation. The first of these deals with the responses of 
unicellular organisms, together with the (usually vague) appli- 
cation of principles derived from cells of more or less generalized 
type. The second method has consisted in the analysis of 
sensations. Yet, if we consider such questions as those involved 
in taste stimulation, we find that no definite conclusions have 
been reached. The very complexity of organization in ‘simple’ | 
forms is probably in the main responsible for the results obtained 
in the first field, results which we find stated in such weak generali- 
zations as occur in Verworn’s book (’13, pp. 41, 86, 94). ‘The 
second procedure, the study of human sense organs, has inevit- 
ably come to be clouded by psychological interpretations. 

It appears that the possibility of arriving at some understand- 
ing of what occurs in stimulation has not yet been sufficiently 
tested out from the standpoint of sense organ irritability as 
revealed in animal reactions. This paper is a contribution 
toward that end; it deals with the problem presented by the 
taste of acids. I am indebted to Prof. G. H. Parker for his 
criticism of the manuscript. 


1 Contributions from the Bermuda Biological Station for Research, No. 46. 
453 


454 W. J. CROZIER 


II. Practically the whole of the difficulty is contained in the 
often considered case of acetic and hydrochloric acids. A compar- 
ative study of the reactions to these and other acids as given by 
the earthworm (Hurwitz, ’10; Crozier, ’16), leads to a preliminary 
simplification of the matter, namely to the formulation of two 
separate problems: these acids stimulate, they also result in 
a sour taste. These two problems must be considered inde- 
pendently; we must distinguish between (a) that property of 
an acid which causes it to be efficient in stimulation and (b) 
that which determines the sourness of an acid solution. 

The relative efficiency of an acid as an agent of stimulation 
may be measured by the liminal effective dilution. The point 
I wish to make is that the property of hydrochloric acid which 
causes it to be tasted in solutions about five times as dilute as 
that limiting the tastability of acetic acid may have no immedi- 
ate connection with the common cause of the sourness perceived 
when these acids act upon the tongue. 

This point of view contains the possibility of an explanation 
for (and is, reciprocally, in part justified by) the fact that acids 
exhibit a characteristic astringency in solutions so dilute that 
they are no longer sour. There is some evidence, also, that the 
sourness and astringency may be separated experimentally, 
as by the use of cocaine. There is therefore ground for the 
opinion that in taste excitation by acids two processes occur, 
and that the production of a sour taste is the secondary one. 

The points which require solution with respect to human 
taste excitation by acetic and hydrochloric acids are: 

(a) Both these substances stimulate. 

(b) They stimulate in different degrees, in dilute solutions 
the stimulus from hydrochloric acid being the more intense. 

(c) They both result in a sour taste. 

(d) They fail to stimulate when too greatly diluted. 

(e) The dilutions which limit the capacity of the two acids 
to stimulate are very different (HCl n/900 =; acetic, n/200 =). 

Each of these points—except (c)—is closely paralleled by the 
details of earthworm reactions to acids (Crozier, 716), and also 


THE TASTE OF ACIDS 455 


by the reactions of AZolosoma, a freshwater oligochaet (Kribs, 
“AD ): | 

It adds somewhat to the clearness of the discussion if, as I 
have previously proposed (Crozier, ’14, p. 16), we restrict the 
word ‘stimulus’ to mean the change induced in a receptor by 
the action of a stimulating agent. The explanation of the acid 
taste is made easier by the fact that one is not called upon to 
account for a sour taste resulting from heterologous stimula- 
tion; there is no good evidence for the existence of a sour taste 
not directly produced by acid. The extreme specialization 
of the acid taste makes it a very favorable case for analysis. 

III. It is sufficiently obvious that only the surface of the 
receptor is immediately concerned in stimulation. This view 
is reasonable upon purely morphological grounds, such as the 
modifications of the exposed ends of sensory cells? and the rela- 
tions of nerve fibrils to the surface of secondary sense cells. 
The way in which this conception of the cell surface as the organ 
of irritability has been elaborated by R. 8. Lillie and others 
need not be discussed here. But it may be pointed out that the 
method of interpreting mechanical excitation has been indicated 
by Osterhout (15) and by such observations as that of Evans 
and Winternitz (Evans and Schulemann, ‘14, p. 453). Un- 
equivocal evidence in this direction is afforded by Harvey’s 
work on the penetration of cells by alkalies (Harvey, ’14 c), 
by his experiments with acids (Harvey, ’14 ¢), and by my own 
study of this latter subject (Crozier, °15, °16); the evidence - 
referred to is, in brief, to the effect that penetration of the body 
of the cell substance is altogether too slow a process.to account 
for the rapidity of taste stimulation. The slowness of acid 
penetration is due to the resistance offered by the cell surface, 
a resistance which practically disappears with the death of the 
cell. An examination of cell penetration by various acids, 
employing a wide range of dilutions, should therefore yield 
valuable information regarding the composition and behavior 
of the resistant cell surface. In several cases a study of this 


2 The propable chemoreceptors of the earthworm have been described by 
Miss Langdon (’95) and Bovard (’04). 


THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL. 26, NO. 4 


456 W. J. CROZIER 
kind has been made possible by the discovery of intracellular 
indicators sensitive toward acids (Harvey, ’14b, Crozier, 715). 
Measurements of cell penetration by acids show, so far as 
concerns the present problem: 1) that both acetic and hydro- 
chlorie acids are able to penetrate the cell; 2) other things being 
/60 
140 


/20 


/00 


gO 


60 


Penetration time, minutes 


40 


Zot 6 


0 100 200 300 400 500 £600 
Dilution 
(Liters containing one equivalent) 


Fig. 1 The average time of penetration of cells by hydrochloric, butyric, 
and acetic acids. 


1, Acetic acid ) 
8, Butyrie acid i integument of Chromodoris zebra, a nudibranch. 
§, Hydrochloric acid : 


2, Butyric acid 

4. Hydrocloric¢ acid 

iO manoma@rozter (2116). ‘ab 2 ie 

2, 4, from Harvey (14¢c), at 28°. 

Butyric acid is introduced for comparison, since Harvey (loc. cit.) gives no 
data for dilutions of acetic; the principle is the same for both these acids. It 
should be noted that in Harvey’s experiments the acid was dissolved in a balanced 
salt solution, in mine in distilled water. 


f testis epithelium of Stichopus ananus, a holothurian. 


THE TASTE OF ACIDS 457 


equal, the speed of this penetration varies with the acid and its 
concentration; 3) the curves relating penetration time to con- 
centration are different for the two acids, dilute solutions of hydro- 
chloric penetrating much more easily than acetic. Points of re- 
semblance, in fact of identity, are found among the characteris- 
tics of stimulation by these acids, and have been enumerated as 
a, b, e, in section II. The closeness of the parallelism indicates 
that the stimulus is due to the penetration of the surface layer 
of the cell. In stimulation by these acids the essential step is 
the union of the stimulating agent with some constituent of 
the ceptor surface. 

According to Becker und Herzog (’07), the decreasing order 
of intensity of taste sensation due to the acids they tested was 
HCl, HNOs, trichloroacetic, formic, lactic, acetic, butyric. These 
acids penetrate the indicator-containing cells of Chromodoris 
zebra in the order HCl, HNQOs;, formic, monochloroacetic, lactic, 
butyric, acetic (considering a range of concentrations; Crozier, 
16). There is reason to believe that trichloroacetic penetrates 
more speedily than monochloroacetic; and with this in mind, 
the parallelism between these two series becomes surprisingly 
close. This particular case is quoted with the purpose of point- 
ing out that a small series of acids, such as that just cited, may 
to all appearances support the view that the intensity of taste 
excitation is proportional to the ionization strength of the acids; 
similarly, a penetration series may be chosen which seems to 
prove a corresponding relation. Yet, further examination of a 
large number of acids shows that ionization is but one of the 
factors determining cell penetrating power, and the existence 
of influences in addition to ionization has also been suspected 
with reference to acid stimulation; direct evidence upon this 
point has hitherto been lacking. 

IV. Analysis of penetration data obtained with eighteen acids 
has shown (Crozier, 716) that in any given case ability to pene- 
trate the cell appears to depend upon two (or more) factors: 
one is the ionization strength of the acid; the second depends upon 
some other aspect of the acid molecule. To account for these 
results it is further necessary to assume that the cell surface is 


458 W. J. CROZIER 


of heterogeneous constitution. This conclusion may be extended 
to include the surface of chemoreceptors, since it enables us to 
account not only for the reactions of the earthworm to dilutions 
of acids (Crozier, 716), but also clears up a well known anomaly 
in regard to human taste. 

It was shown by Richards (’98) and by Par ibe: (98) that 
the hydrogen ion must in some way be responsible for the sour 
taste. Yet acetic acid, which ceases to be tasted at dilutions 


60 


© 


Penetration time, minutes 


A 
os 


P 


SD 


10 5) di = = 96) s5c100 


Per cent ionization 


Fig. 2 The relation between percentage ionization and time of cell pene- 
tration from solutions of hydrochloric, butyric, and acetic acids. 

1, Acetic acid 

3, Butyrie acid integument of Chromodoris zebra. 

§, Hydrochloric acid 

2, Butyric acid 

4, Hydrochloric acid 

1, 3, 5, from Crozier (’16). 

2, 4. from Harvey (’14c). 


} testis epithelium of Stichopus ananus. 


THE TASTE OF ACIDS 459 


below about n/200 is only 6 per cent dissociated at that con- 
centration, whereas hydrochloric acid, fully dissociated (99 per 
cent), is tasted down to about n/900. The actual hydrogen 
ion concentrations in these solutions limiting tastability are: 
HCl (n/900), C,, = 0.00119 N; acetic (n/200), C, = 0.00035 N. 
Acetic acid is more stimulating than would be calculated from 
its dissociation. 

Exactly similar conditions are found in studying the pene- 
tration of cells by acids, as will be clear from an inspection of 
the two sets of curves in figure 2. In table 1 acid solutions giving 


TABLE 1 
The ionization of hydrochloric, acetic, and butyric acids, in solutions which pene- 
trate cells within equal times. For sources of data see corresponding reference 
numbers in figure 2. 


IONIZATION, PER CENT 


TIME 
OF PENETRATION, Hydrochloric Acetic Butyric 
MINUTES 
(5) (4) [1] (3) [2] 
5 95.0 94.5 2.0 
10 98.5 96.7 2.0 2.1 
15 98.8 97.4 2.1 3.3 2.5 
20 99.0 97.5 3.0 4.0 2.6 
50 4.0 5.5 3.5 


the same penetration time appear in the same horizontal row; it 
is seen that the penetration of cells by acetic acid is much more 
efficient than if its dissociation were the deciding influence. 

It follows that the stimulus produced by acetic acid is due, 
in the first place, to its union with the cell surface; this pene- 
tration of the plasma membrane is more efficient than if dissocia- 
tion strength were the determining factor. The indication is 
that this acid dissolves in a fatty substance located at the cell 
surface (for further evidence, see Crozier, ’16). The stimulus 
due to hydrochloric acid also depends upon its penetration of 
the cell surface, but there is some evidence that it involves 
proteins. 

It is not clear to what extent stimulation by acetic and hy- 
drochlorie acids may be interpreted in terms of the effects which 


460 W. J. CROZIER 


these substances have upon surface colloidal conditions, and 
thus upon cell permeability (Spaeth, 716). Herlitzka (’10) 
has discussed this question with reference to the taste of salts, 
but measurements of the effect of acid on permeability (Oster- 
hout, 714) have not considered the possibility that characteristi- 
eally different results may follow from the action of diverse 
acids; Osterhout’s (14) measurements deal only with hydrochloric 
acid. 

V. There remains to be accounted for the production of a 
unitary taste quality, the sourness of acid solutions. This must 
be related to ionizable hydrogen (Kahlenberg, ’98, Richards, 
98). But it has just been shown that hydrogen ions outside 
the cell surface cannot be the effective agents, since stimulation 
by these acids parallels so closely the peculiarities of their pene- 
tration of the cell. The hydrogen ion must therefore act after 
the acid has united with the ceptor surface. At least two 
possibilities are obvious: (1) the presence of potentially ionizable 
hydrogen within sufficiently concentrated undissociated mole- 
cules is enough to produce the sour taste, or (2) the acids ionize 
secondarily after entering. The first suggestion is not so far 
fetched as it may at first seem, though it is not in accord with 
current teaching. Numerous parallels could be drawn from 
the taste of organic substances (Francis and Fortescue-Brick- 
dale, ’08, pp. 331 et seq.). I believe that this view is possibly 
correct, although something could also be said for the hypothesis 
of secondary ionization following some reaction involved in 
gaining admission to the outer layer of the cell. It is in this 
latter direction that the explanation (Beutner, ’14) must be 
sought for the source of electromotive effects accompanying 
stimulation in general, which presumably occur also in taste 
excitation.’ This interpretation disposes of the difficulty (very 
conspicuous in the case of taste) which is encountered by those 
who would make the process of stimulation and the production 
of electromotive effects completely identical in every case. 


3 A taste cell bathed by a stimulating solution is in a condition very similar 
indeed to that of the various tissues experimented upon by Loeb and Beutner 
(Loeb, 715). 


THE TASTE OF ACIDS 461 
SUMMARY 


An attempt has been made to account for the action of acetic 
and hydrochloric acids upon the sense of taste. It is shown 
that the stimulus due to these substances probably depends 
upon their union with diverse constituents of the ceptor surface. 
It is thus possible to explain, by comparison with the penetration 
of cells by these acids: how they stimulate at all, why they 
stimulate to different degrees, and why acetic acid gives a more 
powerful stimulus than if it acted by dissociation alone. Why 
these acids produce a unitary (sour) taste quality is a problem 
of a different order; it is possible that this depends upon the 
presence of potentially ionizable hydrogen within undissociated 
acid molecules, though secondary ionization may also play a part. 


AGAR’S ISLAND 
BERMUDA. 


LITERATURE CITED 


Becker, C. T. unp Herzoa, R.O. 1907 Zur Kenntnis des Geschmackes. I. Mit- 
teilung. Zeits. f. physiol. Chem., Bd. 52, p. 496-505. 

BeEuTNER, R. 1914 Studies on a new kind of E. M. F. Jour. Amer. Chem. 
Soc., vol. 36, p. 2040-2045. 

Bovarp, J. F. 1904 The distribution of the sense organs in Microscolex elegans. 
Univ. Calif. Publ., Zodl., vol. 1, p. 269-286. 

Crozimr, W. J. 1914 The orientation of a holothurian by light. Amer. Jour. 
Physiol., vol. 36, p. 8-20. 
1915 On cell penetration by acids. Science, N.S8., vol. 42, p. 735-736. 
1916 Cell penetration by acids. Jour. Biol. Chem., vol. 24, p. 255-279. 

Evans, H. M., AND ScHULEMANN, W. 1914 The action of vital stains belong- 
ing to the benzidine group. Science, N. 8., vol. 39, p. 443-454. 

Francis, F., AND Fortescur-BrICKDALE, J. M. 1908 The chemical basis 
of pharmacology. London, 8°, xii + 372 pp. 

Harvey, E. N. 1914a The relation between the rate of penetration of marine 
tissues by alkali and the change in functional activity induced by the 
alkali. Publ. Carnegie Instn. Wash., No. 183, p. 131-146. 
1914b Cell permeability for acids. Science, N.S., vol. 39, p. 947-949. 
1914¢ The permeability of cells for acids. Intern. Zeits. Physik.- 
chem. Biol., Bd. 1, p. 463-478. 

HeruitzkKa, A. 1910 Contributo all’analisi fisico-chimica del sapore dei sali. 
Archivo d. fisol., vol. 7, p. 557-578. 

Hurwitz, 8.H. 1910 The reactions of earthworms to acids. Proc. Amer. Acad. 
Arts and Sci., vol. 44, p. 67-81. 


462 W. J. CROZIER 


KAHLENBERG, L. 1898 The action of solutions on the sense of taste. Bull. 
Univ. Wisconsin, Sci. Ser., vol. 2, p. 1-31. 

Kriss, H. G. 1910 The reactions of Molosoma to chemical stimuli. Jour. 
Exp. Zodl., vol. 8, p. 43-74. 

Lanapon, Fanny 1895 The sense-organs of Lumbricus agricola, Hoffm., 
Jour. Morph., vol. 11, p. 193-234, pl. 13. 

LARGUIER DES BANCELS, J- 1912 Le godt et l’odorat. Paris, 8°, vii + 94 pp. 

Lors, J. 1915 Electromotive phenomena and membrane permeability. 
Science, N. S., vol. 42, p. 643-646. 

Lucas, K. 1909 The evolution of animal function, II. Science Progress, 
vol. 3, p. 321-331. 

OsterHouT, W. J. V. 1914 The effect of acid on permeability. Jour. Biol. 
Chem., vol. 19, p. 493-501. 
1915 The nature of mechanical stimulation. Science, N. §., vol. 41, 
p. 175. 

Ricuarps, T. W. 1898 The relation of the taste of acids to their degree of 
dissociation. Amer. Chem. Jour., vol. 20, p. 121-126. 

Spartu, R. A. 1916 The vital equilibrium. Science, N. S., vol. 438, p. 502- 
509. 

Verworn, M. 1913 Irritability. New Haven, 8°, xii + 264 pp. 


THE NERVOUS SYSTEM OF PYCNOGONIDS 


WILLIAM A. HILTON 


Zoological Laboratory of Pomona College, Claremont, Cal. 
TWENTY-ONE FIGURES 


The nervous system of pyenogonids presents many peculiari- 
ties. It is rather difficult to find the counterpart of this sys- 
tem in other arthropods. The nervous system of some Crus- 
tacea suggests it, especially in those forms with an elongated 
thoracic region and reduced abdomen. ‘The general arrange- 
ment of the ganglia is totally unlike the central nervous system 
of arachnids although the general form of the body of ‘sea spiders’ 
strongly suggesis arachnid relationships. The rather small 
supraesophageal ganglion and the well developed chain of ven- 
tral ganglia suggest a rather primitive type of nervous system, 
but the innervation of the pharynx and proboscis presents com- 
plex and apparently unique conditions. 

Although there is an extensive_literature on the classification, 
structure and development of pyenogonids, there is little or 
nothing on the structure of the nervous system. 

The general form of the ganglia with their chief branches 
is quite well known, for nearly every paper on the classifica- 
tion of the group contains a more or less detailed sketch of the 
animals deseribed with the nervous system shown in place. 

The supraesophageal ganglion seems to contain but two pairs 
of ganglia recognized by early authors in other arthropods as 
the protocerebrum and deutocerebrum, the tritocerebrum found 
in some arthropods being absent. This is but one of several 
structures that point to a closer relationship with arachnids 
than with Crustacea. However, without going into further 
reasons at this time, I am inclined to side with Dohrn and con- 
sider Pyenogonida a separate class. 

463 
THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL. 26, NO. 5 
OCTOBER, 1916 


464 WILLIAM A. HILTON 


As the tendency has been to regard these animals as arachnids, 
it may be worth while to glance through the neurological litera- 
ture on this group. 

Among the earliest work on the nervous system of arachnids 
was that of Treviranus in 1812. No hint of pyenogonids is 
given in this paper, nor is there any mention of these animals 
in the work of B. Haller just a century later. There is no refer- 
ence to pyenogonids in the extensive work of Saint Remy, ’90. 
Dahl, in 19138, gives a brief summary of the work of Dohrn 
in connection with various types of arachnids. If we go through 
the extensive literature on the pyenogonids as a group we find, 
it is true, little of the structure of the nervous system, but much 
about the arrangement of the ganglia composing it. 

From the works of Hoek, ’78, ’81, Dohrn ’70, ’81, Sars 791, 
Meinert ’98, and a number of others, as well as from the study 
of Pacific coast forms, we learn that the central nervous sys- 
tem consists of a supraesophageal ganglion and a ventral chain 
of from four to five chief ganglia. The smaller number of gan- 
glia we find when the body is less elongate. The supraesopha- 
geal ganglion has a ventral median nerve to the proboscis, 
nerves to the eyes and a pair to the chelifori. Each ventral 
ganglion has at least one main branch. Three branches from 
the first ventral ganglion are as follows: 1) A small pair or two 
pairs to the proboscis; 2) a pair to the palps; 3) a pair to the 
ovigers; 4) if the first ganglion is fused with the second as it is 
in those with four ganglia, then there is also a pair to the first 
pair of walking legs. 

Figures 1 to 7 show different types of nervous systems from 
Pacific species of pyenogonids. The method by which the 
nervous system was studied by some observers was simply to 
determine the position of the ganglia through the transparent 
body-wall. This was tried with a number of specimens after 
the animals had been fixed in mercuric fluids. In some cases 
the whole animal was stained and mounted in such a way as 
to show the internal ganglia. In some cases the animals to 
be studied were placed for a short time in caustic or acid and 
by one or the other of these methods the internal parts were 


NERVOUS SYSTZM OF PYCNOGONIDS 465 


cleared so that the ganglia might be seen. Serial sections of 
the whole animals were also made for study, but the chitin 
often makes perfect series impossible. Hoek and_ possibly 


} 4 x. 7 - Re = 
Y.  ~eae oR 

YY Pe os 24 sa Sane 
<r me eh Fa RN 

Lae at a agg 
AY 7 " : 

eS fe ™ a 

H ade ah 

A ie 2 3 4 


FL : t sie 
J Tat 
~~ ~~ Le 
a j — 
O_o Lee 
/ \ aie” ee 
Ae wir 


/ se \ a me, 


pes 


Figs. 1 to 7 Drawings from the adult nervous systems of a number of species 
of California pyenogonids. The supraesophageal ganglion is shown at the 
upper end of the figures in every case. The nerves are not all shown in every 
figure. All are shown from the ventral side, the ganglia were exposed by various 
methods and all are not drawn to quite the same scale. 

Fig. 1 Euryecyde spinosa, Hilton. 


Fig. 2. Halosoma viridintestinalis, Cole. 
Fig. 3 Tantystylum intermedium, Cole. 
Fig. 4 Ammothella tubercuata, Cole. 
Fig. 5 Pyenogonium stearnsi, Ives. 
Fig. 6 Palene californiensis, Hall. 

Fig. 7 Anophlodactylus erectus, Cole. 


466 WILLIAM A. HILTON 


some others have used gross dissection with the larger species. 
I also tried this method and found that it was not difficult to 
expose and remove the whole nervous system from even the 
smallest specimens. For the structure of the ganglia serial 
sections were made from these removed ganglia. 

There seems to be some difference of opinion as to number 
and position of abdominal ganglia. There are without doubt 
ganglia in the adult that may be called abdominal, but they are 
often not evident or indicated by shght knobs on the last gan- 
glion. Probably in no case are these little ganglia in the abdo- 
menm(hes: W260, 15 19). 

The special nerve supply to the proboscis has been described 
by Hoek, Dohrn and others. I was able to dissect it out in a 
number of species where I have found essentially the same 
features already described. In the genus Pyenogonum I found 
a similar condition as shown by Hoek. Practically the same 
condition was found in two other genera not before described. 
There are three main branches which run to the three divisions 
of the proboscis: a dorsal branch running from the mid-ventral 
line of the supraesophageal ganglion, and two lateral branches 
springing from the forward part of the first ventral ganglion. 
Each of these branches has numerous small ganglia along its 
course and near the end of each branch there is a much larger 
ganglion. Branches connect the three trunks with each other 
and fine nerves run from each ganglion to adjacent parts of 
the proboscis. Lateral to these three ganglionated branches 
is a more external nerve which sometimes has a separate ori- 
gin from the larger ganglia or from the ganglionated trunks. 
These three more superficial branches appear to fuse in places 
with the deeper branches, but they do not bear ganglia. 

This whole complicated structure seems quite unusual and 
some have seen in this proboscis region the representations of 
other segments of the animals. However I prefer the assump- 
tion of Dohrn that the proboscis represents only a secondary 
growth of the lips of the stomodoeum. I believe the special 
nerves of the proboscis represent the system of frontal nerves 
and ganglia which we find in Insecta and other arthropods. 


NERVOUS SYSTEM OF PYCNOGONIDS 467 


The small gangla of the proboscis are rather new structures, 
but the large, represent the frontal and lateral head ganglia 
of other forms. 

The development of the nervous system of pyenogonids 
has been especially studied by Morgan, ’91, and Meisenheimer, 
03, although a number of others have studied the general life 
histories, or special stages. According to Meisenheimer, in 
the embryos of Ammothea the early development of the ganglia 
is much as in other arthropods, a longitudinal strip or germ band 
enwraps the yolk and paired thickenings of the ectoderm occur 
which represent cerebral and post-oral ganglia. I have not 
followed these earliest stages in any of my material. At the 
time that the free larva is liberated, there is seen a supraeso- 
phageal ganglion and three pairs of sub-intestinal centers such 
as shown in figure 13 of a California form of the same genus. 
The second of these two ganglia is composed of two parts and 
represents the second and third parts fused. This type of larval 
nervous system seems rather typical of this sort of larval form. 
What the changes are from this to the adult are not exactly 
known, but suggestions may be obtained from the study of 
other species. Morgan in Tanystylum gives some idea of the 
gradual development of additional ganglia in the caudal region 
as the larva with three pairs of appendages add a fourth and a 
fifth pair successively and later a sixth pair. At this last period 
the maximum number of ganglia is attained, this number be- 
comes reduced with the growth of the seventh pair of appen- 
dages and the adult condition is reached. During the early 
stages the addition of extra ganglia is probably not so much 
from the backward growth of nervous tissue as from later de- 
velopments from the surface. In Palene the type of develop- 
ment is different because of the large yolk mass. Separate 
ganglia for the segments are developed, each of these has at an 
early period an invagination or ‘ventral organ.’ 

The species whose development I have especially studied 
seems intermediate between the free living larval form and such 
a continuous type of development as shown by Palene. This 
genus Anophlodactylus is more parasitic during larval stages 


468 WILLIAM A. HILTON 


Fig. 8 Central nervous system of A. erectus during first larval stage. X 350. 

Fig. 9 Nervous system of A. erectus during the second larval period. X 350. 

Fig. 10 Section through whole larva of A. erectus during the third larval 
period. X 350. 

Fig. 11 Central nervous system of A. erectus at about the third larval stage. 
Drawn from a whole mount which did not show as much as some others. X 300. 

Fig. 12 Longitudinal section through the central nervous system of A. erectus 
during the last larval stage. » 350. 


Fig. 13 The central nervous system of the first larval stage of Ammothella. 
x 350. 


NERVOUS SYSTEM OF PYCNOGONIDS 469 


than the others mentioned. The first larval stage soon attaches 
itself to, and enters hydroids. It has three appendages in the 
first larval stage, one pair is chelate, the last two have long ten- 
dril-like extensions. At such a period the nervous system is 
not easily made out from surface views, but it is much like that 
of Ammothella. Figure 8 shows three parts, a larger thicker 
portion which has nerves to the larger first appendages, and 
on each side back of this a group of cells corresponding to the 
other appendages. A moult within the hydroid gives rise to 
a small larval form without the long appendages and it is at 


~ nN 


= ae ei, 
J eee a 
» i Wales 
‘ " S - = 
i = Preeti Af } 

f \ eae Nee bY \ 

i i vo } 

“ koe OS 

J ‘ = J } 

¢ if pees / 

i ¢ y 
} pS MS 
14 
“A aa 7 i “A 
15 
AS ‘ 7 8 § 

i ee | 
fe hers nee ep” 
} f IO) \ er wae : a 

i aR a] \ Ge = > 

j \ { ~ Ge ‘ 

\ \ a i 8 
/ w° 62 « 
f \ SS JIS ¥ 
t } Qe 
1 / fi 

( | ie 

Ses Pi 17 18 
{ ya 
Nee ) eee 4 
16 


Fig. 14 Outline of ventral view of larva and ganglia from below A. erectus, 
third larval stage. X 85. 

Fig. 15 Outline of a ventral view of a later stage larva than figure 14 of A. 
erectus. X 85. 

Fig. 16 Outline of a dorsal view of a larva of A. erectus about the same stage 
as figure 14. The brainisshown. X 80. 

Fig. 17 Fourth stage larva of A. erectus from below. X 35. 

Fig. 18 Central ganglia of a larva of A. erectus with three pairs of walking 
legs. The drawing is from below. The upper area without nerves in the figure 
is the supraesophageal ganglion. X 36. 


470 WILLIAM A. HILTON 


such a period that new ganglionic material seems to be developed. 
Figures 9, 11 and 14 are drawn from such early stages. At 
a later moult more ganglia are evident, as in figures 10, 15, and 
16. The ventral ganglia at first are mere groups of cells, as 
is Shown in the frontal section from which figure 10 was taken. 
As may be seen from the figures 10 and 15, ganglia are devel- 
oped in each segment, a pair for each appendage and several 
for the cephalic region and a common mass of cells for the ab- 


DO, 
eto 08 


es) 
ats 
Shut 
oan 
SOL OO? 
OSI ey a), 2) 
eee oe 2! 
porn. 
a 3.58 


Fig. 19 A longitudinal section through the central ganglia of Lecythorhynchus 
marginatus, Cole. Two small abdominal ganglia show at the end of the last 
thoracic ganglion. 35. 

Fig. 20 A longitudinal section through the supraesophageal ganglion of L. 
marginatus. The dorsal side is up, the cephalic side to the right. X 210. 


dominal. In a stage just before this there are two pairs of 
ganglia on the dorsal side of the larva; these are shown in fig- 
ure 16. They represent the brain. 

At about the third moult, as shown in figures 12 and 17, the 
ganglia have developed central fibers, but still show their paired 
nature. There seems to be some indication of more ganglia 
than there are appendages, some of the caudal elements may not 
be evident in later stages, and the first ventral ganglion seems 
composed of two small pairs of elements. In the proboscis 


Fig. 21. Drawing of the nerves and ganglia of the proboscis of L. marginatus. 
Slightly diagrammatic. No structures shown in the proboscis but nerves and 
ganglia. The drawing was made by Miss M. L. Moles from the first sketch taken 
from the dissection. Much enlarged. 


471 


472 WILLIAM A. HILTON 


of this stage there seem to be two small pairs of ganglia. The 
dorsal ganglia are not shown in figure 17. 

When the larvae moult again and leave the cavity of the 
hydroids they have all but one pair of legs. Figure 18 shows 
the whole central nervous system from below at such a stage. 
The brain above the esophagus is at the upper end of the figure, 
then follow the ventral ganglia, seven paired masses and a small 
unpaired caudal ganglion. There is a gradual fusion of these 
ganglia until the adult condition shown in figure 7 is attained. 

The structure of the adult nervous system of pycnogonids 
is quite simple. There is the same general arrangement of 
cells that we find in other arthropods. The ventral ganglia 
have few cells on the dorsal side, but many on the lateral and 
ventral sides. The supraesophageal ganglion is sheathed in 
cells on the lateral and dorsal sides. Nerve fibers connect the 
ganglia and certain regions but in no place is there a concen- 
tration of the fibers. The fibrous mass is not particularly 
dense at any point. There do not seem to be many long tracts 
and the supraesophageal ganglion is not more complicated 
than other parts so far as could be determined. There are 
no marked decussations of nerve fibers and the nerve cells pre- 
sent a uniform appearance. Among the nerve cells are many 
nuclei of neuroglia networks which form the framework of the 
ganglia especially in the area of the cells. 

Although there are indications of special groups of cells and 
fibers, there was no indications of mushroom bodies. 

The animals do not seem to have a special brain. The su- 
praesophageal ganglion is not a very special center. The move- 
ments of the animals agree with this; they move sideways, 
forwards or backwards when stimulated. No part of the body 
seems to lead in the locomotion. 


NERVOUS SYSTEM OF PYCNOGONIDS 473 


BIBLIOGRAPHY 


ALDERZ, G. 1888 Bidrag till Pantopodernas Morphologi och Utvecklings 
historia. Bihang till k. Svenska Vetenskap Akad. Hand., Bd. 13, 
afd 4, no 11, Stockholm. 

Dau, F. 1913 Vergleich. Phys. und Morph. der Spinnentiere, Erster Teil, 
Jena. 

Dourn, A. 1870 Untersuch. tiber Bau und Entw. der Arthropoden. 2. Pycno- 
goniden. Jen. Zeit. f. Nat., Bd. 5. 

1881 Die Pantopoden des Golfes von Neapel. Fauna und Flora des 
Golfes von Neapel. Monog. 3, Leipzig. 

Hatter, B. 1912 Uber das Zentralnervensystem des Scorpions und der Spinnen. 
Arch. f. Micr. Anat.. Bd. 79, Abt. 1. 

Hauwez, P. 1905 Observations sue le parasitisme des larves de Phoxichilidium 
chez Bouganvilla. Arch. zool. exp. et gen., 4me serie, t. 3. 

Hittron, W. A. 1916 The life-history of Anoplodactylus erectus Cole. Jour. 
ent. and zool., vol. 8, no 1, March. 

Hoek, P. P.C. 1881 Report on the pyenogonida. Voyage of H. M. 8. Challen- 
ger, Zoology, vol. 3. 

1881 Nouvelles études sur les pyenogonids. Arch. Zool. exp., t. 9, 
Paris. 

Hopae, G. 1862 Observations on a species of pyenogon (Phoxichilidium coc- 
cineum) with an attempt to explain the order of its development. 
Ann. mag. nat. hist. (3), vol. 9. 

Kroyer, H. 1842 Notes sur les metamorphoses des pyenogonides. Ann. 
sc. nat: Ze ser., t. 17. 

LENDENFELD, R. von 1883 Die Larvenentwicklung von Phoxichilidium plumu- 
larie. Zeit. Wiss. Zool., Bd. 38. 

MEISENHEIMER, J. 1902 Beitrage zur Entwick. der Pantopoden. Zeit. f. 
Wiss. Zool., Bd. 72. 

Mertens, H. 1906 Eine auf Tethys leporina, parasitisch lebende Pantopoden 
larva (Nymphon parasitica n. sp.). Mitt. aus der Zool. Stat. Neapel, 
Bd. 18. 

Meinert, Fr. 1898 Pyecnogonida den Daniske Ingolf-expedition. 

Morgan, T. H. 1891 Contribution to the embryology and phylogeny of pycno- 
gonids. Stud. biol. lab. Johns Hopkins Univ., vol. 5. 


EVIDENCE OF A MOTOR PALLIUM IN THE FORE- 
BRAIN OF REPTILES! 


J. B. JOHNSTON 
University of Minnesota 


ONE FIGURE 


At the title indicates, the purpose of this note is not to describe 
the structure of the reptilian motor cortex nor to discuss the 
localization of function within this area. ‘The purpose is only 
to give such evidence as is now at hand that a specialized area 
comparable to the mammalian motor cortex probably exists 
in the reptilian brain and to indicate the general position and 
extent of this area. 

The writer has indicated in a previous paper (‘15 b) that the 
rostral portion of the dorsal pallium in the turtle differs struetur- 
ally from the rest and has suggested the possibility that this 
area may correspond to the motor cortical area in the mammalian 
brain. 

This hypothesis is being tested by degeneration methods and 
by cortical stimulation. Sufficiently definite results have been 
obtained by the latter method to indicate that motor and sensory 
fields can be distinguished in the reptilian pallium. 

Three species of turtle (Chelydra serpentina, Cistudo carolina, 
and Chrysemys marginata) and one lizard (Gerrhonotus) have 
been studied. Two methods of stimulation were used: induction 
shocks with two-point electrodes, the two points closely approxi- 
mated; induction shocks with one point electrode, the second 
electrode being formed by a copper plate covered with moist 
cloth on which the body of the animal rested. ; 

An important factor in the experiments is the degree of anaes- 
thesia employed. The clearest results have been obtained with 
very deep anaesthesia. Some of the turtles were given a dose 
of morphine by hypodermic injection, and all were anaesthet- 


1 Neurological Studies, University of Minnesota, no. 22, June 29, 1916. 
475 


476 J. B. JOHNSTON 


ized by means of chloroform placed on cotton in the mouth or 
injected directly into the trachea or both. Although the resist- 
ance of these animals was well known, in several cases the 
anaesthesia was not carried far enough and the reactions to 
cortical stimulation were prolonged contractions of neck muscles 
or struggling movements of the limbs indicative of pain. In 
these cases occasional responses, as of the eye muscles or temporal 
muscle, appeared to be due to local excitation of the cortex but 
the more extensive and prolonged movements were elicited from 
any part of the cortex and were not distinguishable from the 
responses to stimulation of the dura mater, the lower brain 
regions or even the tissues exposed in the head. Under con- 
ditions of deep anaesthesia the responses consisted of contraction 
of a small set of muscles, and of short duration, and these were 
obtained from a certain region of the pallium only. The lizards 
are not so difficult to anaesthetize, but long continued appli- 
cation of chloroform is necessary (one and a half to two hours 
in a closed dish containing a wad of cotton wet in chloroform). 

Another factor presenting difficulty is the small size of the 
brain in these animals. This makes it necessary to use a single 
point electrode or to have the two points very close together. 
It is of the greatest importance that the meninges be carefully 
removed from the whole hemisphere and that the brain and 
surrounding tissues be kept as dry as possible without injury 
to the brain tissue. If the electrode comes too close to the dura 
mater or if fluid facilitates the spread of the current to the dura, 
the muscles, thalamus or midbrain, the results are vitiated 
because of responses coming from two or more sources of stimu- 
lation. The responses which follow stimulation of the dura or 
brain stem are very different from those produced by cortical 
excitation, being usually more vigorous and prolonged as well 
as more extensive. In the lizards and in several turtles, after 
the dura was laid back and the olfactory and optic connections 
were cut, the entire forebrain was raised out of the skull so as 
to be free from contact with other tissues. Under these con- 
ditions the entire surface of the hemisphere was explored with 
the electrodes. 


MOTOR PALLIUM IN REPTILES 477 


With the induced current minimal stimuli were applied at 
first and increased later in the experiment when it became 
necessary in order to obtain responses. In many cases the 
brain fatigued quickly and in some it recovered with equal 
promptness. Individuals differed greatly as to the recovery. 
Often when the strength of stimulus was increased struggling 
movements resulted and the experiment had to be discontinued. 
In some eases also, the animals gave definitely localized responses 
to the first stimulation and later, probably because of recovery 
from the anaesthesia, gave irregular and prolonged reactions. 


CORTICAL REACTIONS IN TURTLES 


If we take into account only the short contraction of restricted 
groups of muscles, the following parts of the hemisphere have 
been found to give muscular responses: dorsal surface of the 
olfactory bulb, retraction of the neck, extension of the legs, 
movements of eyeball and eyelid; dorsal surface of pallium near 
olfactory peduncle and lateral border of pallium in the anterior 
one-half or two-thirds of hemisphere, movements of eyes, jaw, 
neck, legs and tail; striatal area, movements of all parts. No 
responses were obtained from any other part of the dorsal sur- 
face or the medial wall or from the tuberculum olfactorium, the 
amygdaloid eminence or elsewhere, except that contractions 
typical of thalamic stimulation were sometimes obtained when 
the caudal pole or amygdaloid eminence was being explored. 
These were probably due to spread of current to the thalamus. 

The responses from the striatal area were presumably due to 
the direct stimulation of descending fibers in the crus. ‘The 
responses from the olfactory bulb may be due to the close proxim- 
ity of the ‘motor area’ which is indeed overlapped by the caudal 
border of the olfactory formation. Thus it appears that a 
somewhat comma-shaped area involving the rostral and lateral 
borders of the general pallium (fig. 1) may be regarded as a 
motor area in the turtle’s pallium. This area corresponds roughly 
to the ‘pallial thickening’ described in a previous paper (715 b). 

Not enough experiments have been conducted under uniform 
conditions to furnish a basis for the discussion of localization 


478 J. B. JOHNSTON 


within this area. As the whole area is only a few millimeters in 
extent, it is obvious that great care must be exercised in accepting 
as conclusive any apparent indications of special function of 
particular parts of this area. For this reason I mention only 
tentatively and with reserve that leg movements have been 
obtained most often from the anterior part of this area and eye, 
jaw and neck movements from the lateral portion. Further 
experiments will be made to determine whether a localization 
within the motor area clearly exists. In certain experiments 


Fig. 1 Sketches of the dorsal surface of the right hemisphere of the turtle 
(left) and lizard (right) on which the motor area is shaded. 


the eye and jaw movements were homolateral only or more 
frequently, while the neck and leg movements were mostly 
heterolateral. 

Excitation of the dorsal ventricular ridge in the turtle was 
earried out by cutting away the whole dorsal pallium and apply- 
ing the electrodes to the ventricular surface of the ridge. _Muscu- 
lar responses were often obtained, such as retraction of the eye- 
ball, contraction of temporal muscle and twitching of fore leg, 
but there is the double danger of spread of stimulus to the 
closely adjacent pallial thickening and of stimulation of the 
underlying crus. 


Lo) 


MOTOR PALLIUM IN REPTILES 47 
MOTOR RESPONSES IN THE LIZARD 


Only four animals have been examined, two of which gave 
much clearer and more definite results than those obtained from 
most of the turtles. From the area shown in the accompany- 
ing sketch one animal gave movements of the fore leg, jaw and 
eye muscles, neck and throat muscles, and anterior body muscula- 
ture. Stimulation of the olfactory tubercle also caused contrac- 
tion of small muscles in the throat, which was felt as the head 
was held in the fingers. Stimulation of the crus in the striatal 
area produced movements of legs and body muscles. Move- 
ments of the fore leg were regularly heterolateral, although 
homolateral movements occurred occasionally. Just the op- 
posite was true of the eye and jaw muscles. The other animal 
responded to stimulation of the anterior part of this area by 
strong contraction of pelvic muscles and movements of the hind 
legs and by weak movements of the fore legs. Stimulation of 
the lateral border of the right pallium in this animal produced 
a definite torsion of the fore part of the body. 

Gerrhonotus is a rather large lizard and its brain is fairly 
satisfactory to work with. The results in these two cases were 
so definite that I regard them as strong evidence that a motor 
area exists in the lizard pallium essentially similar in position 
and extent to that in the turtle. 

Early in my studies I examined the brain of a single specimen 
of Alligator about a foot and a half in length, and obtained 
motor responses from the anterior part of the dorsal pallium, 
but I have no written record of this experiment. 

From these experiments, together with the study of the struc- 
ture of the hemisphere already reported, I am strongly inclined 
to believe not only that reptiles possess a general or somatic 
pallium which has been in dispute until recently, but also that 
in that pallium are to be distinguished definite sensory and motor 
areas in the sense in which those terms are commonly used of 
the mammalian pallium. Localization within these areas and 
the significance of these areas in the evolution of the mammalian 
pallium are subjects for further study. 


THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL. 26, NO. 5 


nay 


ek jee Spe 
eg,” 


THE DEVELOPMENT OF THE DORSAL VENTRIC- 
ULAR RIDGE IN TURTLES! 


J. B. JOHNSTON 
Universily of Minnesota 
TWENTY-SEVEN FIGURES 


In the adult turtle (Johnston, 15 b) the large ridge project- 
ing into the ventricle from the lateral wall of the hemisphere 
has strong fiber connections with the cerebral peduncle and 
has obvious close relations with the cortical layers of the pallium 
in the caudal pole. Indeed, in its caudal part this ridge appears 
to be a fold of the whole thickness of the brain wall and is 
marked externally by the ‘amygdaloid fissure.” Here the cell 
layers of the general pallium turn in to become continuous with 
the cell masses of the ridge. It is evident from these facts that 
the relations of the dorsal ventricular ridge are chiefly with the 
general pallium. Its position lateral and dorsal to both the 
caudate and lentiform nuclei shows that it does not belong 
to the corpus striatum. Its independence from other cell 
masses in its rostral part gives one the impression (from the 
study of adult material alone), that the ridge has been formed 
by infolding from the caudal region (15 b, p. 417). 

The origin of this ridge in the embryo should throw light 
not only on its relations in the turtle but also on its significance 
in the evolution of the mammalian brain. I am able to give 
the main outlines of the development on the basis of a series 
of early embryos of Chelydra serpentina, for which I am in- 
debted to Dr. C. E. Johnson of this University; and of sec- 
tions of 17, 20, and 28 mm. embryos of Chrysemys, belonging 
to the Anatomical Laboratory of Washington University for 
the loan of which I am indebted to Dr. Edwin A. Baumgartner. 


' Neurological Studies, University of Minnesota. No. 23, July 1, 1916. 
481 


482 J. B. JOHNSTON 


The material in both cases was beautifully sectioned and stained. 
The brains of freshwater turtles are so similar that the younger 
stages of Chelydra can be readily compared with the later stages 
of Chrysemys for the rather general purposes of this study. 


AREAS OF PROLIFERATION IN THE EARLY EMBRYO IN RELATION 
TO CELL MASSES OF THE ADULT BRAIN 


Broadly speaking, three periods are to be recognized in the 
development of cell masses in the telencephalon: a period of 
evagination of the hemispheres and formation of indifferent 
cells by mitosis of the germinal cells; a period of form changes 
and further proliferation of cells in which the successive forma- 
tion of functionally ind pendent cell masses is noticeable; and 
a period during which the growth of fiber bundles and other 
factors bring about the definitive form of the adult cell masses. 

In the first period the wall of the telencephalon consists of 
a deep layer of spongioblasts and dividing germinal cells, a 
thick layer of indifferent cells and a superficial clear layer or 
marginal veil of very varying thickness. The thickest part 
of the marginal veil forms the floor of the groove along the 
lower border of the hemisphere, at its junction with the brain 
stem. This is the course of the future lateral forebrain bundle 
or crus cerebri. The indifferent cells form a continuous layer 
with no indication of individual masses. 

The second period begins with the appearance of independent 
cell masses separated by cell-free zones which show in sections 
as clear lines or areas. Such cell-free zones in adult brains have 
been commonly regarded as boundary lines between individual 
cell masses which are presumed to have different functions; 
i1.e., centers or nuclei. Similar cell-free zones in the brains of 
lower vertebrates have been used in recent years as landmarks 
between important morphological regions, as the pallial and 
basal areas. The writer has pointed out the need of caution 
in interpreting these zones in this connection (713 b, p. 382). 
It is now to be noted that the formation of individual masses 
of cells is dependent upon the proliferation of cells from differ- 
ent regions of the germinal cell layer and upon the proliferation 


DORSAL VENTRICULAR RIDGE 483 


of cells at successive periods of development. Thus the pro- 
liferation of cells in the dorsal wall gives rise to pallium, that in 
the basal wall gives rise to various olfactory centers, corpus 
striatum, ete. Also, certain centers owe their origin to the 
proliferation of cells in an early stage, other centers to prolifera- 
tion in later stages. The former come to lie superficial to the 
latter and the two are separated by cell-free zones. A further 
factor is the shifting or spreading of cell masses, whether due 
to cell-migration or to mechanical forces. These several fac- 
tors must be taken into account in any attempt to use the em- 
bryological method in the study of the significance of cell-group- 
ing in the brain. 


CHELYDRA SERPENTINA 


In the oldest Chelydra embryo studied, it is possible to recog- 
nize several of the important cell masses and fiber bundles of 
the telencephalon and in part the boundary lines between pal- 
lial and basal areas. Examination of the model (figs. 1, 2) and 
of sections (figs. 6 to 14) shows that the telencephalon and dien- 
cephalon are narrow from side to side and high dorso-ventrally. 
The hemisphere is a simple sac which projects much farther 
rostrad than caudad. The interventricular foramen is. still 
very large. Behind it, the thin wall which represents the choroid 
plexus extends nearly half way to the caudal pole. This stage 
illustrates very clearly, both in its general form and its inter- 
nal structure, how much the rostral pole of the hemisphere 
precedes the caudal pole in development. The outer surface 
of the brain shows few of the landmarks which are seen in later 
embryos and adults. The stem-hemisphere sulcus is of course 
the most prominent. The olfactory peduncle is not yet formed, 
but the broad depression on the lateral surface may represent 
the beginning of constriction. The fissura prima and diagonal 
band are recognizable at the medio-basal angle in front of the 
preoptic recess. There is a shallow furrow over the rostral 
part of the foramen which may be the beginning of the sulcus 
fimbrio-dentatus. 


484 J. B. JOHNSTON 


The ventricle shows a very sharp ventral groove leading 
forward from the foramen and much deepened in the region 
of the future tuberculum olfactorium. The middle ventricular 
groove is clearly marked in the region rostral to the foramen, 
while the dorsal groove has not yet appeared. 

In spite of the simple form of this brain, the internal structure 
shows considerable advance in differentiation. 

The sections illustrated in figures 6 to 14 can be understood 
best by beginning with that which passes through the rostral 
border of the foramen interventriculare (fig. 9). In the medial 
wall above the foramen are recognized the primordium hippo- 
campi and the fimbria, one of the early fiber tracts to be formed 
in the embryo. The dorsal wall is occupied by pallium, the 
medial portion by hippocampal, the lateral portion by general 
pallium. In the lateral wall the pallium seems to be limited 
by a ventricular groove and by a cell-free space. This bound- 
ary line strongly reminds one of the lateral zona limitans of 
the selachian and freg brains (Johnston, ’lla). The later 
embryos show this groove to be the middle ventricular groove 
of the adult brain and the clear area becomes the cell-free zone 
separating the dorsal ventricular ridge and the nucleus lenti- 
formis in the adult. In the lower half of the lateral wall the 
following structures are seen in the figure: an active prolifera- 
tion which gives rise to the nucleus lentiformis, the nucleus 
caudatus, the lateral and medial forebrain bundles, the anterior 
commissure bundle, the fiber layer of the diagonal band, and 
the lateral olfactory area. Although I have not sufficient stages 
to enable me to follow the history of these structures, there is 
evidence that the superficial cell layer which represents the 
lateral olfactory nucleus has been derived from the layers of 
indifferent cells in the early embryo and that the mass of the 
caudate nucleus represents a later proliferation. ‘The embryo 
shows clearly that the lentiform nucleus is beginning to form 
when the caudate is already completely formed or nearly so. 
As the other sections are studied it will become evident that the 
lensiform nucleus is formed by a proliferation distinctly dorsal 
to the caudate and that the changes in later development bring 


DORSAL VENTRICULAR RIDGE 485 


it into a position lateral to the caudate. There is some indi- 
cation of the superficial layer of the olfactory area spreading 
upward beyond the middle ventricular groove (at the point 
marked l.pyr). 

Passing caudad from this level, in the next section drawn 
(fig. 8) the same structures are to be seen in the lateral wall. 
The middle ventricular groove has almost disappeared, however, 
and the proliferating area for the lentiform nucleus is much 
smaller. The next section drawn (fig. 7) is the last one in which 
the lentiform proliferation can be recognized and with it the 
locus of the pallial border. In the lower part of the hemisphere 
the lateral olfactory area is represented by a large collection 
of cells closely related to the caudate nucleus and the diagonal 
band. This is the nucleus of the lateral olfactory tract and the 
next section shows the close relation of both this and the diago- 
nal band to the stria medullaris. Caudal to this the nucleus 
of the lateral olfactory tract soon disappears and the hemisphere 
wall presents the appearance of an undifferentiated pallium, 
except for the fimbria and the thin choroidal area in the medial 
wall. 

The next section rostrad from the one first described passes 
through the anterior commissure (fig. 10). The lentiform pro- 
liferation appears larger, the diagonal band fibers approach 
those of the olfacto-hypothalamic system. In figures 9, 10, 
and those of sections farther rostrad, one of the most noteworthy 
features is the mass of cells proliferating from the deep layers 
of the pallum. The medial border of the pallium does not 
show active proliferation but in the lateral three-fourths many 
new cells are forming and these appear to be streaming laterad 
into the thick wall just dorsal to the middle ventricular groove. 
Indeed the proliferation and streaming of these cells is the active 
cause for the thickening of this part of the wall and the forma- 
tion of the middle ventricular groove. Further thickening of 
this part of the wall deepens this groove and produces the dor- 
sal ventricular groove at the point indicated in figure 9. 

In figures 9 to 12 appear important relations of the dorsal 
border of the olfactory area to the pallium. Here it is quite 


486 J. B. JOHNSTON 


clear that the superficial layer of cells of the olfactory area ex- 
tends up on the outer surface of the pallium and that it is ac- 
companied by a special bundle of fibers. The crowding of 
cells in this stage is such that 1t is impossible to determine whether 
the pyriform lobe cells have spread or migrated up from the 
lateral surface in the basal region or have been formed here by 
the proliferation which at this stage is giving rise to the pal- 
lium. Farther caudad and in later stages the relations are such 
as to make the former appear more probable. The cells of the 
pyriform lobe probably all come from the layer of indifferent 
cells seen in earlier stages and formed by proliferation in the 
lateral and basal region before the pallial proliferation seen 
in this stage began. Pallium and pyriform lobe are still visible 
in a section (fig. 14) through the peduncle in which the olfactory 
formation appears in the dorsal wall. The olfactory formation 
extends far back on the dorsal surface, as is already well known. 


CHRYSEMYS 


Models were made of the 17 mm. and 28 mm. stages. Most 
of the chief landmarks of the adult brain are already visible 
in the 17 mm. embryo (figs. 3, 4, 15, 16, 17). The caudal pole 
still lags behind the rostral portion in development. ‘The cau- 
dal pole projects much farther back than in the oldest Chelydra 
embryo and in the 28 mm. stage the olfactory area (lobus pyri- 
formis) extends relatively farther back than in the 17 mm. 
stage. The tuberculum olfactorium is well developed and 
the ventral groove of the ventricle dips deep into it. The mid- 
dle ventricular groove (fig. 4 and sections) is very sharp and 
deep in the middle part of the brain and stops abruptly just 
caudal to the level of the foramen interventriculare. The dor- 
sal ventricular groove appears in the 17 mm. embryo (figs. 4, 17) 
as a slight groove in the lateral wall some distance from the 
medio-dorsal angle of the ventricle. 

Three sections caudal to the foramen are drawn from the 
17 mm. embryo and nine sections from various leve's of the 
28 mm. embryo. The internal differentiation has already pro- 
gressed so far in the 17 mm. embryo that it will be unnecessary 


DORSAL VENTRICULAR RIDGE 487 


to speak of it separately. In both stages many features of 
the sections can readily be compared with the structures seen 
in corresponding sections of the adult brain. For this pur- 
pose the reader should consult the description of the cell masses 
in the brain of Cistudo in this Journal for October, 1915. The 
necessary description of individual sections is given in connec- 
tion with the figures. Here I shall take up the relationships 
of the lateral margin of the pallium and the dorsal ventricular 
ridge. 

Corpus striatum and lateral olfactory area. In the younger 
embryos the loosely arranged cells of the lateral olfactory cen- 
ters cover the outer surface of the caudate and lentiform nuclei 
and form a continuous layer over the lateral surface of the 
hemisphere (except dorso-caudally). This is shown in figures 
7 to 12, and this interpretation is given to the models. In the 
28 mm. embryo there is a small area opposite the foramen in 
which the striatum and the crus bundle contained in it come 
to the surface (fig. 21) and the olfactory centers form the pyri- 
form lobe above and the diagonal band below. This small 
area is painted blue in the model and appears white in figure 5. 
Sections show that it is the elbow of the crus, with certain large 
cells contained in it ('15 b, p. 405), which comes to the surface 
here. This is presumably due to the rapid growth of both 
ascending and descending fibers in the crus. The crowding 
thus produced leads to the shifting of the cells of the olfactory 
center upward and downward. This process must go much 
further in later development until the large striatal area is left 
exposed in the adult. The embryonic history thus seems to 
support very clearly the hypothesis (15 b, p. 428, 429) that 
the exposed striatal area in the turtle had been covered at an 
earlier stage of evolution by the lateral olfactory area. 

Relations of pyriform lobe and pallial margin. In the caudal 
half of the hemisphere the pyriform lobe forms a layer external 
to the corpus striatum and extends up to a thin edge superficial 
to the dorsal ventricular ridge and the margin of the pallium. 
This overlapping of the pallial margin by the pyriform lobe 
persists in the adult. At about the level of the interventricular 


488 J. B. JOHNSTON 


foramen the pyriform lobe begins to extend higher on the lateral 
surface. At the same time thickening of the lateral border 
of the pallium moves dorsally, becomes crowded and folded 
on itself and sinks in, forming a projection into the ventricle 
(figs. 22 to 25). These changes take place rapidly from be- 
hind forward so that in thirty or forty sections of 10 microns 
the place occupied by the thick border of the pallium comes 
to be taken by the pyriform lobe. From this point forward 
the pyriform lobe forms a characteristic ridge filled with a dense 
layer of cells, very much as in the adult. 

Origin of dorsal ventricular ridge. In the Chelydra embryo 
above described the dorsal part of the lateral wall is thickened 
by cells coming from a proliferating area in the dorso-lateral 
wall. The groove which bounds this thickening was identified 
with the middle ventricular groove of the adult. In the older 
embryos of Chysemys the groove is readily identified but has 
been very greatly deepened by further thickening of the mass 
just above it. Likewise the thickening of this mass has pro- 
duced a groove above it, the dorsal ventricular groove. The 
thick mass, then, is the dorsal ventricular ridge and in earlier 
stages it is indistinguishable from the dorsal pallium. In the 
caudal part of the hemisphere of the older embryos the relations 
of the dorsal ventricular ridge are not essentially changed from 
those seen in the Chelydra embryo. Caudal to the level of 
the stria medullaris (figs. 18, 19) this ridge bulges into the ven- 
tricle and is covered outside and below by the thick mass of 
the nucleus of the lateral olfactory tract. In the caudal pole 
the ridge merges gradually with the general pallium caudal to 
the level at which the pyriform lobe ends In the rostral part 
of the hemisphere (figs. 22, 23) the dorsal ridge has become prac- 
tically independent of the pallial thickening, very much as it is 
in the adult. From these embryos it is quite clear that the 
dorsal ventricular ridge arises as a thickening of the lateral 
border of the pallium and that its continued growth causes 
is to project into the ventricle and produces the middle and dor- 
sal ventricular grooves. In the 28 mm. embryo the mass of 
cells which in the adult was called the core-nucleus of the ridge 


DORSAL VENTRICULAR RIDGE 489 


is clearly recognizable in the rostral part, back to the level of 
the caudal border of the foramen interventriculare. 

Between the pallial thickening and the cells of the dorsal 
ventricular ridge appears in the sections a clear space filled by 
fibers. In the caudal part of the hemisphere, where the dorsal 
ventricular groove is already formed, this bundle of fibers is 
seen to hold the same position as an important bundle in the 
adult. In the adult many fibers from the crus enter this bundle 
or, more probably, pass through it on their way to the pallium. 
Here in the embryo the crus is not yet sufficiently developed 
to enable one to trace it up through the striatum to this level. 
The strong development at this stage of the bundle here men- 
tioned indicates that it is composed chiefly of fibers connecting 
parts of the hemisphere; an association bundle, in other words. 

Relations of dorsal ventricular ridge to basal structures. In 
the rostral part of the hemisphere the ridge is from the first 
clearly marked off from the striatum below by the middle groove 
and by a prominent cell-free zone. In later embryos this cell- 
free zone becomes compressed into a narrower line as seen in 
section and scattering cells are found in it. There is, however, 
always a clear boundary between the ridge and the developing 
lentiform nucleus. In section the cell-free zone inclines dorsad 
toward the lateral surface and passes above the lateral olfac- 
tory area. In the rostral part, where the pyriform lobe crowds 
far dorsal this limiting zone is bent into a semicircular or U- 
shaped line (figs. 28 to 27). When the writer first studied 
the adult turtle brain he found it impossible to establish a direct 
comparison between the lateral zona limitans of the selachian, 
frog and other simpler vertebrates, and either of the cell-free 
zones seen in the lateral wall of the turtle brain. It is now 
evident that the clear zone which in the adult brain passes 
from a point near the middle ventricular groove toward the 
lateral surface and then bends dorsad to reach the surface in 
the sulcus rhinalis above the lobus pyriformis, is the zona limitans 
of the embryo and corresponds to the zona limitans lateralis in 
the selachian and frog. (See Johnston, 715 b, figs. 19 to 22 and 
compare with ’1la, fig. 75). Since the relations of the pyri- 


490 J. B. JOHNSTON 


form lobe, corpus striatum and pallium in the turtle are readily 
comparable with those in the mammal, the identification of 
the lateral zona limitans in the turtle completes the history of 
this important landmark throughout the series of vertebrates. 
The boundary lines between the pallial and basal portions of 
the hemisphere are now clear in both medial and lateral walls 
in fishes, amphibians, reptiles and mammals (Johnston, ’11 a, 
Altovlon Ilkley 

The relation of the dorsal ventricular ridge to the amyg- 
daloid complex requires further comparative study. In the 
adult, it will be remembered, the caudal part of this ridge is 
separated from the rest by a broad shallow groove which con- 
tinues in the general direction of the deep middle ventricular 
groove (15), figs. 4 and 10). This caudal portion of the ridge 
is made up of the medial amygdaloid nucleus and of an apparent 
infolding of the general pallium. Even in the oldest of these 
embryos the dorsal ventricular ridge extends into the tip of 
the caudal pole as a simple thickening of the general pallium. 
It seems probable, although the evidence is not sufficient for 
a definite conclusion, that the caudal portion of the ridge in- 
cluding the large-celled medial amygdaloid nucleus is formed 
from this thickening of general pallium. 

For the photographs of the models I am indebted to Dr. W. F. 
Allen of the University of Oregon. 


LITERATURE CITED 


JounsTon, J. B. 191l1a The telencephalon of selachians. Jour. Comp. Neur., 
vol. 21. 
1913 b The morphology of the septum, hippocampus, and _ pallial 
commissures in reptiles and mammals. Jour. Comp. Neur., vol. 23. 
1915 b The cell masses in the forebrain of the turtle, Cistudo carolina. 
Jour. Comp. Neur., vol. 25. 


DORSAI VENTRICULAR RIDGE 


491 


FIGURES 


ABBREVIATIONS 


a.p., area parolfactoria 

b.0., bulbus olfactorius 

c.a., commissura anterior 

c.h., commissura hippocampi 

ch. op., chiasma opticum 

c.p.a., commissura pallii anterior 

crus, crus cerebri or lateral forebrain 
bundle 

c.st., corpus striatum 

d.b., diagonal band of Broca 

d.v.r., dorsal ventricular ridge 

fi., fimbria 

f.o., formatio olfactoria 

f.p., fissura prima 

for.i., foramen interventriculare 

f.rh., fissura rhinalis 

g.p., general pallium 

g.s., gyrus subcallosus 

h., hippocampus 

hy., hypothalamus 

l. pyr., lobus pyriformis 

l.t., lamina terminalis 

m.fb.bdl., medial forebrain bundle 


n.c., nucleus caudatus 

n.d.b., nucleus of the diagonal band 

n.l., nucleus lentiformis 

n.o., nervus olfactorius 

n.rot., nucleus rotundus 

n.tr.olf.lat., nucleus of the lateral ol- 
factory tract 

pa., pallium 

pa.th., pallial thickening 

p.h., primordium hippocampi 

p.o., pedunculus olfactorius 

r.p., recessus praeopticus 

r.s., recessus superior 

s.f-d., sulcus fimbrio-dentatus 

s.m., stria medullaris 

s.v.d., dorsal ventricular sulcus 

s.v.m., middle ventricular sulcus 

s.v.v., ventral ventricular sulcus 

thal., thalamus 

t.o., tuberculum olfactorium 

tr.olf-hy., tractus olfacto-hypothalam- 
icus 

tr.op., tractus opticus 


492 J. B. JOHNSTON 


Fig. 1 Chelydra serpentina. The size of this embryo is indicated by the 
owner by the words ‘8.5 mm. carapace.’ Lateral view of a model of-the right 
hemisphere. Description in the text. 

This and the following models have been merely rubbed with a blunt instru- 
ment to smooth the surface. The external surface has been painted so that the 
olfactory centers and hippocampus appear dark and the general pallium light. 
The diencepha'on has been left with the gray color of the modelling paper and 
the ventricu‘ar surface of the hemisphere has had no treatment whatever. The 
lamina terminalis and line of attachment of the choroid plexus have been painted 
white. All the models were made at a magnification of 50 diameters and are 
reduced one-third in reproduction. 


DORSAL VENTRICULAR RIDGE 493 


2 


Fig. 2. Medial view of the model shown in figure 1. Description in the text 


494 


J. 


B. 


JOHNSTON 


DORSAL VENTRICULAR RIDGE 495 


Fig. 3 Chrysemys embryo of 17mm. Medial view of a model of the right 
hemisphere. The narrowing of the interventricular foramen seems to have 
taken place from above downward on account of the expansion of the hippo- 
campus and general pallium. The white area labelled ¢.h., is the cut surface of 
the primordium hippocampi in the median plane. The thin portion of the medial 
hemisphere wall which will form the choroid fissure and plexus is bounded by a 
white line. 

Fig. 4 Same model as figure 3, with the thalamus, hippocampus and primor- 
dium hippocampi removed. The lateral wall of the ventricle shows the dorsal 
ridge bounded below by the deep middle ventricular groove as described in the 
text. The dorsal groove is shallow. There is no evident boundary between the 
dorsal ridge and the pallial thickening at the rostral end. 

Fig. 5 Chrysemys, 28mm. Lateral view of a model of the right hemisphere. 
Note the great elongation of the pyriform lobe in the caudal pole as compared 
with the condition in figure 1. With regard to the exposure of the corpus striatum 
on the lateral surface, compare figures 1 and 5 with the figures in ’15 b. 


THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL. 26, NO. 5 


496 J. B. JOHNSTON 


Figs. 6 to 14 Transverse sections through the right hemisphere of the 
Chelydra embryo illustrated in figures 1 and 2. 

Fig. 6 Section just caudal to the foramen interventriculare. This is the 
most caudal section in which the nucleus caudatus can be recognized and here 
it practically fuses with the nucleus of the lateral olfactory tract. Both are 
undoubtedly parts of the primitive basal olfactory centers which have been car- 
ried out into the caudal pole of the hemisphere evagination. There is in this 
section no apparent boundary between these nuclei which enter into the amygda- 
loid complex and the general pallium. 

Fig. 7 Section through the caudal part of the foramen, 120 microns rostral 
to figure 6. The nucleus caudatus is much more distinct and the proliferation 
of cells for the nucleus lentiformis begins to appear. There is now a slight indi- 
cation of the latera. boundary of the pallium. The letters s.v.m. in this and 
figure 8 indicate the position in which the middle ventricular sulcus is to be 
formed, although it is not actually present in these sections. 

Fig.8 Section through rostral part of the foramen, 100 microns rostral to igure 
7. The lateral olfactory area is represented chiefly by cell masses adjacent to 
the diagonal band. The nucleus caudatus begins to be separated from this area 
by the crus entering the hemisphere. The lentiform proliferation is larger and 
the clear space just above it, which corresponds to the zona limitans lateralis in 
lower vertebrates, is present from this level forward. 

Fig.9 Section through the rostral wall of the foramen. The position of the 
foramen is represented by the light space at the medio-ventral angle of the 
ventricle. The crus has now separated the nucleus caudatus from the nucleus 
of the diagonal band and both are somewhat removed from the larger mass of 
the lateral oifactory area which is here labeled /.pyr. The middle ventricular 
sulcus which is present in this section extends farther caudally in order embryos 
and adult. 


497 


498 J. B. JOHNSTON 


Fig. 10 Section through the anterior commissure. Although the lateral ol- 
factory area is continuous in these sections and even the nucleus caudatus is 
nowhere wholly separated from it, one can clearly see the beginnings of the 
segregation of the pyriform lobe, diagonal band and caudate nucleus by reason 
of the enlargement of the lateral forebrain bundle (crus). This process will be 
completed in later stages by the development of the lentiform nucleus and the 
further enlargement of the crus. 

Fig. 11 Section just rostral to the preoptic recess where the diagonal band 
turns up in the medial wa.l. The dark mass on the medial surface of the area 
parolfactoria represents the beginning of the nucleus of the diagonal band, 
which corresponds to the gyrus subcallosus of the mammalian brain. The 
nucleus caudatus is connected beneath the ventricle with the area parolfactoria. 
From this level rostrad the olfactory area or lobe is quite continuous from its 
lateral border at l.pyr. to its medial border near fi. 

Fig. 12 Section 60 microns rostral to figure 11, where the diagonal band fibers 
rise up in the medial wall to join the fimbria. Although the cells of the pyri- 
form lobe and pallium lie immediately in contact, there appears in many sec- 
tions a very sharp line as if it were caused by a delicate septum running dorsad 
from the small fiber bundle beneath l.pyr. 

Fig. 13 Section through the tuberculum olfactorium. Note that the ventral 
ventricular suleus here descends nearer to the surface than elsewhere so that 
the somewhat bulging tuberculum contains a large pouch of the ventricle. 

Fig. 14 Section through the olfactory peduncle. The ventral part of the 
section cuts the rostral wall of the tuberculum while its dorsal part cuts the 
olfactory formation. 


DORSAL VENTRICULAR RIDGE 499 


500 J. B. JOHNSTON 


Figs. 15, 16, 17 Three transverse sections of the right hemisphere of the 
17 mm. Chrysemys. Figure 17 falls in the caudal part of the foramen and figure 
15 represents next to the last section in which the connection of the hemisphere 
with the thalamus is seen. Figure 16 is between these two. 

In this embryo the mass formed by the pallial proliferation of the earlier stage 
has become divided into a general pallium and the dorsal ventricular ridge. The 
ridge projects into the ventricle forming a deep middle ventricular suleus and a 
shallow dorsal suleus. Note that the dorsal sulcus is quite independent of the 
dorso-medial angle of the ventricle. The dorsal ventricular ridge therefore is 
clearly a thickening of the lateral wall above the lateral zona limitans. This 
zona limitans is the ight streak in the drawings running diagonally between 
d.v.r. above and n.l. and l.pyr. below. The nucleus of the diagonal band is very 
large and definite in figure 17, while in figure 15 and 16 its substance is fused with 
the nucleus of the lateral olfactory tract. The lentiform proliferation is still 
in progress, especially from the dorsal lip of the middle sulcus. The peculiar 
arrangement of ependyma cells seen at this point in the adult already begins to 
be apparent here. 

Figs. 18 to 27. Transverse sections of the right hemisphere of the 28 mm. 
embryo of Chrysemys. 

Fig. 18 Section at about the point where the fimbria curves down behind the 
choroid area. Caudal to this the nucleus of the lateral olfactory tract soon 
disappears and the dorsal ridge continues on into the caudal pole as a simple 
thickening of the lateral walls. At g.p.? is a partly segregated mass of cells 
which probably represents general pallium extending from the caudal pole for- 
ward to this point in the basal wall as it does in the adult. 


DORSAL VENTRICULAR RIDGE 501 


502 J. B. JOHNSTON 


Fig. 19 Section 120 microns rostral to figure 18. In both these sections the 
connection of the dorsal ridge with the pallium is clear and they illustrate the 
more embryonic conditions which prevail in the caudal pole than in the rostral 
part of the hemisphere. 

Fig. 20 Section at the point of connection of the hemisphere with the thala- 
mus where the stria medullaris passes into the thalamus. In this and the next 
two figures the lateral zona limitans appears as a diagonal light line running 
dorsolaterally from s.v.m. to the outer surface. 

Kig. 21 Section through the foramen interventriculare. The lentiform and 
caudate nuclei, neither of which was distinguishable in figure 20, are here quite 
distinct. The crus in this figure separates the caudate from the nucleus of the 
diagonal band. From this point rostrad for thirty to forty sections the crus 
lies practically exposed on the lateral surface. This condition is represented by 
a small striatal area in the model of this brain (fig. 5). 

Fig. 22 Section through the anterior commissure. Here the general pallium 
begins to be decidedly thickened. The motor area of the adult includes this 
region and extends farther caudad. Forward from this level the olfactory area 
is again continuous on the lateral, basal and medial surfaces. 


DORSAL VENTRICULAR RIDGE 503 


504 J. B. JOHNSTON 


Fig. 23 Section through the rostral wall of the preoptic recess and the nucleus 
of the diagonal band (gyrus subeallosus). Between the level of figure 22 and this 
the pyriform lobe has rapidly pushed up over the outer surface of the pallial 
thickening as described in the text. The dorsal ridge is decreasing in size and 
the arrangement of its cells in this and in figure 22 suggests the core-nucleus of 
the adult. The dorsal ridge and the pallial thickening form a common large 
ridge, while in the adult they are separated by a groove (see 715 b, fig. 10). 

Fig. 24 Section through the rostral end of the dorsal ridge and through the 
fissura prima, in which the fibers of the diagonal band are turning up into the 
medial wall. Both pallial thickening and pyriform lobe are massive. 

Figs. 25, 26 and 27. Sections through the anterior part or motor area of the 
general pallium. The whole of what is called general pallium here becomes more 
massive in the adult and is included in the pallial thickening. An important 
feature of these sections is that the general pallium especially at the dorso-medial 
angle is composed of conspicuously large cells. If the results of experiments re- 
ported elsewhere in this journal are to be credited, this is certainly a part of the 
motor area and probably is concerned with the control of the limbs. The pres- 
ence of the large cells is readily understood if this is true. 


DORSAL VENTRICULAR RIDGE 505 


THE MORPHOLOGY AND MORPHOGENESIS OF THE 
CHOROID PLEXUSES WITH ESPECIAL REFER- 
ENCE TO THE DEVELOPMENT OF THE 
LATERAL TELENCEPHALIC PLEXUS 
IN CHRYSEMYS MARGINATA 
PERCIVAL BAILEY 


From the Anatomical Laboratory of the Northwestern University Medical School 
TWENTY-SEVEN FIGURES 


CONTENTS 


MET RUCHOULOULOTI one tap 6 ice a akbar b ninth wee wise ides iwletahe eee eer ae 507 
ANS UOE Ye 26 osa/5, abs stevials wah b & >, Sea etane nb v'S der via a) Ah ce eee > 508 
Material and Methods........... ate ar Sasin ha ta SRO RC ee a 519 
IDEROTIPULON:.« >. 02 ss'< > « Pe idle OE aR hoa caren Ue ahaun b jars St RRR nectar  aeee e 520 
MDIACUESION: aces 2 ec0s Lei eer PO na ee Pe 525 
GONG IBLOUS ee is ol vans v's wp un ey ea Wy) -0) - Fal e 528 


INTRODUCTION 


In a recent communication, the author (Bailey, 715) presented 
an interpretation of the lateral telencephalic choroid plexus in 
the human embryo based on ontogenetic and phylogenetic evi- 
dence. The phylogenetic evidence presented consisted of the 
most accurate descriptions available of the development of this 
structure in the lower vertebrates, viz., Warren’s (’05) account 
of Necturus maculatus and Tandler and Kantor’s (’07) account 
of Platydactylus mauritanicus. With a view to confirming, and 
strengthening if possible, the phylogenetic evidence it was deter- 
mined to investigate more definitely the development of the 
prosencephalic choroid plexuses in others of the lower verte- 
brates. The following scheme taken from Wilder was accepted 


1 Contribution No. 42. Submitted for publication July 20, 1916. 


507 


50S PERCIVAL BAILEY 


to represent as accurately as possible the phylogenetic stages of 
greatest value: 

Amphioxus 

Cyclostomes 

Selachians 

Ganoids 

Urodeles (for Stegocephali) 

Chelonia (for Theromorphs) 

Monotremes (for Pantatheria) 

Marsupials (for primitive Insectivora) 

Insectivora 

Lemurs (modern Mesodonta) 

Cercopithecoidae (tailed monkeys of Old World) 

Tailless Apes (Gorilla, ete.) 

Pithecanthropus (extinct) 

Homo primigenius (extinct) 

Homo sapiens 


Below the Urodeles, the lateral telencephalic choroid plexus 
is very rudimentary if it appears at all. Its condition in the 
Urodeles has been thoroughly elucidated by Warren’s (’05) ac- 
count of Necturus maculatus. For the Chelonia, Chrysemys 
marginata was selected because it was readily accessible and also 
because a great part of the work had already been admirably 
done by Warren (711) in the course of his investigation of the 
‘“‘Paraphysis and Pineal Region in Reptilia.”’ It stands a little 
closer to the line of ascent of the mammals than the lizard, 
Platydactylus mauritanicus, described by Tandler and Kantor. 
Of the forms above the Chelonia, only Didelphys virginiana 
representing the Marsupials, was available. The present com- 
munication is concerned with Chrysemys marginata and a review 
of the literature, leaving Didelphys for a later paper. 


HISTORY 


There are no choroid plexuses in Amphioxus (Burckhardt, ’94). 

Ahlborn (’83) describes well developed plexuses in the roof of 
the hindbrain of Petromyzon planeri, and also in the roof of the 
midbrain. The plexus of the diencephalic roof is not so well 
developed and no reference is made to any telencephalic plexuses. 
Burckhardt (94) writes of Petromyzon fluviatilis: 


MORPHOGENESIS OF THE CHOROID PLEXUSES 509 


Auch hier wie bei Teleostiern bleiben die Plexus inferioris (median 
telencephalic choroid plexus) nur in Gestalt einer Querfalte nach- 
weisbar, die Paraphyse ist eine blosse Kuppel, der caudalen Rand wir 
als rudimentidres Velum auffassen. 


There are no lateral telencephalic choroid plexuses in Cyclo- 


stomes. 

Among the Selachians, however, there is a well formed plexus 
inferioris in Notidanus according to Burckhardt (94), but no 
lateral telencephalic plexuses appear. On the other hand, I infer 
from Kappers and Carpenter (711) that in Chimaera monstrosa 
the lateral plexuses are present. 


Der ependymale Theil der Schluszplatte w6lbt sich in ihrem frontal- 
sten Abschnitt etwas iiber dem Niveau des Gehirnes hinaus eine Art 
paraphyse darstellend um sich dann pl6étzlich wieder einzufalten und 
in den unparen Ventrikel eindringend den plexus chorioideus ventriculi 
imparis zu bilden, wovon auch geringe Ausliufer in den schmalen 
Seitenventrikeln des Gehirns eindringen. 


And Minot (’01) in deseribing Aecanthias remarks: 


The velum has now distinctly the character of a choroid plexus, being 
very irregular in the form of its surface, rich in blood vessels, covered 
by a thin ependyma and projecting far into the cavity of the brain. 
Laterally the projections from its surface are much more developed 
and as the organ has grown forward alongside the median paraphysal 
arch, it has produced what we can now easily identify as the plexus of 
the lateral ventricle. These plexuses are therefore to be interpreted 
morphologically as secondary modifications or appendages of the pri- 
mary velum transversum. . . . . Attention should be paid to the 
two lateral projections, L.ch., of the ependyma on the anterior surface 
of the velum, because these projections not only fix the lateral bound- 
aries of the paraphysal arch but also are the anlages of the choroid 
plexuses of the lateral ventricles. These anlages from this stage (28.0 
mm.) on rapidly increase both in size and in complication of form. 


D’Erchia (96) shows in Torpedo the velum transformed into 
a plexus and the tela chorioidea diencephali practically non- 
existent. In fact, this seems to be the tendency of these two 
structures in the entire Selachian group. There is a plexus in 
the roof of the fourth ventricle in all Selachians. 

No plexus develops in the tela chorioidea telencephali medi 
in Ganoids, but in Acipenser, von Kupffer (’93) figures a plexus 
formation arising from the anterior wall of the velum trans- 


510 PERCIVAL BAILEY 


versum, and Terry (10) describes such a formation in Amia, as 
well as folds which appear on the caudal wall of the velum. 
The tela chorioidea diencephali itself forms a large thin-walled 
dome with no plexus formation. ‘There is a plexus in the roof 
of the fourth ventricle. Lateral telencephalic choroid plexuses 
according to Burckhardt (’94) are absent, but Hill (94) writing 
of Amia says: “It (the paraphysis) may be thought of as an iso- 
lated portion of the roof of the fore-brain which owes its existence 
to the formation of the folds marked Pl.chr. in figure 20, and 
which are themselves the representatives of the choroid plexuses 
of the lateral ventricles.” 

Of the Urodeles, Burckhardt (’91) has described extensive 
plexuses developing both from the tela chorioidea telencephali 
medii and from the tela chorioidea diencephali in Ichthyophis. 
Warren (’05) shows the enormous diencephalic plexus of Necturus 
absorbing also the entire caudal wall of the velum transversum. 
There is a plexus in the roof of the fourth ventricle. The lateral 
telencephalic plexuses are present and arise from the base of the 
median telencephalic plexus (plexus inferioris) as has been at- 
tested by Mrs. Gage (93), Studnicka (’93), Warren (’05) and 
Burckhardt (91). Burekhardt (91) says of Ichthyophis: ** Die 
Plexus der Hirnhemisphiren aber spalten sich je in zwei Stimme 
von denen der eine sich gegen das Zwischenhirn ausstreckt, und 
in der Folge zuerst sich in Zweige spaltet, indess der andere in 
den Hemisphirenventrikel eindringt und sich sodann in zwei 
Zweige spaltet, einen nach riickwarts umbiegenden, welcher den 
Ventrikeltheil des Temporallappens und einen, welcher das 
iibrige Vorderhirn versorgt.’”’ And Warren (’05) writes con- 
cerning Necturus: “The telencephalic plexus develops from the 
paraphysal arch. .’ “The plexuses of the hemi- 
spheres arise on either side from the origin of the telencephalic 
plexus and pass into the lateral ventricles.” 

Concerning the Chelonia, Humphrey (’94) shows a plexus aris- 
ing from the tela chorioidea telencephali medi in Chelydra ser- 
pentina, as does also C. L. Herrick (91) in Cistudo. Warren 
(11) states that the plexus is not present in Chrysemys marginata 
and with this statement my observations agree. The “two 


‘MORPHOGENESIS OF THE CHOROID PLEXUSES 511 


paired masses growing backward from the origin of the lateral 
plexus into the diencephalon” of which Warren writes do not 
seem to me to be at all homologous with the median telencephalic 
plexus. They arise too far posterior on the plexus; are at no time 
connected with the roof plate; and are merely prolongations of 
the lateral plexus. The diencephalic plexus is not so well devel- 
oped as in Urodeles but is present in all forms, as is also the plexus 
of the fourth ventricle. Humphrey (94) shows in Chelydra 
serpentina the lateral telencephalic plexus arising from the base 
of the median telencephalic plexus and passing into the lateral 
ventricle. According to Warren (11), in Chrysemys marginata, 
“The plexus chorioideus lateralis springs from the paraphysal 
arch immediately in front and lateral to the mouth of the para- 
physis, figure 25, and invaginates the dorso-mesial wall of the 
hemisphere.” 

In considering the Mammalia, considerable space will be given 
to interpreting accurately the velum and puraphysis. This is 
rendered necessary if one is profitably to homologize the lateral 
telencephalic plexus of Mammalia with the same structure in 
the lower vertebrates. Since, as we shall see later, the paraphy- 
sis never appears in Mammalia except as an arch of the roof 
plate of the telencephalon in early stages of development, it will 
be referred to as the paraphysal arch. Minot (’01) used the term 
‘to apply to the entire roof plate of the telencephalon between 
the velum transversum and the lamina terminalis, a sense in 
which it is no longer used. 

The only observations of value dealing with the Monotremes 
are those of Th. Ziehen (’05) on Echidna hystrix and G. Elliot 
Smith (97) on Ornithorhynchus, and these leave much to be 
desired. Smith describes the origin of the lateral telencephalic 
plexuses and a structure which he says ‘“‘constitutes the para- 
physis of Selenka.”” I do not know what evidence he had for 
stating that the structure he describes is the paraphysis of Selenka 
for so far as I have been able to determine, Selenka has nowhere 
a description of the paraphysis but merely the bald statement 
that it is present in Marsupials. Smith’s statement probably 
implied no more than that he interpreted this structure as the 


THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL. 26, NO. 5 


512 PERCIVAL BAILEY 


paraphysis. Extracts from this paper follow and will be dis- 
cussed later. 


The lamina supraneuroporica takes a sudden bend backwards (fig. 3) 
to form a horizontal band, which gives origin in many lowly verte- 
brates to the plexus inferioris, and in higher animals to the plexus 
lateralis as well, or exclusively. In the specimen under consideration, 
however, although the plexus laterales do not actually spring from this 
lamina, they are formed from the caudal prolongation of its lateral 
parts. . . . . In describing the structures met with in a medial 
sagittal section it was mentioned that the dorsal part of the (actual) 
anterior wall of the median cavity of the forebrain was bulged out to 
form a large sac. The corresponding structure is well seen in the early 
embryo of Parameles. . . . . In the Platypus embryo, however 
(fig. 2), a well developed choroidal fold extends from the superior 
commissure to the lamina from which the lateral plexus arises, com- 
pletely invaginating the paraphysis (figs. 7, 9 and 15) in the middle 
line. In Platypus the transition from optic thalamus to paraphysis 
is a very gradual one, so that in examining a series of coronal sections 
the lateral walls of the diverticulum would seem to be merely the 
forward continuation of the ependymal layer of the Fligelplatten (fig. 
15). 


As a matter of fact, the appearances in this series of coronal 
sections are not deceptive at all, the sac being undoubtedly. just 
what it appears to be, an anterior pouch of the choroid plexus 
of the diencephalon, the lateral walls of the pouch being actually 
the forward continuation of the Fliigelp’atten. A glance at figures 
1, 2 and 3 will make this quite apparent. The similarity would 
be much more obvious if figure 3 were from a coronal instead of 
a transverse section. In front of this pouch lies the velum 
transversum and then the roof plate of the telencephalon (lamina 
supraneuroporica as he calls it), from the caudal prolongations 
of the lateral parts of which arise the lateral telencephalic plex- 
uses. In a later article, Smith (03) reaffirms his belief that the 
anterior extremity of the lateral telencephalic plexus arises from 
the roof plate. The embryo is doubtless too far advanced to 
show the true paraphysal arch. 

Ziehen’s work on Echidna is eminently unsatisfactory from the 
standpoint of this discussion, because the sections he shows 
invariably skip the anterior end of the lateral telencephalic plexus. 
He has but one suggestive statement referring to this region: 


MORPHOGENESIS OF THE CHOROID PLEXUSES 513 


. Im Bereich des Sulcus hemisphaericus ist die mediale Hemisphiren- 
wand verdiinnt und taschwartig in das Hemisphérenlumen eingestiilpt. 
Diese Tasche entspricht dem Plexus chorioideus ventriculi lateralis. 
Sie 6ffnet sich also in den Sulcus hemisphaericus (und zwar in seine 
laterale Wand) in der Decke des Foramen Monroi und communicirt 
sowohl mit der Sichelspalte wie mit dem hinten obsteigenden zwischen 
Zwischenhirn und Hemisphirenhirn gelegenen Abschnitt des Sulcus 


hemisphaericus. 


Fig. 1 Coronal section from the forebrain of an embryo of Parameles nasuta. 
Copied from Smith. Labels mine. 
Fig. 2 Coronal section from the forebrain of a foetal Ornithorhynchus. 


Copied from Smith. Labels mine. 


Fig. 3 Transverse section through the forebrain of a 19 mm. human embryo. 


H 173, Chicago Embryological Collection. 


Slide 21, Section 11. X 13}. 


REFERENCE LETTERS 


a.a.c.t.l., anterior area chorioidea tel- 
encephali lateralis 

c.o., chiasma opticum 

c.p., commissura posterior 

c.s., corpus striatum 

d-t.gr., di-telencephalic groove 

ep., epiphysis 

f.c., fissura chorioidea 

f.M., foramen interventriculare 

hem., hemisphere wall 

hy., hypothalamus 

l.t., lamina terminalis 

mes., mesencephalon 

n.a.t., nucleus anterior thalami 

n.0., nervus opticus 

p.a.c.t.l., posterior area chorioidea tel- 
encephali lateralis 


p.a., p.t.l., pars anterior plexus telen- 
cephali lateralis 

par., paraphysis 

p.p.,p.t.l., pars posterior plexus telen- 
cephali lateralis 

r.inf., recessus infundibuli 

r.m., recessus mamillaris 

r.post., recessus postopticus 

r.pre., recessus preopticus 

t.c.d., tela chorioidea diencephali 

t.c.t.m., tela chorioidea telencephali 
medii ' 

t.f., taenia fornicis 

th., thalamus 

t.r-p., telencephalic roof-p!ate 

t.t., taenia thalami 

v.t., velum transversum 


514 PERCIVAL BAILEY 


The ‘Sichelspalte’ I understand to be the cleft between the 
hemispheres in front of the diencephalic roof. The lateral telen- 
cephalic plexus, then, must extend anteriorly beyond the velum 
transversum into the roof plate of the telencephalon. 

Ziehen also is inclined to interpret the anterior extremity of 
the diencephalic roof in Echidna as the paraphysis. He remarks: 


Das ganze Bild erinnert an die Paraphyse mancher Reptilien. Ich 
stehe auch nicht an, diese leichte fordere Zuspitzung des hinteren 
Kuppelgebietes der Paraphyse homolog zu sitzen, wie im vergleichenden 
Abschnitt specieller erértert werden soll. Eine tiefere Einsenkung hinter 
der Zuspitzung—etwa im Sinne eines Velum transversum—fehlt. Vor der 
Zuspitzung beginnt sofort die Kinsenkung der Fossa interhemisphaerica. 


Fig. 4 Section from the forebrain of a 63 mm. embryo of Echidna hystrix. 
xX 30. Copied from Ziehen. Labels mine. 

Fig. 5 Transverse section from the forebrain of a 14 mm. pig. no. 18, Chicago 
Embryological Collection. Slide 5, section 19. x 25. 


In an older embryo of which he says, “‘Sehr bemerkenswerth 
is wiederum das Verhalten des Kuppelgebietes des Zwischenhirns 
gegen die Fossa praediencephalica hin,” he describes nothing 
resembling a velum transversum and a comparison of figures 4 
and 5 will show the resemblance of his ‘Kuppelgebietes’ to the 
anterior pouch of the tela chorioidea diencephali of a pig embryo. 
If the plane of section in figure 5 were homologous with that of 
figure 4 the resemblance would be much more striking. If it 
be really the paraphysis, it forms a good transition stage to the 
condition found in higher Mammalia. 


MORPHOGENESIS OF THE CHOROID PLEXUSES 5a 


Observations on the Marsupial brain are remarkably meagre. 
I have been able to find but two statements bearing on the sub- 
ject. Broom (’97) says of a 14.8 mm. embryo of Trichosurus 
vulpecula: ‘‘ The choroid folds into the lateral ventricles, is partly 
formed, and the paraphysis well marked.’ He has no figures of 
sections. Selenka (’91) has a simple statement that the paraphy- 
sis is present in Marsupials, presumably in the opossum. “Bis 
jetzt habe ich die Paraphyse bei embryonen von Haifischen, 
Reptilien, und Beuteltieren beobachtet, zweifle jedoch nicht, 
dass sie allen Wirbeltieren gemeinsam ist.’’ That is all, but I 
hope soon to fill this gap. 


Fig. 6 Sagittal section of the forebrain of a 15 mm. pig. Copied from John- 
ston. XX 10. Labels mine. 

Fig. 7 Sagittal section of the forebrain of a 11 mm. embryo of Erinaceus 
europaeus. Copied from Grénberg. X 10. Labels mine. 


Thanks mainly to Grénberg, the brain of the Insectivora is 
better known. His work was done with Erinaceus europaeus. 
Ziehen (’06) says of Tupaja, ‘‘ Die Verhiiltnisse gleichen den fiir 
den Igel beschriebenen in hohem Masse.” 

Groénberg’s figures 25, 26, and 27 show clearly the velum trans- 
versum and paraphysal arch although he is loath to call them 
so. A comparison of figures 6 and 7 will make this plain. He 
describes also plexuses in the diencephalon and fourth ventricle. 
Of the lateral telencephalic plexus he writes: 


516 PERCIVAL BAILEY 


Die Adergeflechtsfurche entsteht am frithesten. Sie ist schon bei 
meinem Stadium C (11 mm.) vorhanden und sowohl ausgebildet, dass 
ihre erste Entstehung sicher auf einem bedeutend jiingern Stadium zu 
suchen ist. Doch findet sich auf Stadium B (8 mm.) noch keine Spur 
einer Faltenbildung. Die Form und das Aussehen der Falte ergeben 
sich aus den Figg. 52 und 53. Man sieht, dass sie in ihrem vordern 
Theil weiter in die Hemisphirenhohle hineinreicht, als es mehr caudal- 
wirts der Fall ist. Es zeigt sich auch, wenn man eine Schnittserie 
durchmustert, dass die Falte nach hinten allmahlch kleiner wird und 
schliesslich nur eine leichte Einbuchtung darstellt, welche nach hinten 
ganz allmiihlich verstreicht. Der vordere Theil dagegen verhallt sich 
ganz anders. Die Falte ist hier tief und erstreckt sich mit ihrem freien 
Rand weiter nach vorn als die mit der itibrigen Hirnwand in Verbindung 
stehende basis. . . . . Auf den folgenden Stadien vergréssert sich 
die Adergeflechtsfalte bedeutend, besonders in ihr vorderer, freier 
Theil, : 


Fig. 8 Cross section through the forebrain of a 15 mm. embryo of Erinaceus 
europaeus. Copied from Grénberg. X 10. Labels mine. 

Fig. 9 Cross section through the forebrain of a 32mm. humanembryo. H41, 
Chicago Embryological Collection. X 25. Slide 29, Section 6. 


One would conclude from this description that the anterior 
end of the lateral plexus had arisen the earlier, and that it arises 
from the roof plate of the telencephalon is clearly shown in his 
figure 54, which is here reproduced with a section from a human 
embryo for comparison (figs. 8 and 9). Again if the section of 
the human embryo were a coronal instead of a transverse section, 
the similarity would be more obvious. 


MORPHOGENESIS OF THE CHOROID PLEXUSES 517 


Of the forms between Insectivora and Man nothing of value is 
known. Ziehen (’06) shows sections of the lateral telencephalic 
plexus of Tarsius spectrum, a prosimian, but not in the proper 
plane to be of value, and again describes the anterior pouch of 
the diencephalic roof as the paraphysis. 

Sehr beachtenswert ist auch, dass unmittelbar hinter der Fossa prae- 
diencephalica die epitheliale Decke des Zwischenhirns sich zu einer 
stulen Falte, welche an die Paraphyse erinnert, erhebt und dass erst 


einige Schnitte weiter occipitalwiirts diese steile Falte durch den mediane 
Plexus des 3. Ventrikels eingestiilpt wird. 


In an earlier paper, the author (Bailey, ’15) called attention 
to the true homologue of the paraphysis in the human embryo 
and insisted upon its position in the telencephalic roof plate just 
anterior to the velum transversum. Previously, Streeter (12), 
Francotte (’94) and D’Erchia (’96) had written of the paraphysis 
in the human embryo. Streeter homologizes the anterior pouch 
of the diencephalic roof with the paraphysis, and from a com- 
parison of Francotte’s figures with sections of embryos of approxi- 
mately the same stage, I am convinced that he mistook the same 
structure for the paraphysis. D’Erchia shows a section from a 
30 mm. human embryo in which he labels a structure ‘paraphy- 
sis’ which seems to me to be merely an oblique section of the 
lateral telencephalic plexus. 

In his model of a 13.6 mm. human embryo, His shows clearly 
the lateral telencephalic plexus coming off lateral to the paraphy- 
sal arch, but his statement (’04) of the origin takes no account 
of this fact. ‘‘Sein dem Thalamus angehefteter Randstreifen 
bleibt ependymal und in ihm bildet sich die Fissura chorioidea, 
von der aus die Epithelfaltungen des Corpus chorioideum in den 
Seitenventrikel sich einstiilpen.”’ 

Hochstetter (13) in his account of the development of the 
lateral telencephalic plexus in the human embryo shows one 
figure (fig. 6) of a section through the plexus which passes also 
through the telencephalic roof plate, but his description contains 
nothing concerning the origin of this part of the plexus. 

D’Erchia (’96) considers the lateral -telencephalic plexus to be 
derived from the velum transversum: “‘Per questa parte volga 


518 PERCIVAL BAILEY 


tutto quello che si e detto per la cavia. Dal velum per diverti- 
coli laterali di formano i plessi emisferici inferiori destro e sinistro, 
dei quali uno e disegnato nella fig. 28.” 

Finally the author (Bailey, 715) has presented an account of 
the development of the lateral telencephalic plexus which it 
is the purpose of the present paper to supplement. 

Concerning these structures in other groups of vertebrates 
this discussion is not so much concerned, but a few words at least 
about the lateral plexus in other vertebrates will not be out 
of place. ‘This plexus is not present in Teleosits (Burckhardt, ’94) 
nor Anura (Herrick, 710). In Lacertilia 1ts development has 
been described by Tandler and Kantor (07) and Warren (11) 
and agrees in all essential respects with that hereinafter detailed 
for Chrysemys marginata. In examining some young alligators 
in the collection of Dr. Elizabeth Crosby, the posterior part of 
the lateral telencephalic plexus was found to be present, but 
very small and poorly developed. ‘The brain of an adult pigeon 
was also examined and the lateral telencephalic plexus was 
seen arising from a plexus formation in the roof of the telen- 
eephalon, but of the posterior part of the plexus there was no sign 
except a plexus of blood vessels along a thin but uninvaginated 
medial hemisphere wall. Dr. C. Judson Herrick suggested that 
this may represent a stage in the phylogenetic development of 
the plexus. Mrs. Gage (’95) figures the brain of the embryo 
of the English sparrow. Both the figures and the context show 
that the’ lateral telencephalic plexuses arise from the median 
telencephalic plexus and just lateral to the paraphysis, appar- 
ently not invaginating the medial hemisphere wall. She writes: 
“In the young sparrow ( g. 2) the paraphysis occurs in the midst 
of a mass which gives off the paraplexuses, and it opens directly 
dorsad of the portas, i.e., into the aula. . . . It is notice- 
able in the older embryo that the union of the auliplexus with 
the paraplexuses lies dorsad of the porta (fig. 14).’’ The works 
of Neumayer (’99) and Hochstetter (98) contain nothing of 
value for the present discussion. The two invaginations which 
they describe are temporary and unimportant. A similar con- 
dition is shown in my figure 19 (Bailey, ’15). D’Erchia (96) 


MORPHOGENESIS OF THE CHOROID PLEXUSES 519 


considers the lateral plexus in the guinea pig to be derived 
from the velum transversum. “E dal velum che si origina il 
plesso del III ventricolo e i plessi coroidei emisferici destro e 
sinistro.’’ It is interesting also to note that although Tilney (’15) 
correctly labels the paraphysal arch in models of the brains of 
young cat embryos, in later stages he invariably transfers the 
label to the anterior pouch of the diencephalic roof. Johnston 
(09) has correctly interpreted the paraphysis in pig embryos, 
but his statement that the lateral telencephalic plexus appears 
as a folding of the anterior wall of the velum transversum is 
misleading for he immediately follows with the statement that 
it is separated from the velum by the paraphysal arch. 


In the angle between the vesicle and the diencephalon appears the 
chorioid plexus pushing into the lateral ventricle. It appears as a fold- 
ing of the anterior wall or limb of the velum transversum and its lateral 
prolongation. In this way appears the chorioid fissure whose further 
history need not be traced. Near the median line the plexus appears 
as a fold projecting into the interventricular foramen and separated 
from the velum transversum by the paraphysal arch. 


MATERIAL AND METHODS 


The material on which this study is based consists of a series 
of twelve embryos of Chrysemys marginata from my own col- 
lection, ranging in size from 5.1 mm. greatest length to one having 
a carapace 10.6 mm. long, and an 8.8 mm. embryo from the 
Harvard Embryological Collection, very kindly loaned to me 
by the Harvard Laboratory. All of my material was fixed in 
Zenker; stained in bulk with borax carmine; embedded by the 
celloidin-paraffin method; cut at 10. in transverse series; and 
counterstained on the slide with orange G. The excellent pres- 
ervation of the form relations of the delicate membranous por- 
tions of the brain may be attributed to the method of embedding. 
The embryos were all passed from 95 per cent alcohol to ether- 
alcohol; then through 0.5, 1 and 2 per cent celloidin; hardened 
in chloroform alcohol, and cleared in benzol, before embedding 
in paraffin. The Harvard embryo is cut in 10 » sagittal sections 
and stained with borax carmine and eosin. : 


520 PERCIVAL BAILEY 


The forebrains of four embryos—l1 a, greatest length 5.1 mm.; 
H. E. C. no. 1433, greatest length 8.8 mm.; 5b, carapace 8.6 
mm.; and 4b, carapace 10.6 mm.—were reconstructed by the 
Born method at a magnification of 100 diameters. Millimeter 
plates were used and every section drawn. It was necessary 
to dissect the models rather extensively to expose: the fissura 
chorioidea. The models were stacked from a side view of the 
head drawn from a photograph, with the exception of the Har- 
vard embryo. This latter embryo was cut sagittally and the 
stacking was guided by the epiphysis and paraphysis. Since 
the embryo was not cut in an exactly sagittal plane, after the 
paraphysis and epiphysis had passed out of the plane of section, 
most of the lateral telencephalic plexus had been stacked and 
the remaining sections were added with no other guide except 
comparison with other models. 


DESCRIPTION 


This description will be confined to the lateral telencephalic 
plexus, since concerning the other plexuses I find no reason to 
differ from Warren’s account, with the exception noted in the 
history. 

The main landmarks of the region in which the lateral telen- 
cephalic plexus develops are already laid down in an embryo of 
5.1 mm. greatest length. Figure 26 shows the roof plate of the 
forebrain in such an embryo and figure 10 shows the same region 
schematically represented. The roof plate of the telencephalon, 
back of the lamina terminalis, appears as a triangular area (only 
half of it shown in figure 26) its base formed by the velum trans- 
versum and its apex lying at the posterior end of the lamina 
terminalis while from its center arises the paraphysis. The 
lateral sides of the triangle are formed by the taeniae fornices. 
At the lateral angles of the triangle, taenia fornicis, velum trans- 
versum, taenia thalami and di-telencephalic groove meet. It 
is not possible accurately to determine this point in the embryo 
under consideration because the roof-plate has been but imper- 
fectly differentiated histologically, but in later stages these 
angles of the triangle may be easily located. | 


MORPHOGENESIS OF THE CHOROID PLEXUSES 521 


The earliest unquestionable appearance of the lateral telen- 
cephalic plexus is in the Harvard embryo, 8.8 mm. in length. 
It appears here as a crescent-shaped ridge projecting into the 
lateral ventricle, with two more strongly developed points 
. showing as small elevations (fig. 22). The anterior extremity 
of the plexus lies clearly in the roof of the telencephalon lateral 
to the paraphysis and medial to the taenia fornicia. The taenia 
fornicis appears in figure 22 as a ridge apparently in the medial 
hemisphere wall. It bears here a very close superficial resem- 
blance to the hippocampal ridge in a young human embryo, but 
its later development and internal structure show plainly that 


Fig. 10 Diagram of the region around the paraphysis in the roof of the fore- 
brain of a 5.1 mm. embryo of Chrysemys marginata. 

Fig. 11 Diagram of the same region in an embryo of 8.8 mm. 

Fig. 12 Diagram of the same region in an embryo having a carapace of 8.6 mm. 


it is the taenia fornicis. Its apparent position in the medial 
hemisphere wall is an illusion produced by the invagination 
of the plexus between it and the paraphysis. The triangular 
area of the telencephalic roof plate is well marked and the lateral 
angles can be determined. A sagittal section through the tri- 
angle is practically a straight line, except for the evagination 
of the paraphysis (figs. 13 and 21) as are also parasagittal sec- 
tions (fig. 14). Nevertheless the lateral angles are depressed 
and the sides of the triangle are concave outward. Transverse 
sections through this region are therefore curved and convex 
upward (fig. 21). The velum transversum sharply delimits the 
telencephalic roof plate at the base of the triangle (fig. 14). 


522 PERCIVAL BAILEY 


Figure 15 shows a parasagittal section still farther laterally 
through the triangle. The velum transversum is obvious, the 
plexus invaginating the roof plate medial to the taenia fornicis 
and tilting the portion of the roof between itself and the taenia. 
Above and lateral to the taenia fornicis the brain wall is his- 
tologically differentiated; medial it is ependymal. This differ- 
ence is much more apparent in later stages (fig. 17). Poste- 
riorly the plexus is less well developed and as we approach the 
lateral angles of the triangle, plexus, velum, and taenia fornicis 
tend to merge into one groove which becomes continuous with 


Fig. 13 Sagittal section no. 135 from an 8.8 mm. embryo of Chrysemys mar- 
ginata. Harvard Embryological Collection, no. 1433. X 33}. 

Hig. 14 Parasagittal section no. 130 from the same embryo. 

Fig. 15 Parasagittal section no. 124 from the same embryo. 


the di-telencephalic groove. Lateral and anterior to the point 
where the taenia thalami meets the di-telencephalic groove and 
velum transversum the hemisphere wall is massive and unin- 
vaginated. It would seem, then, that in this embryo, the 
lateral telencephalic plexus lies entirely in the roof plate of the 
telencephalon. The condition is diagrammatically represented 
in figure 11. 

Figures 23 and 24 depict the region around the foramen of 
Monro in an embryo with a carapace 8.6 mm. long, about 14 mm. 
greatest length. Figure 23 looks at the foramen of the right 
half of the brain from the medial side. The cephalic end of the 


MORPHOGENESIS OF THE CHOROID PLEXUSES 523 


model is, consequently to the left. The paraphysis which 
had been removed is represented by a line of white dashes. 
Figure 24 looks at the same foramen from the lateral side. The 
model is therefore reversed and the cephalic end points to the 
right. 

The condition of the roof-plate is diagrammatically represented 
in figure 12. The differentiation between hemisphere wall and 


Fig. 16 Transverse section from the forebrain of an embryo of Chrysemys 
marginata having a carapace 8.6 mm. in length. Embryo 5 b, slide 13, sect. 11. 
X 33}. 

Fig. 17 Transverse section of the same embryo; section 16, slide 13. X 33}. 


roof plate is clear and the taenia fornicis obvious (figs. 16 and 
17). The anterior extremity of the lateral telencephalic plexus 
arises plainly from the telencephalic roof plate lateral to the 
paraphysis and medial to the taenia fornicis (fig. 16). Figure 
17 shows a section posterior to figure 16 through the main body 
of the plexus, the plexus still lying in the roof plate. Figure 
18 is of a section still farther posteriorly. The plexus is here 
shown crossing the taenia fornicis into the medial hemisphere 
wall. The taenia fornicis is now medial to the plexus and drop- 


524 PERCIVAL BAILEY 


ping down to meet the anterior nucleus of the thalamus, which 
it does in a few sections and becomes continuous with the taenia 
thalami. In figure 19, much farther posteriorly, the taenia 
thalami is present, the plexus being entirely in the medial hemi- 
sphere wall. The portion of the plexus arising from the medial 
hemisphere wall is very poorly developed (fig. 19) but its area of 
invagination is extensive (fig. 24). 

In later stages this posterior part of the plexus develops more 
rapidly and overshadows the other. Figure 25 shows a lateral 


%, 


& 


a, 


Fig. 18 Transverse section of the same embryo as figure 16. Slide 13, sec- 
HONEA DOOR. 
Fig. 19 Transverse section of the same embryo; section 5, slide 14. X 333. 


view of the region around the foramen of Monro in an embryo 
with a carapace of 10.6 mm. A pen sketch of the entire model 
is appended (fig. 27) showing the region represen ed in figure 
25. (Figure 24 is of a homologous region in a younger embryo.), 
The fissura chorioidea appears merely as a big hole in the medial 
hemisphere wall; all the landmarks are lost. In still later stages, 
this hole becomes reduced to a long narrow slit. 

The development of the plexus in size and shape is so well 
discussed and figured by Warren (711) that it will not be con- 
sidered here. 


MORPHOGENESIS OF THE CHOROID PLEXUSES 525 


DISCUSSION 


From the foregoing history and description, and from an 
analysis of the remaining literature, for in the history are in- 
cluded only the most important papers and especially those 
dealing with the lateral telencephalic plexuses, it will be seen 
that there are certain definite regions of the brain wall wherein 
choroid plexuses develop. These regions and the plexuses which 
develop from them may be tabulated as follows: 

Tela chorioidea telencephali medii—Plexus telencephali medius 
Anterior area chorioidea lateralis telencephali—Plexus telencephali lateralis 
(below Chelonia) 
Anterior area chorioidea lateralis telencephali—Plexus telencephali 
lateralis (Chelonia and above) 
Posterior area chorioidea lateralis telencephali (hemispherici)—Plexus 
telencephali lateralis (Chelonia and above) 
Velum transversum—Plexus velares 
Tela chorioidea diencephali—Plexus diencephali 
Tela chorioidea mesencephali—Plexus mesencephali 
Tela chorioidea myelencephali—Plexus myelencephali 


Tilney (’15) has suggested that the saccus vasculosus should 
be reckoned with the choroid plexuses in the forms where it is 
present and in this opinion I concur. It is best developed in 
those forms in which the diencephalic plexus is rudimentary or 
absent, i.e., in Cyclostomes, Selachians and Ganoids. The plexus 
formation is very poorly developed in Urodeles and never again 
present. There should be added to the above therefore: 


Recessus infundibularis (posterior wall)—Saccus vasculosus. 


The myelencephalic plexus arises in the roof of the fourth 
ventricle, tela chorioidea myelencephali, in every known verte- 
brate above Amphioxus. 

The mesencephalic plexus is found only in Petromyzon, where 
it arises from the mesencephalic roof, tela chorioidea mesen- 
cephali. 

The diencephalic plexus arises from the tela chorioidea dien- 
cephali. Appearing first in Cyclostomes, but very poorly 
developed, it disappears almost entirely in Selachians, where 
the tela chorioidea diencephali is almost completely absorbed 


526 PERCIVAL BAILEY 


by, the overgrowth of the velum transversum. In Ganoids, the 
tela chorioidea diencephali begins to emerge, forming a thin 
walled sac, and in Urodeles is again invaginated by an enormous 
diencephalic plexus. It may be that the diencephalic plexus 
in Selachians and Ganoids is represented by the choroidal folds 
on the posterior limb of the velum, developing from the anterior 
portion of the tela chorioidea diencephali which has been drawn 
down into the velum transversum by the overgrowth of the 
latter. The diencephalic plexus is present in all forms above 
Urodeles but is never again so well developed. 

The velar plexuses are present only in Selachians, Ganoids, 
and Urodeles. They may involve either the diencephalic limb 
or telencephalic limb, or the entire velum transversum. The 
choroidal folds on the diencephalic limb have been homologized, 
in Selachians and Ganoids, with the diencephalic plexus of other 
forms, and with some reason. The choroidal folds on the ante- 
rior limb are not homologous with either the median or lateral 
telencephalic plexuses. 

The median telencephalic plexus arises from the tela chorioidia 
telencephali medu, just in front of the paraphysis (paraphysal 
arch of Mammalia), from the Selachians to the Chelonia, in- 
clusive, with the apparent exception of the Ganoids. It is 
not constantly present in Chelonia and is never found in Mam- 
malia. Its development seems to be in inverse ratio to the 
degree of development of the lateral telencephalic plexus and 
the latter in direct ratio to the size of the hemispheres. 

The lateral telencephalic plexus is found in all groups of verte- 
brates from the lowest to the highest in the line of ascent of 
Mammals with the exception of the very lowest, Amphioxus 
and the Cyclostomes. Apparently sporadically and imperfectly 
developed in Selachians and Ganoids, it is present constantly 
thereafter. Wehave stated previously that it is absent in Teleosts 
and Anura, with which we are not concerned. In all forms below 
Chelonia, it develops in what I have called elsewhere (Bailey, 715) 
the anterior lateral telencephalic chorioidal area, in the roof . 
plate of the telencephalon between the paraphysis and the taenia 
fornicis of the medial hemisphere wall. Where the median 


MORPHOGENESIS OF THE CHOROID PLEXUSES Hoe 


telencephalic plexus is strongly developed the lateral plexus 
appears to be an appendage of it (Necturus, Warren); where 
the velum is involved in an extensive plexus formation, the 
lateral plexus appears to be an appendage of it (Acanthias, 
Minot); but always arises from the region of the telencephalic 
roof plate described above, and medial to the taenia fornicis. 

With the Chelonia comes a change. The lateral plexus arises 
in the anterior area chorioidea lateralis telencephali as has been 
previously described but in its later development crosses the 
taenia fornicis and invaginates also the posterior area chorioidea 
lateralis telencephali in the medial hemisphere wall. Such a 
condition is found also in the Gecko brain (Tandler and Kantor). 
This involvement of the medial hemisphere wall comes more and 
more to predominate in the development of the lateral plexus as 
the hemispheres come more and more to dominate the develop- 
ment of the telencephalon, but even in the highest Mammalia 
the anterior extremity of the lateral choroid plexus develops 
from the roof plate of the telencephalon and the evidence may 
briefly be summarized here: 

(a) Elliot Smith (97) in deseribing the brain of a foetal Orni- 
thorhyneus described the lateral choroidal plexus as arising from 
the continuation backwards of the horizontal part of the lamina 
supraneuroporica, which term he used to inelude all the roof 
plate of the telencephalon between the velum transversum and 
the lamina terminalis. Again in 1903, he reiterated his belief 
that the anterior extremity of the lateral choroidal plexus arises 
from the roof plate. 

(b) Th. Ziehen’s figures of Echidna do not show the anterior 
extremity of the lateral plexus, but he states that it opens into 
the Sichelspalte, that is, into the cleft between the hemispheres 
anterior to the diencephalon, as well as the suleus hemisphaericus 
(di-telencephalic groove). 

(c) Observations on the Marsupials up to present reveal 
nothing. 

(d) Grénberg states of Erinaceus europaeus that the anterior 
extremity of the lateral plexus is better developed in his earliest 
stage than the posterior extremity and apparently arises first. 


THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL. 26, NO. 5 


528 PERCIVAL BAILEY 


This holds true of all mammals and is what one would expect 
from the phylogenetic history of the plexus. His figure 42 of 
a section anterior to the velum shows clearly the anterior extrem- 
ity arising from the roof plate of the telencephalon. 

(e) The earliest appearance of the plexus in the human embryo 
occurs in His’ embryo of 13.6 mm. and his model shows clearly 
the plexus arising just lateral to the paraphysal arch. 

(f) Hochstetter (13) figures a section of a human embryo 
showing the anterior extremity of the plexus, and the author 
recently (Bailey, 715) has described three human embryos show- 
ing the plexus extending into the roof plate alongside the para- 
physal arch anterior to the velum transversum. 

(g) Collateral evidence is derived from the development of the 
lateral plexus in other mammals, for example, Johnston (’09) 
states that in the pig the plexus arises from the anterior limb of 
the di-telencephalic groove but is separated medially from the 
velum transversum by the paraphysal arch. 


CONCLUSIONS 


1. The lateral telencephalic plexus of Chrysemys marginata 
arises from the anterior area chorioidea lateralis telencephali in 
the roof plate of the telencephalon between the paraphysis and 
the taenia fornicis of the medial hemisphere wall and in its later 
development oversteps the taenia fornicis just anterior to the 
point where taenia fornicis, taenia thalami, velum transversum 
and di-telencephalic groove meet, and invaginates also the pos- 
terior area chorioidea lateralis telencephali in the medial hemi- 
sphere wall, just anterior to the di-telencephalic groove. The 
pars anterior plexus telencephali lateralis which develops from 
the anterior area chorioidea lateralis telencephali in early stages 
is much larger and better developed than the pars posterior 
plexus telencephali lateralis, but in later stages develops much 
less rapidly than the pars posterior. 

2. All the data at present available support the conclusion that 
the lateral telencephalic plexus arises from the anterior area 
chorioidea lateralis telencephali in the roof plate of the telen- 
cephalon between the paraphysis (or paraphysal arch) and the 


MORPHOGENESIS OF THE CHOROID PLEXUSES 529 


taenia fornicis in all vertebrates where it is present, but that in 
Chelonia and the forms above, it oversteps the taenia fornicis 
and invaginates the posterior area chorioidea lateralis telencephali 
in the medial hemisphere wall and that this latter portion comes 
more and more to dominate its development as we ascend the 
vertebrate scale. 


I wish here to express my gratitude to Dr. 8. Walter Ranson 
for his unvarying kindness and interest throughout the progress 
of this work. 


530 PERCIVAL BAILEY 


LITERATURE CITED 


Aunxtporn, F. 1883 Untersuchungen tiber das Gehirn der Petromyzonten, 
Zeits. wiss. Zool., Bd. 39. 

BaiLtey, P. 1915 The morphology of the roof plate of the forebrain and the 
lateral choroid plexuses in the human embryo. Jour. Comp. Neur., 
vol. 26. 

Buake, J. A. 1900 The roof and lateral recesses of the fourth ventricle, con- 
sidered morphologically and embryologically. Jour. Comp. Neur., 
vol. 10. 

Broom, J. 1897 The embryology of Trichosurus vulpecula. Proc. Roy. Soe. 
of N. S. Wales, vol. 27. 

Burckuarpt, R. 1891 Untersuchungen am Gehirn und Geruchsorgan yon 
Triton und Ichthyophis. Zeitschr. f. wiss. Zool., Bd. 52. 
1894. Der Bauplan des Wirbeltiergehirns. Morph. Arbeiten von G. 
Schwalbe, 4. 

D’Ercuia, Fu. 1896 Contributo allo studio della volta del cervello intermedio 
e della regione parafisaria In embrioni di pesci e di Mammiferi. Moni- 
tore Zoologico, vol. 7, pp. 118 and 201. 

FRANcoTTE, P. 1894 Note sur l’oeil parietal, ’epiphyse, la paraphyse et les 
plexus choroides du troisieme ventricule. Bull. de lAcad. roy. de 
Belg., no. 1. 

Gaae, Mrs. 8. P. 1893 The brain of Diemyctylus viridescens. Wilder Quarter- 
Cent. Book. - 1898. 
1895 Comparative morphology of the brain of the soft-shelled turtle 
(Amyda mutica) and the English sparrow (Passer domesticus). Proc. 
Am. Mier. Soe., vol. 17. 

GRONBERG, GOstTa. 1901 Die Ontogenese eines niedern Sdiugergehirns nach 
Untersuchungen an Erinaceus europaeus. Zoolog. Jahrb., Abth. Anat., 
Bas 15s» 26s 

Herrick, C. J. 1910 The morphology of the forebrain in Amphibia and Reptilia. 
Jour. Comp. Neur., vol. 20. 

Herrick, C. L. 1891 and 1893 Topography and histology of the brain of certain 
reptiles. Jour. Comp. Neur., vols. 1 and 3. 

Hitt, C. 1894 The epiphysis in teleosts and Amia. Jour. Morph., vol. 9, p. 237. 

His, Wituetm. 1904 Die Entwicklung des menschlichen Gehirns. Leipzig. 

Hocusterrer, F. 1898 Beitriige zur Entwickelungsgeschichte des Gehirns. 
Bibliotheca Medica., Abthg. A., Heft. 2. 
1913 Uber die Entwickelung der Plexus Chorioidei der Seitenkammern 
des menschlichen Gehirns. Anat. Anz., Bd. 45, pp. 225-238. 

Humpnrey, O. D. 1894 On the brain of the snapping turtle. Jour. Comp. 
Neur., vol. 4. 

Jounston, J. B. 1909 On the morphology of the forebrain vesicle in vertebrates. 
Jour. Comp. Neur., vol. 19. 

KAPPERS AND CARPENTER 1911 Das Gehirn von Chimaera monstrosa. Folia 
Neuro-Biologica, Bd. 5. 

Kineassury, B. F. 1897 The encephalic evaginations in Ganoids. Jour. Comp. 
Neur., vol. 7. 


MORPHOGENESIS OF THE CHOROID PLEXUSES 531 


Kuprrer, C. von 1893 Die Entwicklung des Kopfes von Acipenser sturio. 
Studien zur vergleichenden Entwicklungsgeschichte des Kopfes der 
Kranioten. Heft. I., Miinchen. 

1903 Die Morphogenie des Nervensystems. Oskar Hertwig’s Hand- 
buch. 

Minot, C. S. 1901 On the morphology of the pineal region based upon its 
development in Acanthias. Am. Jour. Anat., vol. 1. 

Neumayer, L. 1899 Studien zur Entwicklungsgeschichte des Gehirns der 
Siugetiere. Festschr. z. 70 Geburtstag von C. von Kupffer. Jena. 
Fischer. 

Sevenka, E. 1891 Das Stirnorgan der Wirbeltiere. Biol. Centralbl., Bd. 10. 

Smitu, G. Exvuiort 1897 The brain of a foetal ornithorhynchus. Quart. Jour. 
Micros. Sci., n. s., vol. 39. 

1903. On the morphology of the cerebral commissures. Trans. of 
the Linnaean Soc., London., vol. 8, part 12. 

Srerzi1,G. 1907 Il sistema nervoso centrale dei vertebrate. Vol. 1, Ciclostomi. 

Srreeter, G. L. 1912 The development of the central nervous system. Keibel 
and Mall’s Human Embryology. 

Srupnicka, F. K. 1893 Zur Losung einiger Fragen aus der Morphologie des 
Vorderhirns der Cranioten. Anat. Anz., Bd. 7. 

TANDLER, J. AND Kantor, H. 1907 Beitrige zur Entwickelung des Vertebrat- 
engehirns. I. Die Entwickelungsgeschichte des Geckohirnes. Anat. 
Hefte, Bd. 33, Heft 101. 

Terry, R. J. 1910 The morphology of the pineal region in teleosts. Jour. 
Morph., vol. 21. 

Tinney, Frepertck 1915 The morphology of the diencephalic floor. Jour. 
Comp. Neur., vol. 25, no. 3, pp. 213-282. 

WarrREN, Jonn 1905 The development of the paraphysis and pineal region in 

Necturus maculatus. Am. Jour. Anat., vol. 5, pp. 1-29. 

1911 The development of the paraphysis and pineal region in reptilia. 
Am. Jour. Anat., vol. 11, pp. 313-392. 

ZinHEN, Tu. 1904 Verhand. Kon. Ak. van Wetensch. te Amsterdam. 26 Noy. 
(pp. 331-340). 

1905 Zur Entwickelungsgeschichte des Centralnervensystems von 
Echidna hystrix. Jenaische Denkschriften, Bd. VI, 2. Theil. 

1906 Morphogenie des Centralnervensystems der Siugetiere. Oskar 
Hertwig’s Handbuch. 


PLATE 1 
EXPLANATION OF FIGURES 


The reference letters for all figures are found on page 513. It is impossible 
to portray all the topography of the region around the foramen interventriculare 
by means of drawings. For aid in orientation, the position of the principal 
landmarks is indicated on the models as follows: the taenia fornicis by a line of 
dashes; the velum transversum by a line of crosses; the taenia thalami by dots; 
and the di-telencephalic groove by dots and dashes alternating. The cross 
sections will also aid. On figures 22 and 23 the positions of the sections are 
indicated by arrows. Since the models are tilted, the arrows indicate accurately 
only the points where the sections cross the lateral telencephalic plexus. 

20 Median view of the forebrain of a 5.1 mm. embryo of Chrysemys marginata. 
Embryola. X 50. 


532 


MORPHOGENESIS OF THE CHOROID PLEXUSES PLATE ft 
PERCiVAL BAILEY 


533 


PLATE ‘2 
EXPLANATION OF FIGURES 


21 Median view of the forebrain of an 8.8 mm. embryo of Chrysemys margi- 
nata. Embryo 1433. H.E.C. X 40. 

22 Lateral view of the forebrain of an 8.8 mm. embryo of Chrysemys mar- 
ginata. Embryo 1438. H.E.C. X 40. Lateral hemisphere wall removed. 


5384 


MORPHOGENESIS OF THE CHOROID PLEXUSES PLATE 2 


PERCIVAL BAILEY 


te.t.m. 


a hem 


‘inf 992 


539 


MORPHOGENBSIS OF THE CHOROID PLEXUSES PLATE 3 
PERCIVAL BAILEY 


Lelim = 3 soe he ge 2 Bia alo tic 


oo\td 


at 


ISS1O, 17. 18 19 
ae i 23 


EXPLANATION OF FIGURES 


23 Median view of the region around the foramen interventriculare in an 
embryo of Chrysemys marginata having a carapace 8.6 mm. in length. Embryo 
5b. > 100. In order to show the fissura chorioidea it was necessary to remove 
the paraphysis and represent the fissure slightly higher than it really is. 


536 


MORPHOGENESIS OF THE CHOROID PLEXUSES PLATE 4 
PERCIVAL BAILEY 


EXPLANATION OF FIGURES 


Lateral wall of the hemi- 


24 Lateral view of the same region as in figure 23. 
horioidea. 


sphere and plexus telencephali lateralis removed, exposing the fissura ¢ 
Embryo 5b. X 100. 


537 


PLATE 5 
EXPLANATION OF FIGURES 


25 Lateral view of the region around the foramen interventriculare in an 
embryo of Chrysemys marginata having a carapace 10.6 mm. in length. Embryo 
4b. & 662. Lateral hemisphere wall and lateral telencephalic plexus removed, 
exposing the fissura chorioidea. 

26 Dorsal view of the forebrain of a 5.1 mm. embryo of Chrysemys marginata. 
Embryola. X 80. 

27 Pen sketch of model of entire forebrain of embryo 4b, showing the area 


~ 


from which figure 25 was taken. XX 25. 


538 


PLATE 5 


MORPHOGENESIS OF THE CHOROID PLEXUSES 
PERCIVAL BAILEY 


THE STRUCTURE OF THE THIRD, FOURTH, FIFTH, 
SIXTH, NINTH, ELEVENTH AND TWELFTH 
CRANIAL NERVES 


SUMNER L. KOCH 
From the Anatomical Laboratory of the Northwestern University Medical School 


FIVE FIGURES 


Following the demonstration of unmyelinated fibers in the 
spinal nerves and in the vagus nerve by means of the pyridine- 
silver technique (Ranson, ’11 and ’12; Chase and Ranson ’14), 
Professor Ranson suggested the application of the same method 
to the study of certain of the cranial nerves, with especial refer- 
ence to the presence or absence of unmyelinated fibers. 

The nerves studied were the oculomotor, trochlear, trigem‘nal, 
and abducens of the dog and of man; and the glossopharyngeal, 
accessory, and hypoglossal nerves of the dog, the cat and the 
rabbit. The nerves were obtained by lifting off the skull cap 
and following them distally from their cerebral origin by chip- 
ping away the base of the skull about their foramina of exit. 
The dissected specimens were laid on glass slides and prepared 
by different methods. Some were stained by the pyridine-silver 
method; others were placed in 50 per cent pyridine solution for 
seven days, washed, and then treated with silver nitrate, water, 
and pyrogallic acid as in the pyridine-silver method; others were 
stained by the Pa’-Weigert method and the osmic acid method. 
All were cut and mounted serially. 

The oculomotor, trochlear, and abducent nerves form a natu- 
ral group formerly described as purely motor and consisting of 
large and small myelinated axons, but now recognized as con- 
taining somatic afferent as well as efferent fibers. The nerves 
are described as communicating with the ophthalmic division 
of the trigeminal nerve and with the cavernous plexus of the 
sympathetic system. 


1 Contribution no. 40. 
541 


542 SUMNER L. KOCH 


Gaskell (89) studied the cranial nerves with especial reference 
to their relation to a typical spinal nerve. The oculomotor, 
trochlear and abducent nerves he described as pure efferent 
nerves. The oculomotor was composed of large and small mye- 
linated fibers, in the dog 14.4 to 18 micra and 38 to 5 micra in 
diameter. The smaller fibers passed to the ciliary ganglion, the 
larger to the extrinsic muscles of the eye. The trochlear nerve 
consisted of large myelinated fibers, 14.4 to 18 micra in diameter, 
supplying the external oblique muscle, and small myelinated 
fibers 3.6 to 5.4 micra in diameter, of unknown function. The 
abducens consisted of large myelinated fibers 14.4 to 18 micra 
in diameter, and a few smaller ones, but contained no distinct 
group of small fibers as did the IIIrd and IVth nerves. In the 
roots of these nerves he saw fibrillar structures which he inter- 
preted as the remains of ganglion cells present at an earlier stage 
of their development, and representing the afferent portion of 
the nerves. 

Barratt (98, 799, and ’01) described the oculomotor nerve as 
composed of large and small myelinated fibers, 11 to 15 micra 
and 3 to 5 micra in diameter, in about the ratio of three to one. 
He found unmyelinated fibers also, ‘‘both in the fibrous sheath 
and in the main trunk; in the latter situation usually at the 
periphery.”’ He did not find any communication with the cav- 
ernous plexus nor with the ophthalmic division of the trigeminal 
nerve. The trochlear nerve wes composed of large and small 
myelinated fibers 12 to 19 micra and 4 micra in diameter, in the 
ratio of three to one. There were no unmyelinated fibers pres- 
ent. The abducent nerve he described as composed of large 
and small myelinated fibers, 11 to 17 and 3 to 6 micra in diame- 
ter. A few small twigs composed of small myelinated and un- 
myelinated fibers joined it 25 mm. from its superficial origin, and 
left it again 10 mm. distalward. 

Carpenter (06) studied the development of the oculomotor 
and abducent nerves in the chick. He found that the oculo- 
motor nerve was composed of myelinated axons 3 to 15 micra 
in diameter. The trunk was composed of comparatively large 
axons, with a few scattered small axons, and a zone of small 


STRUCTURE OF THE CRANIAL NERVES 543 


axons at the periphery which passed into the ciliary ganglion. 
He found a communicating branch extending from the ophthal- 
mic division of the trigem‘nal nerve to the ciliary nerve, but 
none to the undivided oculomotor nerve. He did not speak of 
a direct communication between the nerves studied and the 
sympathetic system, but referred to Jegorow’s suggestion that 
a distribution of abducent fibers to the eyeball in birds might 
be accounted for by sympathetie fibers joining the abducens as 
it passed through the cavernous sinus. 

Boughton (’06) found an almost regular increase in the number 
and size of myelinated fibers in the oculomotor nerve of the 
white rat and the cat at different ages. In rats of 730 days 
weighing 414 grams the large fibers averaged 8 1 micra in diame- 
ter, the small 4.3 micra. In cats of 2893 grams the large fibers 
averaged 13.5 micra in diameter, the small 7.2 micra. 

Kopsch (’07) described the ceulomotor nerve in man as com- 
posed of about 15,000 mostly large myelinated fibers grouped in 
a number of secondary bundles. In the roots between the fibers 
were isolated, branched, spherical nerve cells. The trochlear 
nerve he described as ecmposed of about 1200 myelinated fibers, 
the abducens of about 2600 myelinated fibers. All three nerves 
received communicating branches from the carotid plexus of the 
sympathetic system, and from the ophthalmic nerve. 


THE OCULOMOTOR, TROCHLEAR AND ABDUCENT NERVES 


Specimens of the oculomotor nerve of the dog and of three 
humen adults were studied in serial sections. Upon dissection 
at Jeast two fine branches from the sympathetic system could 
be seen joining the nerve. Microscopically sections showed in 
each case clear pictures of a typical motor nerve (fig. 1) com- 
posed of large and small myelina ed axons. In the human the 
ratio of large to small axons was about three to one; in the dog 
it varied between two to one and three to one. In osmie acid 
preparations the large myelinated fibers of the dog averaged 12 
to 16 micra, the small 3 to 6 micra. Close to its peripheral origin 
the fibers of the nerve were grouped in one large bundle, with 
incomplete septa extending toward the center of the nerve. As 


THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL. 26, NO. 5 


n44 SUMNER L. KOCH 


the nerve approached the cavernous sinus the septa became 
complete, and divided it into from five to eight fasicles of vary- 
ing size. Nowhere along the course of the undivided nerve were 
unmyelinated fibers seen within its substance. In sections of 
the intra-cavernous portion of the nerve a few clusters of un- 
myelinated fibers were seen approaching it, but could not be 


followed into its substance. The characteristic grouping of 


cy 
ag 
sas 
(7 
SS 
ans 
IA 
S 
22, 
En 
S 
is} 
ee 
ny) 
07, 
Seo 2 
Soe 


(Se & 
CSS SI Mose, . lo, 

oo soe 1S, ok Ye) EY , ee 

Doe ee ee een, 

INE SS ED MO ee 

SOS meOC mesa Des Cp OUl SOUR RODS, 

esccodO/o E09 OOOO 000 DDG 0, 

pioroge Barat ep caus npeeva acon 


y OO, 
MsOscd2 0020) 
pees avasicce 

SO: 


Le? 
e 
Se 
AS 

O} 


o 
a 
Jos 
is) 
On 
0; 


eae COUaaTG 
Beare 
Sosreoa 
G0sI009. 
easter 
SPSS 
Ps 
Gee 
CR 
soe 
Eats 
EN 
she 
Ss 
Seo; 
Con: 
= 
oe 
Coos: 
'O' 
iS ts) 
IS 
ea 
ao: 
iS; oor 
epee 
aes 
(adow 
PISS 
{0 
aS 
rs 
1 
Sal 
LEG 
Se 
Sass 
30% 


sa 

aaao 
agate 
OEIC 

i=] S26) 
St 

co 

ch 


= 

Asks 
Sass 
Qc 

Sx 500 {2} 

Ca) 

eat 

Ps: 

ae 


Ose OF 
Dig7cg Doses enesom Ura trecoos| y 
CTPA Oooo topo cer cote ed | 


O 
a 
5 


i SEO RSD 9 
NOG eosaeee NO DONC OCOsUOMO LY 
\GhessePoreys Hak rs c Ros SOSOe Que 7, 
OC 00S OLS OO OO O59 0525 CACAO U MO OC aN a0S Oo) LA Neto R TOG / 
Nate obo OO ASO OR ORO nem ACC Ue Nan Re cee OO DOITS BODO nse eH 
NOS OG MO we Oe A DSO COD OER DOD COS eater Oe O es 0622 O DLAC DLS COR opr O Osos OM 
No: Cn OO LOLOOO CO as COONS COs se Orne Re DOO CRD), 
paces Bon eRaGrcanerbes Cap \eocoaa arate 


is OIG: 
OSS Oo con Wo gsce ara Toul, Sacog acoso 
Cosa Coe roe OCaeOSBOOacrepotanun pen 
GSU outa ey an pepe 
16) OHO 
a ~OAO, 


x 


Rese 
soon 
) 
rs? 
a 
Lh, 
Xe) 
co 
ae 
i) 
1 
o 
Sl 
SS 
Oo: 
ae 
ey 
ee 
es 
3 
al 
OIeaaeasscs 
ets 
= 
i) 
30) 
6 
Co 
ATG 
ey 
ea 
O; 
7 
5 
(2) 
ga 
LC 
1S 
nO 
5 
Se 
Ne) 
SS 
Xe) 
=a 


Ler / 


Forno! 


(TEX te {ok elp) 
Pear 


AO SOLt 
Oe OC sx Z 
{) 4 0, } 
BCU OUT EOS eR CEOS OSE AN se eco Sob eee 


Fig. 1 Section of the oculomotor nerve of the dog.  Pyridine-silver. Ocu. 
0, Obj. 6. 


small fibers at the periphery of the nerve, noted by Gaskell and 
Carpenter, could be seen most clearly in osmic acid preparations; 
but not all of the small fibers were grouped at the periphery, nor 
did they all leave the nerve to pass to the ciliary ganglion. The 
cellular structures in the roots of the nerves, mentioned by 
Gaskell and other observers, were not observed in our prepara- 
tions, as even in the most proximal sections the roots had already 
united to form a common trunk. 


STRUCTURE OF THE CRANIAL NERVES 545 


Microscopic sections of the trochlear nerves of the dog closely 
resembled those of the oculomotor nerve; large and small mye- 
linated fibers, 11 to 15 and 2 to 6 micra. in diameter, in about 
the ratio of three to one, with a few fibers of intermediate size. 
No communicating branches from the cavernous plexus or the 
ophthalmic nerve were seen. 

The abducent nerve of the dog near the brain stem consisted 
of a single group of large and small myelinated fibers, 11 to 15 
and 3 to 6 micra in diameter, in about the ratio of three to one. 
Within the cavernous sinus the nerve was joined by a large 
bundle of sympathetic fibers, the majority of which formed an 
intimate union with three of the six or seven fasicles which made 
up the nerve at this point (fig. 2)... More distally the main group 
of sympathetic fibers left the nerve to pursue an independent 
course. 

The striking characteristic of these three nerves is their simi- 
larity in structure, both as to the size of the fibers and the propor- 
tion of large and small fibers. Gaskell and Carpenter noted the 
presence of small fibers in the oculomotor nerve, and considered 
the ciliary ganglion as their destination, though they did not 
say that all the small fibers entered the ganglion. Gaskell 
further noted the similarity in structure between the IlIrd and 
IVth nerves, but said the destination of the small fibers of the 
latter was as yet unknown. The VIth nerve he described as 
composed of large myelinated fibers ‘‘with a few smaller ones; 
with no sign of any distinet group of small fibers as in the IIIrd 
and IVth nerves.”’ In our preparations all three nerves showed 
strikingly similar characteristics as to the size of the fibers and 
the ratio of the small and large fibers. The VIth nerve could 
not be distinguished from the other two by a diminished number 
of small myelinated fibers, but was characterized by the pres- 
ence of unmyelinated fibers from the sympathetic system which 
came into intimate contact with the myelinated fibers and 
travelled distalward with them as far as the nerve was followed. 
The only clue to an explanation of this fact is Jegorow’s sugges- 
tion that a distribution of fibers of the VIth nerve to the eyeball 
in the birds might be accounted for by the presence of sympa- 


546 SUMNER L. KOCH 


thetic fibers which joined the nerve as it passed through the 
cavernous sinus. No such communications from the sympa- 
thetic system were seen in connection with the [lIrd and [Vth 
nerves. Branches of the cavernous plexus, which on gross dis- 
section could be seen to join the IIIrd nerve, on microscopic 


Fig. 2. Section of the abducent nerve of the dog, showing bundles of sympa- 
thetic fibers, a, Pyridine-silver. Ocu. 0, Obj. 6. 


examination were seen to accompany the nerve for a short dis- 
tance and again separate from it farther distalward, without 
entering into intimate contact with its fibers. No communica- 
tion whatever between the IVth nerve and the sympathetic 
system was seen. 


STRUCTURE OF THE CRANIAL NERVES 547 


The unmyelinated fibers seen in connection with the IlIIrd 
and VIth nerves nowhere showed the characteristic arrange- 
ment of the unmyelinated fibers seen in the spinal nerves and the 
vagus. In the IIIrd nerve the unmyelinated fibers remained 
grouped in the nerve sheath and did not enter its substance. 
More distally they separated from the nerve completely. The 
unmyelinated fibers which joined the VIth nerve and entered 
into intimate association with the myelinated fibers were not 
uniformly distributed throughout the nerve substance, but re- 
mained at the periphery, definitely grouped in two or three of 
the fasicles of the nerve. In no sections did these unmyelinated 
fibers from the sympathetic system appear evenly and uniformly 
distributed throughout the nerve substance as do the unmye- 
linated fibers in the spinal nerves and the vagus. 


THE TRIGEMINAL NERVE 


The intracranial portion of the trigeminal nerve of the dog 
and the cat, and of one human specimen was studied. Pyridine- 
silver sections showed large and small myelinated fibers, and in 
addition small numbers of unmyelinated fibers, appearing in 
largest numbers in two fasicles of the sensory portion of the 
nerve. No sympathetic fibers were seen joining the nerve within 
the cranium. In osmie acid preparations of the Vth nerve of 
the dog the large myelinated fibers measured 12 to 16 micra in 
diameter, the small 3 to 6 micra. In the cat the small fibers 
measured 4 to 7 micra, the large 12 to 16 micra, with occasional 
fibers 18 micra in diameter. 


THE GLOSSOPHARYNGEAL, ACCESSORY AND HYPOGLOSSAL NERVES 


The glossopharyngeal nerve is usually described as a mixed 
nerve, and is recognized, according to Herrick’s classification, as 
containing both general and special vsceral efferent fibers, and 
general and special visceral sensory fibers, as well as somatic 
fibers with sensory function. The accessory and hypoglossal 
nerves are pure motor nerves, the former containing general and 
special visceral fibers, the latter special somatic fibers. 


d48 SUMNER L. KOCH 


Gaskell described the small myelinated fibers in the glosso- 
pharyngeal nerve as 1.8 to 3.6 micra in diameter, the large as 
not exceeding 10.8 micra. He found that large myelinated fibers 
were present in all the roots of the accessory nerve, but that the 
small fibers were confined to the bulbar and upper cervical roots. 
Barratt described the [Xth nerve as composed chiefly of small 
myelinated fibers 4 micra in diameter. Kopsch described the 
IXth nerve as consisting of a motor and sensory portion, and 


Fig. 3. From a section of the accessory nerve of the dog, showing a bundle 
of sympathetic fibers joining the nerve. Pyridine-silver. Ocu. 0, Obj. 8. 


recelving sympathetic fibers from the superior cervical ganglion. 
He described the accessory nerve as arising in two parts, the 
accessorius vagi, which joined the vagus as the ramus internus; 
and the accessorius spinalis, which as the ramus externus received 
fibers from the jugular ganglion of the vagus. Chase and Ran- 
son found numerous unmyelinated fibers in the roots of the 
vagus but thought the bulbar rootlets of the accessory nerve 
contained few if any unmyelinated fibers. Kopsch described 
the hypoglossal nerve as arising in 10 to 15 root bundles, and 


STRUCTURE OF THE CRANIAL NERVES 549 


forming anastomoses with the vagus, the upper three cervical 
nerves, and the superior cervical ganglion of the sympathetic 
system. 

Pyridine-silver sections of the roots of the glossopharyngeal 
nerve showed large and small myelinated fibers, and a few 


Fig. 4 Section of the accessory nerve of the dog, showing bundles of darkly 
stained unmyelinated fibers. Pyridine-silver. Ocu. 0, Obj. 8. 


unmyelinated fibers. Sympathetic fibers could be followed into 
the trunk of the [Xth nerve between its two ganglia. In the 
trunk of the nerve the small myelinated fibers outnumbered 
the large fibers nine to one. In osmie acid preparations the 
large fibers measured 9 to 12 micra in diameter, the small 3 to 


550 SUMNER L. KOCH 


5 micra. Sections of a pyridine-silver preparation of the [Xth 
nerve of the cat, just proximal to the superior ganglion, showed 
few large myelinated fibers 12 to 15 micra in diameter, many 
small myelinated fibers, 3 to 6 micra, and a few unmyelinated 
fibers. Sections of the IXth nerve of the rabbit showed a 
similar picture. No connections with the sympathetic system 
distal to the petrous ganglion were seen in any preparations. 
The bulbar rootlets of the accessory nerve are composed chiefly 
of small myelinated fibers 2 to 5 micra in diameter. The spinal 


Fig. 5 From a section of the hypoglossal nerve of the dog, showing bundles 
of darkly stained unmyelinated fibers joining the nerve. Pyridine-silver. Ocu. 
O2Obj2 8: 


portion was composed of large and small myelinated fibers in 
about the ratio of five to one. The majority of the large fibers 
were 9 to 12 micra in diameter, with a few larger fibers of 14 to 
16 micra; the small fibers 2 to 5 micra. In several preparations 
at the point of its separation from the vagus, a few unmyelinated 
fibers could be seen in the ramus externus, arranged in three or 
four groups. Bundles of sympathetic fibers could be traced for 
some distance in the sheath of the XIth nerve, finally entering 
it (fig. 3) and forming well marked clusters of fibers within its 


STRUCTURE OF THE CRANIAL NERVES 551 


substance (fig. 4). In Pal-Weigert preparations the bluish-black 
myelin rings were separated in places by small unstained areas, 
corresponding closely with the location of the unmyelinated 
fibers seen in pyridine-silver preparations. No essential differ- 
ences were seen in the XIth nerve of the cat and rabbit. The 
large myelinated fibers of the spinal root of the accessory of the 
cat were the largest of all the fibers measured, some of them 
having a diameter of 18 micra. 

The hypoglossal nerve of the dog showed a picture closely 
resembling that of the oculomotor nerve, large and small mye- 
linated fibers in about the ratio of three to one, measuring 11 
to 15 micra and 3 to 6 micra in diameter. No unmyelinated 
fibers were seen within the roots of the nerve, but slender bun- 
dles of sympathetic fibers could be followed as they approached 
the nerve, entered its sheath, and finally joined the nerve sub- 
stance (fig. 5). The hypoglossal of the cat and rabbit showed 
similar pictures. 

Nowhere did the unmyelinated fibers seen in sections of the 
XIth and XIIth appear evenly distributed among the mye- 
linated fibers, as in the spinal nerves and the vagus; but always 
grouped in clusters, as in the VIth nerve, and in largest numbers 
near the periphery. 


SUMMARY 


1. Unmyelinated fibers are present in the Vth, VIth, [Xth, 
XIth and XIIth eranial nerves. Those in the VIth, XIth and 
XIIth are probably all derived from the sympathetic system. 
In these nerves they have an arrangement characteristic of 
sympathetic fibers; they are grouped in clusters, and are most 
numerous near the periphery of the nerve. 

2. The oculomotor and trochlear nerves are strikingly similar 
in composition; they are composed of large and small myelinated 
fibers, without any accessions or unmyelinated fibers from the 
sympathetic system. 

3. The abducens is similar to the IIIrd and [Vth nerves as 
regards its myelinated fibers, but in addition receives a large 
number of unmyelinated fibers from the sympathetic system, 


552 SUMNER L. KOCH 


some of which enter the nerve sheath and travel distalward with 
the myelinated fibers. 

4. The accessory and hypoglossal nerves are composed of 
large and small myelinated fibers, and are similar in appearance 
and structure to the IlIrd and IVth nerves. Like the VIth 
nerve, they receive considerable numbers of unmyelinated fibers 
from the sympathetic system which can be followed to the ter- 
mination of the nerves in the musc'es which they supply. 


LITERATURE CITED 


Barrarr, J.O. W. 1899 Observations on the normal anatomy of the 9th, 10th, 
11th, and 12th cranial nerves. Arch. of Neur. from the Path. Lab. of 
the London County Asylums, vol. 1, p. 537. 
1899 On the anatomical structure of the 9th, 10th, llth, and 12th 
cranial nerves. Brit. Med. Jour. 1899, part II, p. 837. 
1901 Observations on the structure of the third, fourth, and sixth 
cranial nerves. Jour. Anat. and Phys., vol. 35, n. s. 15, p. 214. 

Bouauton, 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 differ- 
ent ages. Jour. Comp. Neur., vol. 16, p. 153. 

CarRPENTER, F. W. 1906 The development of the oculomotor nerve, the ciliary 
ganglion and the abducent nerve of the chick. Bull. Mus. Comp. 
Zool. Harv. Coll., vol. 48, no. 2. 

Cuases, M. R. anp Ranson, S. W. 1914 The structure of the roots, trunk, and 
branches of the vagus nerve. Jour. Comp. Neur., vol. 24, p. 31. 

GASKELL, W. H. 1889 On the relation between the structure, function, dis- 
tribution and origin of the cranial nerves; together with a theory of 
the origin of the nervous system of vertebrates. Jour. Phys., London, 
vol. 10, p. 153. 

Koprscu, Fr. Rauber’s Lehrbuch der Anatomie des Menschens, Abt 5, Leipzig. 

Ranson, 8. W. 1911 Non-medullated fibers in the spinal nerves. Am. Jour. 
AmiaitenvOle IQ ips Oie 
1912 The structure of the spinal ganglia and of the spinal nerves. 
Jour. Comp. Neur., vol. 22, p. 159. 


THE BRAIN AND THE. FRONTAL GLAND OF THE 
CASTES OF THE ‘WHITE ANT,’ LEUCO- 
TERMES FLAVIPES, KOLLAR 


CAROLINE BURLING THOMPSON 
Department of Zoology, Wellesley College 


TWENTY-SIX FIGURES 


CONTENTS 


MU OCULC DLO IN erate tain ete eee carehoke the fete a ein od x 8S oe he TERM aS ks ex 353 

Material and methods...... Pe Me Sicha se iciwis al GEC TCT © ees ODE 
The members of the colony of L. flavipes 

ie heraciilin CAStes ANGE tA. COUNT. ccs 5 cs.cu. Wek ee ae tes ww ace 555 

2. Distinguishing characters of the five adult castes and of three types 

OF nymphs.:..7,.. PORE ORR CE Lice erg Ss se a 558 

3. Means of distinguishing the sexes... ee res ae ee 560 

Generalssnatomy: ofthe termite’ Drain. 5....-.- vais... sons binlgn sien EMER nes fe oss 561 

The finer structure of the brain...... fa NE ee UR EMS Sc 50's % 564 

ieerlhe DYSIn SHEA. o.<.06.2+<<~ ss PE rt set a ee 564 

2 aL NES MTOCOCETE DIAL LOWES W7 os a oases a ace ave daw a Oe eee ests viola 564 

3. The mushroom bodies...........-..... PP er 7 dt, eee 566 

Bee NBs COMUTAL DOC Vacs shasien Weld > kin uvihis: da su vids ia SC ROR emt SIs Pers a eM) 

DauOcellu ama, OCEUEY NEXVER,: vin cawais sso vce e's 5.nse > 0 lle eRe ie as as os 570 

65) Dhe optic-lobes:<..<........ Ee ERE ho i ee 572 

Met NCTANDEN NEL MODE. sc. kk ith vcs & soit en woe oa Re Ogee eos sok 574 

8. The tritocerebral lobes and the tritocerebral commissure............ 574 

Gre NearrOnUal: SANGO. 5 oa.co25 a8 ve ec ao x.k 2 ile cyte EROS cle ie eos wae 574 

10. The ventral connectives and the subesophageal ganglion............ 576 

11. The frontal or fontanel gland and the fontanel nerve................ 576 

PIERO SARBUT IAN PRS UIAE Va. gh 59d ain. Ls PAs Finn Ee eno hy CR eR tbs cuss vo 23 589 

Obs WEEDS DINE cert co each deni Roca ah dco e Gots Sh ee neg ERIN MOMMIES Lt, ow Ne 592 


INTRODUCTION 


This paper was undertaken with the idea of studying the finer 
structure of the brains of the different castes of the common 
termite or ‘white ant,’ and to compare them with the brains 
of the castes of true ants, the chief point of interest being that 
both termites and ants are social insects, living in communities 


553 


5A CAROLINE B. THOMPSON 


composed of many types of individuals, but differing greatly in 
their degree of specialization and in intelligence. As the work 
progressed it was found necessary to study the frontal gland in 
detail, and it is now my intention to follow this paper with a 
second, dealing with the development of the frontal gland and 
the differentiation of the different types of brains in the recently 
hatched nymphs. 

The only form to be discussed in the present paper is the 
common termite, Leucotermes flavipes, Kol. As is well known, 
this insect is not an ant, but belongs to the order Isoptera, family 
Mesotermitidae, Holmgren (’11). The species, flavipes, was de- 
seribed by Kollar (°37) ; the genus Leucotermes, formerly included 
in the Linnaean genus Termes, was established by Silvestri (01). 

In the northeastern United States there is only this one genus 
and species of termites; in the southern states there is a second 
species, L. virginicus, with three different genera in the Gulf 
states, Snyder, and several other genera occur in California and 
the southwest, Heath (’03). 


MATERIAL AND METHODS 


My material was collected in Wellesley, Mass., in May, 1915, 
and June, 1916, beneath and within some old planks of wood. 
The fixatives used were Bouin’s Fluid, and Gilson’s Fluid. The 
former, however, is decidedly better for termite nervous tissue, 
and three hours in the fluid gives a good fixation. Whole 
mounts of all the heads were made by staining a long time, 
twenty-four to thirty-six hours, in Conklin’s picro-haematoxylin, 
and then destaining for a day or longer in acid alcohol, clearing 
in cedar oil and mounting with the frontal side up. The heads 
of workers and of both types of nymphs, were embedded in hard 
paraffin, three to four hours, and were sectioned at a thickness 
of 6u. The frontal or horizontal plane proved the most satis- 
factory for sections of the head. In the ease of the soldier, where 
the chitin of the head is very thick, the brain was first dis- 
sected out, under a dissecting lens, stained, embedded, and then 
sectioned. This was also done for the worker, in addition to 
sections of the entire head, the great hardness of the chitin making 


BRAIN OF THE ‘WHITE ANT’ 555 


the sections of this caste frequently imperfect. Whole mounts 
of the true adult and other brains were also made. It should 
be stated here that on account of the internal chitin of the 
tentorium, in addition to the chitin in the cuticula of the skin, 
the termite head is a difficult object to section. Cnly the pro- 
longed time in paraffin has given good results. The sections 
were stained on the slide with (1) Ehrlich’s acid haematoxylin 
and eosin, or (2) iron haematoxylin and orange G. 

For the adult sexual forms, I am indebted to Mr. A. D. Hop- 
kins and Mr. T. E. Snyder of the division of forest entomology 
of the U. 8. Department of Agriculture, who have kindly fur- 
nished me with young and old adults of both the ‘true’ and the 
neoteinic type with short wing pads, and with much other 
termite material that will be utilized later. I wish here to express 
my thanks to both Mr. Hopkins and Mr. Snyder for their kind- 
ness in sending me this material. 


THE MEMBERS OF THE COLONY OF L. FLAVIPES 
I. The adult castes and the young 


The colony of Leucotermes flavipes is composed of many 
different kinds of individuals, some of which are adult or final 
stages, the castes, while others are merely the young, or the 
developmental stages of the castes, their presence, in some 
degree, depending upon the season of the year Each caste 
contains both males and females, caste not corresponding with 
sex, as in the bees and ants. 

The term adult is here used in the general and _ biological 
sense to denote an insect which has undergone its last molt, 
and which has, in general, acquired its final definitive form 
and structure. The phrase “‘in general’ is inserted to cover the 
case of those older females whose abdomens have become great- 
ly enlarged in the course of egg-laying. I am aware that some 
biologists would exclude from the category of adult stages the 
neoteinic reproductive forms and the sterile workers and soldiers, 
on the ground that the former do not possess the outer bodily 
structure characteristic of the species, as seen in the ‘true’ adult, 


THOMPSON 


B. 


. 
4 


CAROLINE 


(SSoydUIM) (0 & 
“UU J 0} 9 ‘SLIIP[OR °c 


(SSopsurIM) 9 & 
“UU Q 0} C ‘STOYIOAL “Ff 


pedojoaep JOU SUBSIO XOQ 


(sped Sut ou) 
“uu § 09 4 ‘uavenb 
SUIAR]-380 posIR[Uy “¢ 


(sped SuiM 410Ys) 

“UU § 0} Gy ‘uaenb 

SUIAR]-550 posIvVlUy °“Z 
(SSUIAL 


jo sqnjs)  “urlUr 
GFL 0} Gy ‘usenb 


BUIAR]-3589 posIe[U *T 


(sped SUL ou) 
“wu ¢ 9 6 “SUIIOF 
OINATSGNS IU1d}O9} 

(sped 


BULM JLOYS) “UU Gy 
03 G9 (8 6 ‘SuTIO] 
OINIYSGNS OIUTOIOIN °G 


(SULM 
con 
Bondy, ‘T 


SuOy) “Wu Cc 
¢79 ‘PS ‘SzN] 


0 }é 


“sydUIAU TdIP[ON “G 


‘ulut 4 ‘sped 
BUIM OU YQIM syduAN “¢ 


(sped Sura 4.10ys) 
“uu G*y 07 G°9 ‘ULIOF 
puodos oy} JO SYduUIAN *Z 


(Sped suIM suo) 
“UU GY OF @'9 “ULIOF 
qsdy oy} jo syduiAN ‘Ty 


\ piieice, aie) ets sm impr (dojeaep 
| jou | IA SUvsIO Xos) 
‘sydutrdu soysoyy “fF | sydurdu 


poproy osaery 


eee me enone (CLOTON 


-op ][[IM suBsI0 xaos) 
syduxu poproy yypeurg | 


“sss sydurAu poe 
-1,Ud.LO FJ IPU/]—So95] 


pedoyaAep suvs10 xog 


SaDViIS IVNIA HO LINGY 


SUDVLS IVINGUNdOTSAAGG HO DNOOXK 


“IOM SHUdIAV THA SUNUALOOVAT 


IsSsDLy WOLf pajdopy 


T WIaVo 


BRAIN OF THE ‘WHITE ANT’ a iF 


and that the latter have undeveloped sex organs. One disad- 
vantage of excluding these forms is that we are then left with 
no general term to denote the full grown stage, which after 
all is the original sense of the word adult. 

The adult stages may be divided into two groups: (1) forms 
with functional and mature sex organs, the ‘kings’ and ‘queens,’ 
of which there are three different types, (a) kings and queens ‘with 
long functional wings, the ‘true’ royal pair, (b) kings and queens 
with short wing pads, the neoteinic substitute pair with short 
wing pads; (c) kings and queens with no wing pads, the neoteinic 
substitute wingless pair, sometimes known as ergatoid or worker- 
like; (2) forms with undeveloped sex organs, (d) the ‘workers,’ 
(e) the ‘soldiers,’ making in all five adult castes. To these five 
types of adults, three more, representing phases of a later devel- 
opment, may be added, namely: the enormously distended fer- 
tilized egg-laying queens of the three queen types (table 1). 

The young, or the developmental stages of the five adult 
castes, occur In many stages, the number of these differing with 
the caste. Irom the egg hatch forms which are said to be all 
alike and undifferentiated, Grassi, Snyder, and from these undif- 
ferentiated nymphs, ‘larvae’ of Snyder,!develop nymphs with large 
heads and nymphs with small heads. The large-headed nymphs 
develop ‘‘after a series of molts and quiescent stages of com- 
paratively short duration,” the whole process requiring less than 
one year, Snyder ('16), into the sterile workers and soldiers. 
The small-headed nymphs undergo a more complex development, 
which ‘‘apparently requires two seasons,’’ and develop into: (1) 
nymphs of the ‘first form,’ with long wing pads, which become 
the true royal pair, the sexual adults with long wings, (2) nymphs 
of the ‘second form,’ with short wing pads, which become the 
substitute sexual forms with short wing pads, (3) nymphs with- 
out wing pads, whose development is not fully understood, and 
which become the substitute wingless sexual adults. 

'The term nymph is used in this paper to denote any developmental stage 
of an insect with incomplete metamorphosis, whether the form possesses wing 
pads or not. Snyder, in accordance with the older authors, applies the term 
‘larva’ to the younger nymphs which possess rudiments of wing pads that can 


only be distinguished with magnification. He uses the term nymph to designate 
forms with wing pads that can be distinguished with the naked eye. 


558 CAROLINE B. THOMPSON 


2. Distinguishing characters of the five adult castes and of three 
types of nymphs 


Before beginning the discussion and comparison of the brains 
of the different forms of L. flavipes some descriptions of the 
types themselves and the means of distinguishing them may be 
of interest. 

Forms with sex organs developed in the adult. 1. Nympus. 
Nymphs of the first form, with long wing pads. These nymphs 
are most numerous in the early spring before the transformation 
into the adult has taken place. According to Snyder they attain 
a length of 7.5 mm. before the last molt, but they may be dis- 
tinguished while much shorter. The color is creamy white, 
with a pale pink or rose pigmentation of the compound eyes. 
The hind wing pads extend back as far as the fourth abdominal 
segment. 

Nymphs of the second form, with short wing pads. These 
nymphs are found at the same time and in about equal numbers 
with the long-winged nymphs. They likewise attain a length 
of 7.5 mm. before the last molt. They may be readily distin- 
guished from the ‘first form’ by the short rudimentary wing pads 
which hardly reach the second abdominal segment. The color 
is creamy white and the compound eyes show no trace of pig- 
ment. 

Nymphs with no wing pads. I have never found these nymphs 
and therefore quote from Snyder, ‘‘ Another type of substitute 
or neoteinic reproductive form, which greatly resembles the worker 
(‘ergatoid’) is developed from young larvae of the sexed forms.” 

2. Aputts. The true royal pair, with long wings. The true 
winged adults are found in the nest or in the air in late spring 
and early summer. The head and body are dark brown, of about 
the same size as the nymphs of the first form, the long filmy 
wings are nearly twice the length of the body. The compound 
eyes are black surrounded by a lighter rim of unpigmented skin. 
The two lateral ocelli are visible as small light spots in front of 
the compound eyes and behind the antennae. The opening of 
the frontal or fontanel gland lies in the median line of the frontal 
surface of the head. 


BRAIN OF THE ‘WHITE ANT’ 559 


The egg-laying true queen. The old egg-laying true queens 
are always found within the nest, often in large ‘queen cells’ or 
chambers, and they may attain a length of 14.5 mm., Snyder 
(16). The head and thorax are of normal size and brown in 
color, the abdomen is enormously distended, and is straw-colored 
with intersegmental bars of brown. The broken off stubs of the 
long wings remain attached to the thorax. 

The neoteinic substitute pair with short wing pads. This pair 
is found within the nest and is similar in length to the nymphs 
of the second form. The body color is light brownish and the 
“compound eyes are brown. The wing pads are now merely short, 
transparent chitinous plates. 

The egg-laying substitute queen with short wing pads. This 
old substitute queen with short wing pads attains a length of 9 
mm., Snyder (716). The abdomen is distended and is creamy 
white with darker chitinous bars, the head and thorax are of 
normal size. The compound eyes are brown.? 

The neoteinie substitute wingless pair. As I have never seen 
this form, I will again quote from Snyder (’16, p. 7), ‘‘Of a pale 
yellowish or grayish color, and having no wing pads (fig. 3; fig. 
4c), individuals of this larval supplementary reproductive form 
are apparently blind and never leave the parent colony, except 
by underground tunnels.” 

The egg-laying substitute wingless queen. The old egg-lay- 
ing substitute wingless queen attains a length of 9 mm. and has 
a distended abdomen, Snyder (16). 

Forms with sex organs not developed in the adult. 1. The work- 
ers. Workers are found in termite colonies throughout the year. 
The smallest of all the castes, the body measures only 5 mm. in 
length. The head is considerably broader from side to side 
than that of the nymphs. The abdomen is shorter, softer, and 
covered with a more transparent skin. The color of the abdo- 
men js usually grayish, owing to particles of wood in the ali- 
mentary canal seen through the skin. No eyes are visible in 

2 The above descriptions of the enlarged egg-laying queens are based upon 
the alcoholic specimens sent to me by Mr. A. D. Hopkins and Mr. T. E. Si yder 
of the U. S. Dept. of Agriculture. 


THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL. 26, NO. 5 


560 CAROLINE B. THOMPSON 


the worker until after staining, when small rudimentary com- 
pound eyes may be distinguished. No ocelli are present. 

2. The soldiers. Soldiers are found in the colony throughout 
the year. The entire body measures 6 to 7 mm. in length, but 
the head is much longer and the abdomen shorter than that of 
the worker. No eyes can be seen until after the head is stained, 
when rudimentary compound eyes even smaller than those of the 
worker become visible. The opening of the frontal gland is in 
the median line of the frontal surface of the head. 


Fig. 1 The abdomens of three individuals seen from the ventral surface. 
A, adult male; B, adult female; C, female nymph with long wing pads; 2 to 10, 
2nd to 10th abdominal segments. Obj. 32, oc. 6, reduced one-half. 


ITI. Means of distinguishing the sexes 


The only differentiation between the two sexes in any caste 
of L. flavipes is in the size of the abdomen, and in the relative 
size and arrangement of the sternites, i.e. the ventral parts, of 
the posterior abdominal segments. 

The sex of any termite may be distinguished by the size and 
the shape of the seventh and eighth abdominal sternites.® 

In the male viewed from the ventral surface (fig. 1, A) the 
seventh and eighth abdominal segments measure about the same 
in length, i.e., in an antero-posterior direction, and the seventh 


3 The ventral surface of the first abdominal segment is entirely covered by the 
metathorax, so that the actual seventh segment is apparently the sixth. 


BRAIN OF THE ‘WHITE ANT’ 561 


segment is shorter than in the female (B). The posterior mar- 
gin of the ninth segment is concave and two slender cerci are 
present on the lateral surfaces of the tenth segment. 

In the adult female (fig. 1, B) the seventh segment is very 
long, with a rounded posterior edge, and completely covers the 
eighth and part of the ninth segment. Jointed cerci are borne 
on the sides of the tenth segment. 

In all female nymphs (fig. 1, C) and also in the female workers 
and soldiers, the seventh segment, although long, does not com- 
pletely cover the small indented eighth segment. Cerci similar 
to those of the male are borne on the median surface of the 
ninth segment and also on the lateral surfaces of the tenth 
segment. 


GENERAL ANATOMY OF THE TERMITE BRAIN 


The present paper will discuss the brains of the nymphs of 
the first and second forms, the true sexual adults, and the workers 
and soldiers. 

It is generally thought today that the insect brain includes 
six regions, corresponding to the embryonic segments of the 
head, three of which are supraesophageal, namely: the proto- 
cerebrum, the deutocerebrum, and the tritocerebrum, and three 
subesophageal, the mandibular, the maxillary, and the labial 
ganglia. Holmgren (’09) finds these six regions in the termite 
brain and discusses at some length the origin of the labrofrontal 
nerves, which he decides is in the tritocerebrum. It need only 
be remarked here that the same six regions may be distinguished 
in Leucotermes, and that in agreement with Holmgren I find the 
labrofrontal nerves arising from the tritocerebrum. Table 2 
names the regions and the parts that are present in the brain 
of L. flavipes. 

Figure 2 is a diagram of the brain of L. flavipes, as seen from 
the frontal surface. The supraesophageal ganglion is drawn 
from an entire mount of a brain dissected out from the head of 
a nymph with long wing pads; the ventral connectives, v.c., and 


562 CAROLINE B. THOMPSON 


TABLE 2 
The parts of the brain 


I. The supraesophageal ganglion. 
A. Protocerebrum  _ 
The protocerebral lobes 

f anterior, dorsal 

\ posterior, dorsal 
(anterior roots 

The mushroom bodies eee body roots 
porterior roots 


The protocerebral commissures 


The central body 
The optic lobes 
B. Deutocerebrum 
The antennal lobes 
C. Tritocerebrum 
The tritocerebral lobes 
The tritocerebral commissure (subesophageal) 
II. The subesophageal ganglion 
The ventral connectives 
The mandibular 
The maxillary ganglion 
The labial 
The nerves of the head 
I. A. The optic nerves 
The ocellar nerves 
The fontanel nerve 
B. The antennal nerves 
C. The labrofrontal nerves to {the nerve to the protocerebrum 
the frontal ganglion Pe labral nerves 
the recurrent nerve 
II. The mandibular nerves 
The maxillary nerves 
The labial nerves 


the subesophageal ganglion, sb.g., are drawn from sections of 
the same form.* 


4 The heads of all termites are held in a somewhat slanting position, making 
a large obtuse angle with the long axis of the body, so that the morphologically 
dorsal surface of the head has become frontal or anterior. The slant of the head 
is much less in termites than in ants, and in the termite soldier the head is almost 
horizontal, the so-called frontal surface being practically dorsal, but for the 
sake of clearness the same terms of direction will be applied to all the castes of 
termites. In describing the entire heads the terms anterior and posterior imply 
toward and away from the frontal surface, in the same sense that Hesse (1901 b) 
uses the terms rostral and caudal; in like manner dorsal and ventral imply toward 
or away from the vertex of the head. 


BRAIN OF THE ‘WHITE ANT’ 563 


The insect brain froms a ring of nerve tissue which encircles 
the esophagus and which is composed of the supraesophageal gan- 
glion,. the ventral connectives, and the subesophageal ganglion. 

The three parts of the supraesophageal ganglion, the proto- 
cerebrum, the deutocerebrum, and the tritocerebrum, are merged 
into a single mass which constitutes the principal part of the - 
brain. To the protocerebrum belong the protocerebral lobes, 
p.l., the optic lobes, o.l., connected by the fibers of the optic 


oc. 


Ock- 2.3. 


7 Bo. Se 


Fig. 2 Diagram of the brain of a nymph with long wing pads, as if seen in 
frontal optical section. A.l., antennary lobe; f.g., frontal gland; f.n., fontanel 
nerve; fr.gn., frontal ganglion; l/.n., ‘abrofrontal nerve; lb.n., labial nerve; m.b., 
mushroom body; md.n., mandibular nerve; mz.n., maxillary nerve; oc.n., oc llar 
nerve; 0.l., optic lobe; p.!., protocerebral lobe; sb.g., subesophageal ganglion; 
tr.cm., tritocerebral commissure; v.c., ventral connective. 


nerves with the compound eyes, the two mushroom bodies, m.b., 
the ocellar nerves, oc.n., connected with the two lateral ocelli, 
and the fontanel nerve, f.n., connected with the frontal gland, 
f.g. ‘To the deutocerebrum belong the antennal lobes, a.l., and 
the antennal nerves of the antennae. The tritocerebrum con- 
sists of the tritocerebral lobes, and the tritocerebral commissure, 
tr.cm., beneath the esophagus. The labrofrontal nerves, l.f.n., 
or the tritocerebral nerves, arise on the inner, median surfaces 
of the tritocerebral lobes, and running upward and forward unite 
in the frontal ganglion, fr.gn., which lies anterior to the main 


564 CAROLINE B. THOMPSON 


part of the brain and directly above the mouth opening. The 
connection of the frontal ganglion and the labrofrontal nerves 
is not shown in the diagram nor is the recurrent nerve which 
runs backward from the frontal ganglion. 

The delicate unpaired nerve from the dorsal surface of the 
frontal ganglion to the protocerebral lobes, and the labral nerves 
from the ventral surface of the frontal ganglion, are figured but 
not labeled. Posterior to the tritocerebral commissure the two 
slender ventral connectives, v.c., run first backward (which can- 
not be shown in this diagram), then downward, and unite to 
form the large subesophageal ganglion, a single mass, consisting 
of the fused mandibular, maxillary, and labial ganglia, and from 
which arise the mandibular, maxillary, and labial nerves. Pos- 
terior to this ganglion the thoracic connectives, not shown in 
the figure, pass upward and then backward into the thorax. 


THE FINER STRUCTURE OF THE BRAIN 
I. The brain sheath 


The entire brain is surrounded by a membranous sheath com- 
posed of a single layer of cells resembling mesenchym cells, their 
fibrous processes making a continuous double walled membrane 
between which the cell bodies and the nuclei are situated. 


ITI. The protocerebral lobes 


The protocerebral lobes (fig. 2, p.l.) form the central part of 
the supraesophageal ganglion; they are continuous on their dor- 
sal surface with the mushroom bodies, on their lateral surfaces 
with the optic lobes, and on their ventral surface with the an- 
tennal lobes, and are also connected by a slender nerve with the 
frontal ganglion, fr.gn., which lies anterior and somewhat ven- 
tral to the protocerebrum. ‘The ocellar nerves, from the lateral 
ocelli, and the fontanel nerve, from the frontal gland, also enter 
the protccerebral lobes, and will be considered in more detail 
under those headings. Like other parts of the brain the proto- 
cerebral lobes consist of a central fibrous core and an outer in- 
vesting layer of nerve cells. Most of these nerve cells are small 


BRAIN OF THE ‘WHITE ANT’ 565 


with round nuclei and a round cell body which appears as a 
mere rim of cytoplasm (fig. 3, A,B), but in the median dorsal 
region, the intercerebral region of Haller, and here and there on 
the ventral surface, the cells are larger with pear-shaped cell 
bodies (fig. 3, C, D). The smallest nerve cells of the proto- 
cerebral lobes are, however, almost twice as large as the adja- 
cent nerve cells belonging to the mushroom bodies (fig. 3, £) 
and to the optic lobes. 

In a series of frontal sections, beginning with the frontal or 
anterior surface (figs. 13 to 20) it will be seen that the proto- 


B 


Fig. 3 Nerve cells from the protocerebral lobes and the mushroom body. 
A, small cells, from the lateral region of the protocerebral lobes; B, medium 
sized cells, from the same region; C, large cells, from the median ventral region 
of the protocerebral lobes; D, large cells, from the intercerebral region; EF, cells 
from the mushroom body. Homog. immers. 1.8 mm., oc. 6. 


cerebral lobes are at first entire (fig. 13), then, farther back in 
the series (fig. 14) the two great anterior roots of the mushroom 
bodies penetrate deep into the fibrous core of the protocerebral 
lobes, dividing them into a median and two lateral portions, and 
this division is further continued by the stalks of the mushroom 
bodies (figs. 15, 16). 

a. The protocerebral commissures. About the middle of the 
protocerebral lobes, in the plane of the central body (fig. 15) the 
lateral and median parts of the lobes are again connected by a 
stout fiber tract passing from side to side beneath the central 
body and connected also with the antennal lobes. This tract 
or commissure is homologous with the so called ‘ventral com- 


566 CAROLINE B. THOMPSON 


missure’ of the hymenoptera and other insects. The lateral 
portions of the protocerebral Icbes are also connected by two 
dorsal protocerebral commissures, the anterior, a delicate strand 
of fibers, lying above the central body (figs. 15, 16, a.cm.) the 
posterior, on the same level as the first, lying immediately pos- 
terior to the central body, fig. 18, p.cm. These two commissures 
are homologous with the two dorsal commissures found in ants, 
and it is interesting to note that the posterior one, the ‘Bricke’ 
of the older authors, has the same characteristic wide inverted 
N shevpe. 

In figures 15 to 18, the connection of the optic lobes with the 
protccerebral Jcbes may be seen. 

Figures 19 to 20 show the relation between the protocerebral 
lobes, the deutocerebrum or antennary lobes, and the trito- 
cerebrum. 


ITI. The mushroom bodies 


The two mushroom bodies, or the globuli, as Holmgren terms 
them, are now generally considered the chief centers of intelli- 
gence in insects. They appear, in frontal views of the head 
(figs. 8 to 12, m.b.), as two slightly grooved projections from the 
dorsal surface of the brain, the frontal gland occupying the 
space between them. Each mushroom body consists typically 
of two lobes, which in the termites are not distinctly separated 
but form cone continuous mass of tissue, the lobes being indi- 
cated on the exterior of the brain merely by a slight groove or 
depression of the surface. In sections (figs. 15 to 18) it may be 
seen that the mushroom bodies are composed of an outer nerve 
cell layer and an inner fibrous portion which forms the cups or 
calyces, the stalks, and the three pairs of roots. 

a. The nerve cells of the mushroom bodies. The nerve cells of 
the mushroom bodies are all of the same size and are among the 
smallest cells of the brain (fig. 3, #). They have a round or 
oval nucleus and so small an enveloping layer of cytoplasm that 
it cannot be distinguished, even with the immersion lens, except 
at the distal end where the axon is given off. The chromatin 
masses of the nucleus are all about the same size, forming a 


BRAIN OF THE ‘WHITE ANT’ 567 


peculiar characteristic pattern by which these cells may be 
recognized. 

Although the nerve cells are of similar size throughout the 
mushroom bodies, they are differentiated into groups, or zones, 
according to their position. In the center of each cup or calyx, 
as seen in section (fig. 4), ies an oval mass of cells, 7, which is 
homologous in position with the central oval mass of large cells 
found in the bees, Jonescu (09), and in the ants, Pietschker (’11), 
Thompson (712), and which I have termed Group I. On each 
side of Group I, are broad wedge-shaped masses of cells (fig. 4, 


Fig. 4 Diagram of a mushroom body, showing the nerve cell layer divided 
into groups, the calyces, and the beginning of the stalk. J, cell group I; JJ, cell 
group II; JJ/, cell group IIT; gl., glia cells. 


II), which occupy most of the dorsal surface of each lobe and 
whose inner margins overlap and enclose the central group. I. 
These masses, which appear separate in sections, form a con- 
tinuous zone if seen in surface view, and are homologous with 
Group II of ants. Again, in each lobe, on each side of Group II 
lie smaller masses of cells which form the lateral surfaces of the 
lobes (fig. 4, J7Z). There are only three, instead of four, of these 
cell masses, because the inner lateral surfaces of the two lobes 
are in contact and their cells are continuous. These groups are 
equivalent in position to Groups III and IV of the ants. 

b. The fibrous core of the mushroom bodies. The cups or calyces 
of the inner and outer lobes of the mushroom bodies are com- 
posed (1) of converging bundles of fibers, the axons of the three 


568 CAROLINE B. THOMPSON 


zones of nerve cells just described, and (2) of masses of glia cells 
(fig. 4, gl.) resembling the ‘glomeru'i’ of the antennal Jobes, which 
envelop the fiber bundles-on all sides and unite the calyces of 
the two lobes into one continuous whole. 

The stalks of the mushroom bodies are the continuations 
downward and inward of the fiber bundles of each lobe. These 
lie side by side, the fibers from the inner lobe on the median side 
of those from the outer lobe, and for a short distance they remain 
distinct. The two stalks, surrounded by a delicate sheath, pene- 
trate deep down into the fibrous core of the protocerebral lobes 
and then run forward. There is no ‘decussation’ of fibers such 
as is seen in the hymenoptera. In the same frontal plane with 
the central body (fig. 15) the distal ends of the stalks lie beneath 
the central body, and at this point each stalk gives rise to three 
roots, the anterior, the central body, and the posterior roots. 
The central body roots, which are the shortest, pass upward and 
directly into the central body (fig. 16, c.b.r.). The anterior roots 
which are the longest, bend sharply upward and forward, mak- 
ing an elbow with the stalks from which they have arisen (fig. 
14, a.r.m.b., m.b.s.). In sections that are farther forward in the 
head, the anterior roots may be seen as two great bundles of 
fibers curving latero-dorsally and dividing the protocerebral lobes 
in the manner already described. On reaching the dorso-lateral 
margins of the protocerebral core, the anterior roots curve in a 
posterior direction and are seen in sections as two detached 
masses on the outer sides of the mushroom bodies (figs. 15, 16, 
a.r.m.b.), then (fig. 14) returning again forward, the narrow dis- 
tal ends turn downward and terminate among the nerve cells 
of the anterior part of the protocerebral lobes. 

The posterior roots branch off from the stalks together with 
the central body roots, and, after the latter have passed into 
the ventral surface of the central body, pass dorsalward, posterior 
to the central body and to the posterior dorsal commissure, there 
expanding into two large and very prominent lobes which nearly 
fill the intercerebral region (figs. 19, 20, 28, 26, p.r.m.b.). The 
ventral part of these lobes or roots is connected with the proto- 
cerebral fibrous core, the dorsal part, however, projects back- 


BRAIN OF THE ‘WHITE ANT’ 569 


ward for some distance, evidently giving fibers to and receiving 
fibers from the posterior cells of the mushroom bodies, and then 
ending in about the same plane with the beginning of the trito- 
cerebral lobes. 

c. Comparison of the mushroom bodies of the different castes of 
L. flavipes. The mushroom bodies differ little in the different 
castes except in size, and even in this respect there is not a very 
great dissimilarity (figs. 8 to 12). 

The worker has the largest mushroom bodies, largest by actual] 
measurement and in the estimated number of nerve cells; the 
nymphs of the first and second forms have mushroom bodies 
about sitnilar in size, but slightly smaller than those of the 
worker; the soldier has the smallest mushroom bodies of any 
easte. The cells of the mushroom bodies of the worker and 
soldier penetrate into the intercerebral region (figs. 22, 25), but 
merely border upon it in those of the nymphs of the first and 
second forms. 

My only true adult material, as already stated, was alcoholic 
and much of it seemed to have shrunk. It will be noted that in 
the true adult brain outlined in figure 9 the mushroom bodies are 
farther apart than in the nymph of the first form from which 
this adult has developed (fig. 8) and the frontal gland has in- 
creased in size. Some other better preserved true adult brains 
are considerably larger than the brain of a nymph of the first 
form, and the mushroom bodies are likewise larger. 

Holmgren (’09) emphasizes the great deterioration of the brain 
tissue which he observes in the older enlarged and egg-laying 
queens. My material has been preserved in alcohol and is not 
the best for finer study, but the examination of several old and 
enlarged queens leads me to believe that in L. flavipes this 
deterioration of the brain has not taken place. 

d. Comparison of the mushroom bodies of termites and hymenop- 
tera. A comparison of the Leucotermes mushroom body with 
those of ants and bees shows, as one would naturally expect, 
that the termite mushroom body is much more simple and 
primitive. This primitive condition is apparent in the small 
and uniform size of all the nerve cells, especially in the cell 


570 CAROLINE B. THOMPSON 


group I; in the presence of three zones of cells instead of the 
four found in ants; in the incomplete differentiation of the two 
lobes, whose cells are not separated by a deep furrow, as in ants, 
and whose two cups or calyces are completely fused by inter- 
vening masses of glia cells; in the shallowness of the cups; finally, 
in the smaller size of the entire mushroom bodies and their slight 
differentiation in the different castes. 


IV. The central body 


The central body is situated in the central frontal plane of 
the protocerebral lobes, embedded in their fibrous core, directly 
beneath the intercerebral region. In form as well as in’ position 
the central body of L. flavipes resembles that of bees and ants. 
It has two parts (fig. 15, c.b.), a curved dorsal portion, composed 
of fiber bundles that are radially arranged with intervening 
spaces, and a flatter ventral part, also fibrous. No nerve cells 
are found within the central body, as in ants, where small nerve 
cells cecupy the spaces between the radial bundles of fibers, but 
a few scattered nerve cells are occasionally found along the outer 
surface. 

The structure of the central body is the same in all the castes, 
but the size varies with the size of the different brains (figs. 15, 
BOIS): 

Lying beneath the central body are two small round bodies 
which were formerly known as the ‘“‘tubercles of the central 
body” and also as ‘ocellar glomeruli.’ I have shown, Thompson 
(12, ’14), that in ants and in the bumble bee these previously 
misinterpreted bodies are the posterior roots of the mushroom 
bodies. In L. flavipes these rounded bodies are in connection 
with the central body roots of the mushroom bodies (figs. 15, 
16, c.b.r.), which, it will be remembered, arise from the distal 
ends of the stalks in close proximity to the posterior roots. 


V. Ocelli and ocellar nerves 


Two simple eyes or ocelli are present in the adults and nymphs 
of the sexual forms, and are situated on the lateral surfaces of 
the head, in front of the compound eyes and behind the antennae. 


BRAIN OF THE ‘WHITE ANT’ 571 


In the head of the true adult (fig. 9, 0.c.) the ocelli may be readily 
seen as small colorless spots which stand out sharply from the 
surrounding brown skin, but in the heads of nymphs (figs. 8, 10), 
whose skin is not pigmented, the ocelli are not visible except in 
sections. 

The ocelli are of a very simple and primitive type, Hesse 
(01 6), without a lens and with little or no pigment. The outer 
surface of the hypodermis above the ocelli is slightly convex, 
the inner surface is very concave, the latter caused by the sudden 
thinning of the inner cuticula, and in this concavity the bulb- 
like ocelli are situated. In general contour the ocelli resemble 
the tactile buds found in the skin of Amblystoma punctatum. 
The visual cells are Jong slender and curving, with spaces between 
their bases. As to the finer intracellular structure, I am not 
prepared to make a statement at this time. I shall also leave 
for further study the question whether the distal ends of the 
visual cells lie between the hypodermal cells, a primitive position, 
according to Hesse, or wholly, and secondarily, beneath them; 
although the first view would seem to be upheld by my present 
material. . 

In a frontal section of the anterior part of the brain of a 
nymph with long wing pads, the ocelli (fig. 13, oc.) may be seen 
on each side of the section, just beneath the hypodermis, and 
covered by a thinner layer of cuticula, the inner cuticula, 7.cw., 
being absent at this point. From the ocelli the slender ocellar 
nerves, lying just outside the brain sheath, run in toward the 
median line and enter the nerve cell layer of the protocerebral 
lobes just above the anterior roots of the mushroom bodies. 
Passing down into the intercerebral region, the ocellar nerves 
(figs. 15 to 18, oc.n.) run backward and finally enter the dorsal 
surface of the protocerebral lobes just behind the posterior 
dorsal commissure and in the same frontal section as the fontanel 
nerve from the frontal gland (fig. 19, oc.n., f.n.). At no point 
along their entire Jength do the ocellar nerves expand into ocellar 
lobes such as are present in the bees and ants. After my first 
cursory examination of the brain sections of L. flavipes, I was 
inclined to assign the réle of ocellar lobes to the large lobes, 


572 CAROLINE B. THOMPSON 


p.r.m.b. seen in figures 19, 23, 26, on the ground that these lobes 
occupy a position similar to that of the ocellar lobes of ants and 
are in the neighborhood of the entering ocellar nerves. A very 
careful examination of a large number of sections, however, 
proved that there is no connection between the ocellar nerves 
and these adjacent lobes, and that there is direct connection be- 
tween the mushroom body stalks, the protocerebral lobes, and 
the lobes in question, which are, as shown in a preceding section, 
the posterior roots of the mushroom bodies. 


VI. The optic lobes 


The optic lobes are situated on the lateral surfaces of the proto- 
cerebrum and are present in all the castes of L. flavipes, although 
they are well developed only in the sexual forms (figs. 8 to 12, 
o.l.). They are continuous with the protocerebral lobes and 
consist of an outer layer of very small nerve cells, similar in size 
to those of the mushroom bodies, and an inner fibrous portion 
that is subdivided into the outer, middle, and inner fiber masses, 
and the outer and inner crossings of fibers found in the typical 
insect brain, Berlese (09). 

The optic lobes of the nymph with long wing pads are the 
largest and most highly developed of any caste studied. The 
outer fiber masses. (fig. 17, 0.f.) are slender and elongated in a 
dorso-ventral direction. Nerve fibers pass in to them from the 
compound eyes, and continue inward, as the outer crossing, to 
the middle fiber mass, mf., the largest of the three masses. 
Fibers again cross between the middle and the inner fiber masses, 
i.f., making the inner crossing. The small inner fiber masses 
are directly continuous with the fibrous core of the protocerebral 
lobes. 

In the nymph with short wing pads (fig. 10) the optic lobes 
and the compound eyes are slightly smaller than in the nymph 
just described, but the relative size and arrangement of the parts 
is the same. 

In the worker, although the optic lobes are very much reduced 
in size, they are readily seen in surface views (fig. 11) as small 


BRAIN OF THE ‘WHITE ANT’ 573 


projections from the lateral surfaces of the brain, and a slender 
optic nerve passes from each greatly reduced compound eye to 
the optic lobe. In a frontal section of the worker brain (fig. 25) 
the small inner, middle, and outer fiber masses may be distin- 
guished in the optic lobes, and also the fibers of the distal part 
of the optic nerve. By carefully following the optic nerve in 
subsequent sections it may be traced to the vestigeal compound 
eye. 

The condition of the compound eyes evidently varies con- 
siderably in different termites. Holmgren (’09), describing the 
optic lobes in the worker of Eutermes, writes as follows: 


Die Reduktion der Facettaugen hat eine entsprechende Reduktion 
der Sehganglien mitgefiihrt, Die Sehnerven sind vielleicht vorhanden 
aber enthalten nur wenige Nervenfiidchen und von den bei den Ge- 
schlechtstieren, besonders den jungen, so wohlentwickelten Sehganglien 
kann man gar nichts entdecken. Die Seiten des Protocerebralganglions 
sind somit kreisférmig abgerundet. 


The optic lobes of the soldier, like those of the worker, appear 
as small rounded projections on the lateral surfaces of the proto- 
cerebrum (fig. 12, 0.l.), with optic nerves extending to the still 
smaller vestigeal compound eyes, c.e. On account of the very 
hard chitin of the soldier’s head no attempt was made to sec- 
tion the entire head. The brain was dissected out, under a dis- 
secting microscope, stained, embedded, and then cut in frontal 
sections. A section through the optic lobes (fig. 22, 0.1.) shows 
that the inner, middle, and outer fiber masses, although small, 
are clearly distinguishable, and that fibers enter the outer fiber 
mass from the optic nerve. 

Holmgren (’09) states of the soldiers of EKutermes: 


Die Soldaten sind ja blind und deshalb sind die Sehganglien voll- 
stiindig verschwunden. Man kann von denselben kein Spur ent- 
decken, obwohl ein sehr dinner n. opticus meistens vorhanden ist. 
Dieser fungiert aber als Hautnerv. Zufolge des Fehlens der Sehganglien 
sind die Seiten des Protocerebrums beinahe kreisrund. 


The optic lobes of the true adult have not been studied in 
sections. 


574 CAROLINE B. THOMPSON 


VII. The antennal lobes 


The antennal lobes compose the deutocerebrum, or the second 
brain segment, and are continuous with the anterior and ventral 
part of the protocerebrum (fig. 2, a.l.). In form they are elon- 
gated masses that continue into the antennae as the antennal 
or olfactory nerves. On their inner lateral surfaces the antennal 
lobes are continuous with the third brain segment, the trito- 
cerebrum. 

In sections it may be seen that the antennal lobes have an 
outer nerve cell layer containing both large and small cells, and 
an inner fibrous core that contains scattered masses of glia cells, 
the so-called ‘glomeruli.’ 

The relative size of the antennal lobes differs very slightly in 
the different castes of L. flavipes (figs. 8 to 12). The antennal 
lobes are largest in the nymphs with long wing pads and in the 
worker, but are smaller in the nymph with short wing pads and 
in the soldier. Holmgren states that these lobes are much larger 
in the worker of Eutermes than in the other castes: ‘‘die Deuto- 
cerebralganglien der Arbeiter sind verhiltnissmiassig grésser und 
enthalten eine auch absolut bedeutend gréssere Zahl Ganglien- 
zellen als bei den Geschlechtsindividuen.”’ 


VII. The tritocerebral lobes and the tritocerebral commissure 


The tritocerebral lobes are small lobes extending ventrally 
along the sides of the esophagus and continuous with the inner, 
lateral, surfaces of the antennal lobes (fig. 20). From the inner 
median surfaces of the tritocerebral lobes arise a pair of nerves 
that run forward alongside of the esophagus and unite in the 
frontal ganglion; these are the tritocerebral or labrofrontal nerves 
(fig. 2, l.f.n.). Beneath the esophagus the tritocerebral lobes 
unite in the slender tritccerebral commissure (fig. 2, ér.cm.), 
which is almost devoid of any investing nerve cells. 


IX. The frontal ganglion 


The frontal ganglion is a small mass of nerve tissue situated 
just beneath the clypeus, the sclerite of the head to which the 
labrum is attached, and above the mouth opening, somewhat 
antero-ventral to the brain (figs. 2, 8, fr.gn.). 


BRAIN OF THE ‘WHITE ANT’ 575 


In sections (figs. 13, 21, 24) the frontal ganglion is rather tri- 
angular in appearance, the apex composed of nerve cells and the 
base of nerve fibers. The nerves connected with the frontal 
ganglion are (1) the paired tritocerebral or labrofrontal nerves, 
l.f.n., (2) the paired labral nerves, la.n., (3) an unpaired nerve to 
the protocerebrum, (4) the unpaired recurrent nerve, 7.7. 

The tritocerebral or labrofrontal nerves, l.f.n., arise from the 
tritocerebrum as described under that heading, and their anterior 
ends are connected with the latero-ventral surfaces of the frontal 
ganglion. 

From the median ventral surface of the frontal ganglion a 
bundle of nerve fibers emerges that runs first ventralward for a 
short distance and then divides into delicate right and left 
branches that pass to right and Jeft through some muscle fibers 
in the clypeus, continuing down and forward into the labrum as 
two delicate but distinct nerves. Those branches are distributed 
throughout the labrum. I have termed these nerves the labral 
nerves. In this I am apparently not in agreement with Holm- 
gren (’09) who does not describe the labral nerves as arising from 
the frontal ganglion but speaks as though the labrum were inner- 
vated directly from branches of the labrofrontal nerves, which 
is not the case in L: flavipes. In all the sections which I have 
examined the nerves of the labrum have no direct connection 
with the labrofrontal nerves but arise indirectly, from the base 
of the frontal ganglion as described above. ‘To quote Holmgren: 


Ganglion frontale liegt unmittelbar vor dem oberen Schlundganglion 
und ist mit diesem durch einen unpaaren Nervenfaser, der im Proto- 
cerebrum eintritt, verbunden. Vor dem Ganglion frontale gehen nach 
vorn Nerven aus, welche einige clypealen Muskeln innervieren. 

An den Seiten des Ganglion frontalekommen die tritocerebralen Kon- 
nective ein, und hinten lauft der nervus recurrens aus, um oberhalb des 
Darmes sich nach hinten zu begeben. 

Aus Tritocerebrum werden folgende Teile innérviert: 

1) Die ganze oben beschriebene Labralmuskulatur von n. labro- 
frontalis 

2) Die Labraldriisen von n. labrofrontalis. 


In agreement with Holmgren I find a very delicate unpaired 
nerve running between the dorsal apex of the frontal ganglion 


THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL. 26, NO. 5 


576 CAROLINE B. THOMPSON 


and the ventral surface of the protocerebral lobes, also the un- 
paired recurrent nerve which arises from the posterior surface 
of the frontal ganglion. This latter nerve (figs. 14 to 26, r.n.) 
runs backward above the esophagus into the thorax and at a 
short distance posterior to the supraesophageal ganglion expands 
into the socalled ‘esophageal ganglion.’ 


X. The-ventral connectives and the subesophageal ganglion 


The two slender ventral connectives (fig. 2, v.c.) arise from the 
posterior surface of the tritocer.bral commissure. They con- 
tinue backward side by side, their inner surfaces touching, 
beneath the esophagus and above the brcad tentorial band of © 
chitin that is present in the interior of the head. They become 
still smaller, and pass down through the tentorial aperture, 
finally merging into the dorsal surface of the subesophageal 
ganglion. 

The subesophageal ganglion (fig. 2, sb.g.) is large and consists 
of a fibrous core and a thick investing layer of nerve cells on 
the ventral surface. From the anterior end arise the mandibu- 
lar nerves, md.n., from the posterior end, the maxillary, mz.n., 
and the labial, Jb.n., nerves. The branches of these nerves, 
described by Holmgren, have not been traced. Immediately 
behind the point of exit of the Jabial nerves the subesophageal 
ganglion ends, or rather is continued into the thoracic connec- 
tives, which pass upward and out of the head through the great 
tentorial aperture. 


XI. The frontal or fontanel gland and the fontanel nerve 


The frontal or fontanel gland of the termites has long been 
an object of much interest and speculation. It is situated in 
the median Jine of the frontal surface of the head beneath the 
so-called ‘fontanel,’ which is defined by Brues (715) as ‘“‘a small, 
depressed pale spot on the front of the head between the eyes 
(isoptera).”’ 

In surface views of the heads of the castes of L. flavipes, 
(figs. 8 to 12) the frontal gland appears as a rather triangular or 
rounded mass lying in the space between the two mushroom 
bodies. In the adult sexual forms and in the soldier a tiny round 


BRAIN OF THE ‘WHITE ANT’ 577 


opening may be observed in the cuticula above the anterior end 
of the gland, but I have not been able to distinguish any such 
opening in the worker or in the two nymphs. The opening is 
situated in a shallow depression of the surface, the fontanel, 
which is paler than the surrounding skin. 

The frontal gland differs greatly in size in the various castes. 
It is smallest of all in the worker (fig. 11) and greatly elongated 
in the soldier (fig. 12). In neither of these castes does the 
frontal gland completely fill the space between the mushroom 
bodies. In the sexual forms the true adult (fig. 9) has the largest 
frontal gland; the nymph with long wing pads (fig. 8) has a 
gland of similar proportions but smaller; the nymphs with short 
wing pads (fig. 10) has an even smaller gland. In all the sexual 
forms the frontal gland completely fills the space between the 
mushroom bodies. 

Looking down into the frontal gland as it is seen in surface 
views is like looking into a shallow crater. The edges rise into 
folds which approach each other and nearly meet at the antero- 
ventral end and enclose the central lumen. Sections show that 
the marginal cells of the gland are directly continuous with the 
hypodermal cells and that the cuticula above the center of the 
gland is slightly depressed and thin, consisting only of the pri- 
mary or outer cuticula, the secondary, inner, cuticula ending 
at the gland margin. This depression and thinning of the cutic- 
ula is the structural explanation of the pale ‘fontanel’ spot seen 
on the outer surface of the head. 

In the series of sections seen in figures 13 to 20, the first indi- 
cation of the frontal gland is seen in a group of elongated hypo- 
dermal cells (fig. 13, f.g.) lying in the median line directly beneath 
the cuticula, which is here of normal thickness. This group of 
cells forms the anterior margin of the frontal gland and it should 
be noted that it occurs in the same frontal section as the simple 
eyes, or ocelli, oc., a point which will be further discussed later. 
Figure 14, which is six sections farther back in the series, includes 
a section of the anterior part of the frontal gland, f.g. The 
high columnar cells of the gland are continuous with the low 
cuboid hypodermal cells; the cuticula is slightly depressed and 


578 CAROLINE B. THOMPSON 


the secondary, inner, layer is lacking above the lumen of the 
gland. In figures 15, 16, 17, the deeper, central, part of the 
gland is shown; figures 18, 19, contain the posterior outpocketing 
of the gland, not connected with the margin in the figures, and 
since these sections are considerably farther back in the series, 
in the longest part of the head, the cuticula and hypodermal 


Fig. 5 Frontal section of the frontal or fontanel gland of a nymph with long 
wing pads. B.m., basement membrane; hyp., hypodermis; 7.cu., inner cuticula; 
o.cu., outer cuticula; 1, long slender cells, probably not yet secretory in function; 
2, swollen cells, probably glandular. Homog. immers. 1.8 mm., oc. 6, reduced 
one fifth. 


cells are now some distance from the brain, and, for consider- 
ations of space, are no longer included in the figures. Figure 
20 shows the posterior wall of the frontal gland which ends in 
the next section. 

a. The finer structure of the frontal gland. <A closer study with 
the immersion lens shows that the frontal gland is composed of 


‘ BRAIN OF THE ‘WHITE ANT’ 579 


the following parts, which are present, with slight modifications, 
in all the castes: (1) the epithelial cells, (2) the basement mem- 
brane, (3) muscle fibers attached to the. basement membrane, 
(4) the fontanel nerve, an unpaired nerve running from the 
frontal gland to the protocerebral lobes, and here described for 
the first time. 

1. Nymph with long wing pads. The epithelial cells (fig. 5, 
1, 2) are of two kinds, of which the first far outnumber the 
second: /, long slender columnar cells, with rounded distal ends, 
the distal cytoplasm taking a slightly deeper stain than the 
proximal portion, and the large oval nucleus situated near the 
base, 2, larger swollen cells whose cytoplasm contains vacuoles 
and irregular masses of a deeply staining substance. The nuclei 
of these cells are evidently basal but they are so crowded by 
the adjacent cells that I have been unable to distinguish them. 

It will be noted that the lateral cells are more closely crowded 
together than those on the floor of the gland, in the median line, 
where the outlines of several cells may be clearly traced. It will 
also be noted that the marginal cells of the gland are lower and 
are directly continuous with the hypodermal cells. Mitotic fig- 
ures are present here and there among the cells of the frontal 
gland in both nymphs, but no cuticula on the distal surfaces of 
the cells and no extracellular secretion have been observed. The 
only cells which appear capable of producing a secretion in the 
frontal gland of the nymph with long wing pads are the few 
scattered swollen cells marked 2, and these seem to have not 
yet discharged their secretion. This and the fact that there is 
no external opening may indicate that the frontal gland is not 
yet functional in this nymph. 

The basement membrane (fig. 5, b.m.) upon which the epi- 
thelial cells of the frontal gland rest, is continuous with that of 
the hypodermal cells and is considerably thickened along the 
lower, ventral, surface. This membrane is composed of mesen- 
chym cells and fibers and is beset with many irregular openings. 
In some specimens it is very delicate but thick and firm in others, 
and takes a deep pink stain with eosin, yellow with orange G. 
The ventral surface of the basement membrane near the posterior 


580 CAROLINE B. THOMPSON 


end of the frontal gland is prolonged into a slender median 
process that passes downward posterior to the protocerebrum 
and connects the frontal gland with the great tentorial mem- 
brane that divides the interior of the head into the two parts 
in which lie the supra- and the subesophageal ganglia, and 
through which the ventral connectives pass. The basement 
membrane is similar in structure in all the castes. 

Muscle fibers may be seen attached to the posterior or caudal 
surface of the basement membrane in the true adult, in surface 
views, but I have not observed them in either of the nymphs; 
muscle fibers are also attached to the basement membrane in 
the soldier and in the worker. | 

The fontanel nerve (figs. 2, 19, 23, 26, f.n.) is the slender un- 
paired nerve which makes its exit through apertures of the 
median ventral process of the basement membrane of the frontal 
gland, and which passes vertically downward and enters the 
median dorsal surface of the protocerebral lobes at a definite 
point, namely: between the large posterior roots of the mush- 
room bodies and in the same frontal plane in which the ocellar 
nerves enter the protocerebral fibrous core. This is the first 
time that this nerve has been described or figured in any termite 
brain. It is present in a similar position in all the castes and 
phases of L. flavipes that have been examined, and it is undoubt- 
edly a strand of nerve fibers, although a very delicate one. I 
have provisionally termed this nerve the fontanel nerve, and we 
shall return to it again after a discussion of the frontal gland in 
the other castes. 

2. The true adult. I have not yet succeeded in making good 
sections of the true adult’s frontal gland, but in a surface view 
of an entire brain and frontal gland dissected out and mounted 
it is most evident that the inner surface of the frontal gland is 
lined with a wavy chitinous cuticula, the channels of which 
converge toward the anterior opening on the surface of the head. 

Although I have unfortunately not studied the true adult 
frontal gland in sections and can not therefore describe the epi- 
thelial cells, the facts that it is larger than in the nymph with 
long wing pads, and that it has acquired a chitinous cuticula, 


BRAIN OF THE ‘WHITE ANT’ 581 


and an opening to the exterior, seem to indicate that this gland 
probably produces secretion and is therefore functional. The 
true adult frontal gland in this respect would represent a later 
phase in development than that of the nymph with long wing 
pads. 

3. The nymph with short wing pads. The frontal gland of 
this nymph is similar in all respects except size to that of the 
nymph with long wing pads. Although this gland as a whole is 
smaller (fig. 6, A), its component epithelial cells are similar in 
height and width to those of the gland just ascribed. The two 
types of cells, the numerous long slender ones, 1, and the few 
swollen gland (?) cells, 2, are also present. No cuticula is dis- 


Fig. 6 Sections of the frontal gland. A, the median part of the gland of a 
- nymph with short wing pads; B, the lateral part of the gland of a soldier. Homog. 
immers. 1.8 mm., oc. 6, reduced one fifth. 


tinguishable on these cells and there is no external opening. 
This gland is evidently not yet functional. 

4. The soldier. The cells of the elongated frontal gland of the 
soldier form a continuous syncytium, a part of which is shown 
in section in figure 6, B. The syncytium rests upon a thin base- 
ment membrane, the nuclei are rather widely separated, and the 
distal, inner, surface is bordered by a wavy glassy and porous 
cuticula. The entire inner surface is thrown into folds but the 
average height of the component cells is lower in the soldier than 
in the nymphs. The cells are all of one kind, and are evi- 
dently glandular in function, the abundant cytoplasm contain- 


582 CAROLINE B. THOMPSON 


ing a network of intracellular canals that open into the lumen 
of the gland by pores in the cuticula. In surface views of the 
soldier head strands of muscle fibers seem to be connected with 
the lateral and posterior surfaces of the frontal gland, and, as 
was stated above, an opening in the cuticula of the head above 
the gland may be clearly seen. Although no secretion was 


Fig. 7 Frontal section of the frontal gland of a worker. B.m., basement 
membrane; ms., mesenchyme; J, epithelial cells. Homog. immers. 1.8mm., oc. 6, 
reduced one-fifth. 


actually observed in the lumen of the gland, the intracellular 
canals and the pores in the glandular cuticula seem to indicate 
that the frontal gland in the soldier is in active secretion. 

5. The worker. The frontal gland of the worker as seen in 
surface view (fig. 11) is very small and is surrounded by an 
empty space. Sections (figs. 7, 25) show that this gland, al- 
though bearing a morphological resemblance to that of the 


BRAIN OF THE ‘WHITE ANT’ 583 


nymphs of the sexual forms, is in a less highly developed con- 
dition, a condition that seems to be secondarily, indeed regres- 
sively, modified, rather than primitive. In the area between 
the mushroom bodies we find (fig. 7) a cup formed by base- 
ment membrane, b.m., which, although smaller, is somewhat 
similar in outline to that of the two nymphs just described. 
The space within the basement membrane is not, however, 
entirely filled by the cells of the frontal gland. The cells, /, 
occupy only the central part of the cup, the remainder consists 
of a membranous network of mesenchym, ms., similar to and 
indeed continuous with the tentorial membrane which lies pos- 
terior to the frontal gland and to the supraesophageal ganglion 
in all castes. The cells of the frontal gland as shown in figure 
7, 1, are cut in a plane parallel to their distal upper surfaces, the 
cell bodies polygonal in outline and closely pressed together, 
and only a few nuclei are present. In the following sections 
from deeper portions of the gland, many more nuclei occur. 
The cells are all of the slender elongated type, but much smaller 
in every dimension than the slender elongated cells of the two 
nymphs above described. None of the swollen secretory cells 
are present. From the general appearance it may be inferred 
that this gland in the worker is not only nonfunctional but 
degenerate in structure. 

b. General discussion of the frontal gland. To summarize the 
different structural and functional conditions of the epithelial 
cells of the frontal gland in the various forms of L. flavipes, we 
find that in the soldier the frontal gland is evidently functional 
and in active secretion. An external opening is present and the 
epithelial cells are all glandular, united in a syncytium contain- 
ing a network of intracellular canals and covered by a porous 
cuticula. 

In the true adult the frontal gland is probably functional. 
An external opening is present above the gland; the gland has 
enlarged since the last nymphal phase, and a cuticula resembling 
that of the soldier is present upon the inner surface of the epi- 
thelial cells. No statement as to the cells themselves can be 
made now, since no sections of this gland have been studied. 


584 CAROLINE B. THOMPSON 


In the two nymphs with long and short wing pads the frontal 
gland seems not yet functional. There is no external opening 
above the gland; the cells which appear capable of producing 
secretion are few and scattered, while most of the cells are nar- 
row and slender with no evidence of any contained secretion, 
but resembling rather sensory cells in their form and contents. 
The last nymphal molt will remove the outer cuticula, and the 
absence of the inner cuticula, described above, will produce the 
external opening seen in the true adult. At the same time the 
epithelial cells which were not glandular in the nymphal phases 
may form a cuticula on their inner surface, acquire the power 
of secretion, and become fully glandular in function in the adult 
condition. It would therefore seem that the frontal gland of 
nymphs with long and short wing pads represents an earlier 
phase of development than that of the true adult or soldier. On 
the other hand, it is apparent that the frontal gland of the 
worker is in a secondarily modified and regressive condition. 
It is evidently nonfunctional], the cells showing no signs of secre- 
tion; it is also degenerate in structure, for the cells, although of 
the same form as in the nymphs, are much smaller, are numeri- 
cally fewer, and are partly replaced by a network of mesenchym. 

The study of the frontal gland might end here with the fre- 
quently asked and still unanswered question: What is the function 
of the frontal gland? The question, however, that has presented 
itself most forcibly throughout my study of this organ is, Why 
should most of the epithelial cells of the frontal. gland in the 
nymphal phases bear such a striking resemblance to sensory 
cells? Young gland cells may usually be recognized as such 
even in their early phases, and they are rarely so slender and 
elongated. If, as the biogenetic law teaches us, ancestral struc- 
ture is frequently to be observed in young, that is, in develop- 
ing organs, is this then an instance? It is possible that the 
frontal gland of the termite, whose function is unknown and 
whose structure needs further investigation, may represent an 
ancestral organ, highly developed and secondarily modified in 
some individuals and vestigeal in others, but whose primitive 


BRAIN OF THE ‘WHITE ANT’ 585 


structure and perhaps even whose origin may be revealed in the 
developing nymphal stages. 

This supposition leads naturally to the query, What then was 
the frontal gland if it were originally not a gland? In answer I 
wish to offer the suggestion that the frontal gland of the termites 
may be a modification of the ancestral frontal ocellus or simple 
eye that is found in many insects, and of which Berlese (’09) 
says that if the third eye is not present it should be considered 
as having disappeared. ‘‘Ordinamente sono in numero di tre 
e giacciono sul vertice o sulla fronte—. . . . In taluni casi 
se ne osservano solo due (Grillidae ecc.) ma il terzo devesi con- 
siderare come abortito.” 

Several naturalists of the nineteenth century, notably Joly 
(49) and Lespés (’56), in their study of termites observed the 
pale depressed spot now known as the fontanel and described 
it as the third, median, ocellus. 

Hagen (’55-’60) in his description of Termes flavipes, Kollar, 
states that there is ‘‘Mitten auf dem Scheitel ein flachen Ein- 
druck mit einem erhabenen helleren Fontanellepunkt.”’ Refer- 
ring to the earlier view that this ‘fontanel point’ was a third 
ocellus, he writes: 


Ein drittes Nebenauge fehlt bestimmt; dass das so oft dafiir ange- 
gebene Organ (dem iibrigens stets eine gew’lbte Hornhaut fehlt) kein 
Nebenauge sein kann, hatte man schon aus seiner Lage abnehmen 
~kénnen. Es liegt nimlich immer betrichtlich h5her auf dem Scheitel 
als die Nebenaugen, ein Verhiiltniss, dass bei den Insekten ohne Bei- 
spiel ist; das dritte Nebenauge liegt bei allen bekannten Insekten niiher 
dem Munde, als die seitlichen. 


Nassonoff (’93) described the structure of the frontal gland in 
a termite soldier. 

Czerwinsky (’97) studied a number of termites, including some 
tropical forms. He described the structure of the frontal gland 
as follows: 


Die Stirndriise liegt hinter dem Oberschlundganglion ein wenig tiber 
demselben and gehirt immer zu den mehrzelligen Driisen, in der 
Bildung aber stellt sie mannigfiltige Grade der Entwicklung dar. In 
dem einfachen Fille besteht sie aus einer Schieht in die Linge verzo- 
gener Hypodermiszellen. Solchen einfachen Bau findet man bei vielen 


DS6 CAROLINE B. THOMPSON 


Arbeitern (Eutermes sp., Eutermes Ripperti, Termes Miilleri). Finen 
komplizierteren Bau finden wie bei einer anderen Gruppe der Arbeiter 
(Termes lucifugus, Eutermes capricornis, Termes dirus) und bei den 
gefliigelten Insekten. Die Driisenzellen sind hier stark in der Linge 
ausgezogen und bilden zusammen ein Sackchen. Das Sekret der 
Driisenzellen hiuft sich zwischen denselben und dem Chitin und gelangt 
wahrscheinlich erst durch die Kérperoberfliche nach aussen. Das Chitin 
ist hier sehr dinn und bildet ausserhalb einen weissen Fleck (Fon- 
tanelpunkt der Beschreibungen). Dieser Fontanelpunkt wurde sogar 
von einigen Verfassern als ein drittes Nebenauge angenommen. 

Hagen leugnet die Existenz des dritten Nebenauges, aber erklart 
nicht die Bedeutung des Fontanellpunktes. Der Fontanellpunkt mit 
der anliegenden Oberfliche hildet eine wenig deutliche Vertiefung. 
Durch tiefere Einsenkung der Driisenzellen ins Innere mit der dariber- 
liegenden Cuticula kann sich ein Reservoir bilden aus dem das Secret 
durch eine Oeffnung heraustritt. Eine soleche Bildung stellt die Stirn- 
driise bei den Soldaten dar. Den kompliziertesten Bau der Stirndriise 
finden wir bei den soldaten Nasuti und Arbeitern Nasuti, bei denen 
ausser dem schon beschriebenen Bau ein Ausfiihrsgang in der Nase 
vorkommt. 


Holmgren, ’09, finds the fontanel gland similar in structure 
in the imago and worker of Eutermes chaquimayensis. The 
glands consist of a single layer of elongated cells. ‘Two muscles, 
the mm. retractores fontanellae, are attached near the tip of the 
gland. No actual secretion was observed. In the soldier, the 
frontal gland consists of two clearly defined parts, (1) the glandu- 
lar part, of high cylindrical cells, with a row of short rods (Stab- 
chen) on the distal ends, (2) the duct of low and culticulated 
cells. 

Feytaud (712), in a study of ‘Le termite lucifuge,’ writes: 

La glande fontanellaire ou glande frontale est formée par la differ- 
entiation d’une groupe de cellules hypodermiques au niveau de la partie 
supérieure du front, au-dessous de la zone plus claire désignée sous le 
nom de fontanelle. Dans beaucoup de eas, la glande frontale n’est 
représentée que par une différentiation des cellules hypodermiques, 


sans dépression. Enfin, dans certaines formes, elle est complétement 
absente. 


Feytaud cites the nymph of the second form as an instance of 
a slightly differentiated frontal gland, stating that like that of 
the worker the frontal gland in this nymph is merely a zone of 


elongated hypodermal cells. In the conclusion of this paper 
Feytaud asks ‘‘What is the réle of the frontal gland?” 


| 


BRAIN OF THE ‘WHITE ANT’ 587 


Wasmann (’02), in describing the new genus Speculitermes 
states that it is ‘‘Ausgezeichnet durch die grosse, unpaare Stirno- 
celle, welche nicht nur bei der Imago, sondern auch beim Arbeiter 
vorhanden ist.” 

Now, although I am firmly convinced that the fontanel spot 
with its opening is not an ocellus or simple eye, I feel that there 
is considerable evidence that the frontal gland of the termites 
may have developed phylogenetically from an ancestral median 
eye and its ocellar nerve. In other words, the frontal gland 
may have first arisen as a modification of the hypodermis at the 
point where the ancestral median eye was formerly situated, and 
then may have extended inward into the head along the course 
of the former median ocellar nerve, the proximal part of which 
may still persist in the nerve which I have termed provisionally 
the fontanel nerve. : 

The evidence in support of this view, gained from the study 
of the different forms of L. flavipes, will now be presented. 

1. The anterior surface of the frontal gland of the nymphs 
and adults of the sexual forms lies in the same frontal plane as 
the lateral ocelli. The posterior end of the frontal gland lies in 
the region in which the lateral ocellar nerves enter the proto- 
cerebral fibrous core. The frontal gland has therefore the same 
linear extent as the lateral ocelli and the ocellar nerves. 

2. The nerves from the lateral ocelli run from the ocelli in 
toward the median line on the outside of the brain sheath, directly 
beneath the hypodermis, then piercing the brain sheath, they 
run downward and backward within the nerve cell layer of the 
brain, finally entering the fibrous core of the protocerebral lobes 
at a definite point, namely: immediately behind the posterior 
dorsal commissure. The fontanel nerve from the frontal gland 
enters the protocerebral fibrous core in the median line, imme- 
diately behind the posterior dorsal commissure, and, according 
to the present theory, may represent the median ocellar nerve 
(fig. 19). 

3. The lateral ocelli are derived from the hypodermis, and the 
tips of the visual cells still lie in contact with and perhaps be- 
tween the hypodermal cells. The frontal gland is directly con- 


588 CAROLINE B. THOMPSON 


tinuous with the hypodermis, and its cells are modified hypo- 
dermal cells. 

4. The inner cuticula is thinner at the margins and lacking 
above the center of the lateral ocelli (fig. 13). The same is true 
of the inner cuticula above the frontal gland (fig. 16). 

5. The cells of the frontal gland in the adult stages are secre- 
tory in function and resemble other gland cells in structure, but 
in the developing nymphal stages the cells produce no secretion 
and bear a striking resemblance to the slender elongated visual 
cells found in insect ocelli, Hesse (01 b). According to our 
theory this ontogenetic phase corresponds to a primitive, phylo- 
genetic, condition. In this connection it is significant that in 
very young nymphs the frontal gland is barely distinguishable 
in surface views, while the brain and eyes are quite well devel- 
oped. The further study of these youngest nymphs by means 
of sections may give additional evidence. 

6. The phylogeny of the termites upon the basis of the mor- 
phology of the frontal gland suggested by Holmgren (’09) may 
be referred to as additional evidence. Holmgren has divided 
all adult sexual termites into three groups: The first group, 
including all the higher termites, consists of those forms with a 
sack-like frontal gland (sackf6rmige Druse) and a fontanel open- 
ing; the second group, including part of the lower termites, con- 
sists of those forms with a plate-like nonglandular ‘gland’ (Fon- 
tanellplatte) and no fontanel opening; the third group, includ- 
ing the remainder of the lower termites, consists of forms in 
which both the plate-like nonglandular ‘gland’ (Fontanellplatte) 
and the fontanel opening are lacking. Holmgren believes that 
the sack-like fontanel gland and the nonglandular plate are 
morphologically equivalent, and that both are derived from a 
common ancestor without either gland or plate. He further 
remarks that no fontanel is found in the Blattidae, which in 
this respect are more nearly related to the lower termites. To 
quote his exact words: 


Versuchen wir nun eine Phylogenie der Termiten auf dem Verhalten 
der Fontanelle zu griinden, so mjchte erstens bermerkt werden, dass 
diejenigen Formen, welche Fontanelldriisen (Fontanellplatte und sack- 


BRAIN OF THE ‘WHITE ANT’ 589 


formige Driise) besitzen, aus. Formen ohne driisige Fontanellplatte 
stammen miissen. Da es wenig wahrscheinlich ist, dass eine driisige 
Fontanellgewebe unabhinig bei verschieden Termitengruppen ent- 
standen ist, so miissen wir die jenigen mit Fontanelplatte und die 
jenigen mit schlauchformiger Fontanelldriise von einer gemeinsamen 
Ausgangspunkt ableiten. 


7. It is always questionable whether comparisons should be 
drawn between invertebrates and vertebrates, but it does not 
seem wholly improbable that the frontal gland of the termites 
may have had a phylogenetic history similar to that of the 
vertebrate pineal gland. The disappearance of the simple me- 
- dian vertebrate eye may be understood in conjunction with the 
increasing complexity of the compound eyes, and a brief survey 
of termite habits may throw light upon the causes of the dis- 
appearance of this median invertebrate eye. The true sexual 
adults are the most conservative and primitive in structure and 
in habits of all the castes. They are deeply pigmented, possess 
long functional wings, and for a short time lead an aerial exist- 
ence like other typical winged insects, and unlike all the other 
members of the community. After the short period in the air, 
which termite observers tell us is not a nuptial flight, the true 
adults abandon their normal ancestral mode of life and begin a 
secondarily acquired existence within the total darkness of their 
burrows. Here the long wings are shed, mating takes place, and 
from henceforth they live within the dark and narrow chambers 
of the nest. It is possible that this sudden and marked change 
of environment may account in part for the disappearance of the 
median eye in the ancestral termite and that the frontal gland 
may have arisen in response to some new need. 


SUMMARY OF RESULTS 


The purpose of this study is to compare the brains of the 
different castes of L. flavipes, and, further, to compare them with 
those of the castes of ants, since both termites and ants are 
social insects with a highly organized community life, but with 
a very different degree of specialization and intelligence. 

The forms whose brains are here discussed are: the nymph of 
the first form, with long wing pads, the nymph of the second 


590 CAROLINE B. THOMPSON 


form, with short wing pads, the soldier, the worker, and, to 
some extent, the true adult. 

1. There is no differentiation between the brains of the males 
and females of any caste or stage of L. flavipes. 

2. There is very little differentiation between the brains of 
the different castes and stages of L. flavipes here discussed. 

a. The most marked differentiation is in the optic apparatus, 
a correlation existing between the degree of development of the 
compound eyes and the size of the optic lobes. The latter are 
large in the nymphs and adults of the sexual forms, but are 
greatly reduced in the worker and soldier. 

b. The mushroom bodies differ very little in size, by actual 
measurement and in the estimated number of cells. They are 
largest in the worker, smallest in the soldier, and are inter- 
mediate in size in the sexual forms, although the mushroom 
bodies of the true adult are nearly as large as those of the worker. 

c. The antennary lobes are very similar in size, but are largest 
in the worker and in the sexual forms and slightly smaller in the 
soldier. 

3. The mushroom body stalks do not end beneath the central 
body, as was formerly thought, but divide into three roots, the 
anterior, central body, and posterior roots. The Jatter are ex- 
panded into large and prominent lobes. 

4 The labral nerves arise from the ventral surface of the 
frontal ganglion, and not directly from the labrofrontal nerves. 

5. The termite brain as a whole is very similar in structure to 
the brain of ants, with the notable exception of the mushroom 
bodies which are of a much more simple and primitive type. 
This is apparent in: (a) the nerve cells, which are all very small 
and of equal size, (b) the separation of these cells into three 
groups, instead of the four of ants, (c) the incomplete separation 
of the two lobes. 

Only the two lateral simple eyes or ocelli are present in the 
nymphs and adults of the sexual forms, and none are found in 
the worker or soldier. These ocelli are very simple and primi- 
tive in structure, without lens or pigment, and the ocellar nerves 
do not expand into ocellar lobes as in ants. 


BRAIN OF THE ‘WHITE ANT’ BS 


6. The fronta] gland is a gland found in all the castes of L. 
flavipes, situated on the postero-dorsal surface of the brain in 
the space between the mushroom bodies. The gland is com- 
posed of epithelial cells which are continuous with the hypo- 
dermal cells. A nerve, termed provisionally the fontane] nerve, 
runs from the frontal gland into the brain. 

In the true adult and soldier the frontal gland is evidently 
functional; in the worker it is nonfunctional and degenerate in 
structure; in the nymphs with long and short wing pads it 
is evidently not yet functional, a few cells are glandular in 
appearance with contained secretion, the Jarge majority of the 
cells are slender and elongated and resemble sensory, visual, cells. 

7. The suggestion is made that the frontal gland may have 
arisen phylogenetically from the ancestral median ocellus which 
is now lacking in the termites, and that the ‘fontanel’ nerve 
may be a vestige of the former median ocellar nerve. 

The arguments for this view are based upon the position and 
the structural resemblances of the frontal gland and the lateral 
ocelli, upon the presence of the ‘fontanel’ nerve in the same 
frontal section in which the lateral ocellar nerves enter the 
brain, upon the resemblance of the frontal gland cells of the 
developing nymphal phases to visual cells, and upon the facts 
collected by Holmgren regarding a phylogeny of the termites 
based upon the morphology of the frontal gland. 


WELLESLEY, MASS. 
avuaust 18, 1916 


THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL. 26, No. 5 


592 CAROLINE B. THOMPSON 


BIBLIOGRAPHY® 


Banks, W. 1907 A new species of Termes. Ent. News, vol. 18, no. 9. 

Breriess, A. 1909 Gli Insetti, vol. 1. 

Brugs, C. T. anp MeLanper, A. L. 1915 Key to the families of North Ameri- 
can insects. 

Buanton, E. 1913 La differentiation des castes chez les termites (Nerv.) Bull. 
Soc. Ent. de France, No. 8. 

Czerwinsky, K. 1897 Beitriige zur Kenntniss der Termiten. Zool. Anz., Bd. 
20. 

Frytaup, J. 1912 Contribution a l’étude du termite lucifuge. Archives 
d’anat. microse., T. 13. 

Grasst, B., AND Sanpras, A. 1893 The constitution and development of the 
society of termites; observations on their habits; with appendices on 
the parasitic protozoa of Termitidae and on the Embiidae, translated 
by T. H. Blandford. Quart. Jour. Micros. Sci., vol. 39, n. s., April, 
1897. ' 

Hagen. H. A. 1855-60 Monographie der Termiten. Linn. Ent., Bd. 10, 12, 14. 

Heatu, H. 1903 The habits of California termites. Biol. Bull., vol. 4. 

Hnussr, R, 1901 a Ueber die sogen. einfachen Augen der Insekten. Zool. Anz., 
Bd. 24. 

1901 b Untersuchungen tiber die Organe der Lichtempfindung bei 
niederen Thieren. 7. von den Arthropoden-Augen. Zeit. f. wiss. 
Zool., Bd. 70. 

HormGren, N. W. 1909 Termiten Studien. 1. Anatomische Untersuchungen. 
K. Svenska Vetensk. Akad. Hand., Bd. 44. 

1911 Termiten Studien 2. Systematik der Termiten. K. Sv. V. Ak. 
Hd., Bd. 46. 

1912 Termiten Studien 3. Systematik der Termiten. Die Familie 
Metatermitidae. K. Sv. V. Ak. Hd., Bd. 48. : 

19138 Termiten Studien 4. K. Sy. V. Ak. Hd., Bd. 50. 

Joty 1849 Recherches pour servir 4 l’histoire naturelle et 4 l’anatomie des 
Termites. Mém. Acad. Sc. Toulouse, t. V. 2 

Jonrescu, C. 1909 Vergleichende Untersuchungen iiber das Gehirn der Honig- 
biene. Jenaische Zeit., Bd. 45. 

KouuarR 1837 Naturgeschichte der schidlichen Insekten. 

Lespks, C. 1856 Recherches sur l’organisation et les moeurs du Termite 
lucifuge. Ann. Sci. Nat. Zool., 4 série, t. 5. 

Maruatt, C. L. 1908 The white ant (Termes flavipes, Koll.) U. 8. Dept. Agr. 
Bur. Ent. Cire. no. 50. 

Nassonorr, 1893. Ueber Higenthiimliche auf den Nestenbau beziigl. Organi- 
sationsverhiltnisse bei den Termiten. Entom. Untersuchungen, War- 
saw. 

PintscHKER, H. 1911 Das Gehirn der Ameise. Jenaische Zeit., vol. 47. 

Sttvestri, F. 1901 Boll. Mus. Torino, vol. 16, no. 389. 

1903 Redia, vol. 1, p. 37. ; 


5 More extensive bibliographies of the termites may be found in the works of 
Holmgren (’11), Feytaud (’12), and Snyder (’15). 


BRAIN OF THE ‘WHITE ANT’ 593 


SrrickLanp, E. H. 1911 A quiescent stage in the development of Termes 
flavipes, Kollar. Journal of N. Y. Ent. Soc. 

Snyper, T. E. 1915 Biology of the termites of the Eastern United States, 
with preventive and remedial measures. U.S. Dept. Agr., Bur. Ent., 
Bull. no. 94, pt. II. 
1916 Termites, or ‘White Ants,’ in the United States: their damage, 
and methods of prevention. U.S. Dept. Agr. Bur. Ent., Bull. no. 333. 

Tuompson, C. B. 1913 A comparative study of the brains of three genera of 
ants, with special reference to the mushroom bodies. Jour. Comp. 
Neur., vol. 23. 
1914 The mushroom bodies of the worker of Bombus sp. Jour. Comp. 
Neur., vol. 24. : 

Wasmann, E. 1902 Termiten, Termitophilen und Myrmekophilen. Zool. 
Jahrbuch., Abt. f. Syst., vol. 17. 
1908 Zur Kastenbildung und Systematik der Termiten. Biol. Cen- 
tralblatt., vol. 4§. 

Wueeter, W. M. 1904 The phylogeny of termites. Biol. Bull., vol. 8. 


PLATE 1 
EXPLANATION OF FIGURES 


8 and 9 are drawn from whole mounts of the heads. Figure 10 is a combina- 
tion drawing of the outline of the head and the brain, the latter added from a 


dissected and mounted brain. 


The stippling represents the nerve cell layer, 


the fibrous matter is blank. Obj. 16, oc. 6, stage level, reduced one-third. 
8 The head and brain of a nymph with long wing pads. 


9 The head and brain of a true adult. 


10 The head and brain of a nymph with short wing pads. 


EXPLANATION OF PLATES »% 


All figures are drawn with the Spencer camera lucida and with Spencer lenses 


ABBREVIATIONS 


a.cm.. anterior dorsal commissure 

a.l., antennary lobe 

a.r.m.b., anterior root of mushroom 
body 

b.m., basement membrane 

c.b., central body 

c.b.r., central body root 

c.€., compound eye 

cu., cuticula 

f.g., frontal gland 

f.n., fontanel nerve 

fr.gn., frontal ganglion 

gl., glia cells 

hyp., hypodermis 

i.cu., inner cuticula 

7.f., inner fiber mass 

la.n., labral nerve 

lb.n., labial nerve 

L.f.n., labrofrontal nerve 

m.b., mushroom body 


m.b.s., mushroom body stalk 

md.n., mandibular nerve 

m.f., middle fiber mass 

ms., mesenchym 

mz.n., maxillary nerve 

oc., ocellus 

oc.n., ocellar nerve 

0.cu.,.outer cuticula 

oe., oesophagus 

o.f., outer fiber mass 

0.l., optic lobe 

p.l., protocerebral lobe 

p.r.m.b., posterior root of mushroom 
body , 

r.m., recurrent nerve 

sb.g., subesophageal ganglion 

tr.cm., tritocerebral commissure 

tr.l., tritocerebral lobe 

v.c., ventral connective 


594 


PLATE 1 


BRAIN OF THE ‘WHITE ANT’ 
CAROLINE B. THOMPSON 


PLATE 2 


EXPLANATION OF FIGURES 


11 and 12 are drawn from whole mounts of the head. The stippling represents 
the nerve cell layer, the fibrous matter is blank. Obj. 16, oc. 6, stage level, 
reduced one third., 

11 The head and brain of the worker. 

12 The head and brain of the soldier. 


596 


BRAIN OF THE ‘WHITE ANT’ PLATE 2 
CAROLINE B. THOMPSON 


PLATE 3 


EXPLANATION OF FIGURES 


13 to 20 are taken from a series of frontal sections of the head of a male nymph 
with long wing pads, beginning at the frontal surface. The fibrous core is 
mottled, the nerve cell layer is blank. Obj. 16, oc. 6, table level. 

13. Section through the anterior part of the protocerebral lobes, p.l., showing 
the lateral ocelli, 6c., and the anterior margin of the frontal gland, f.g. 

14 Section showing the connection of the anterior roots, a.r.m.b., and the 
stalks, m.b.s., of the mushroom bodies. 

15 Section through the central body, c.b., and the central part of the mush- 
room bcdies, m.b. 

16 Section showing the entrance of the central body roots, c.b.r., into the 
central body. 


598 


BRAIN OF THE ‘WHITE ANT’ PLATE 3 


CAROLINE B. THOMPSON 


599 


PLATE 4 
EXPLANATION OF FIGURES 


17. Section through the posterior part of the central body, showing the origin 
of the posterior roots of the mushroom bodies, p.r.m.b. 

18 Section showing the posterior dorsal commissure, p.cm. The exit of the 
labrofrontal nerve, l.f.n., from the nerve cell layer is also shown. 

19 Section showing the entrance of the ocellar nerves, oc.n., and of the fon- 
tanel nerve, f.n., into the protocerebral fibrous core. The large lobes of the 
posterior roots of the mushroom bodies, p.r.m.b., are very prominent, and are 
also connected with the protocerebral fibrous core. 

20 Section showing the connection between the protocerebrum, deutocere- 
brum, and tritocerebrum. 


609 


BRAIN OF THE ‘WHITE ANT’ PLATE 4 
CAROLINE B. THOMPSON 


-- €.6. 


PLATE 5 


EXPLANATION OF FIGURES 


21 to 26, obj. 16, oc. 6, table level. 


21 
22 
23 
24 
25 
26 


Frontal section through the anterior part of the brain of a soldier. 
Frontal section through the central part of the brain of a soldier. 
Frontal section through the posterior part of the brain of a soldier. 
Frontal section through the anterior part of the brain of a worker. 
Frontal section through the central part of the brain of a worker. 
Frontal section through the posterior part of the brain of a worker. 


602 


PLATE 5 


BRAIN OF THE ‘WHITE ANT’ 
CAROLINE B. THOMPSON 


ae eta tt he! 


ips “hor Wik 


+ c 
¥ ‘ 5 
P ‘ , gy 
* Pa ce 4 ' me 
A ¥ 1 
\ i ‘ 
3 
' 
} . 
« i 
‘ 7 1 
2 
~ - 
hy iY 
¥ ' 
» - “4 ‘ « 
- Lue ie 
* : 7 ‘ 
. * 
. -} ‘ 
¢ ~, 78 
: nH 
’ r ’ 
‘ 
3 ‘SS ° 
th 
A - : 
x - 
1 
. 
P es 
- ae , 
i —— 


yay \ fe ie 
Ps 


ij : 
. 
La 
r 
WF So 
> « 
, \ 
. . ts 
i =< 
‘ AP ak a en ; oer 
ve ty Ae Hy i 
1 - | 5 14 a 
/ Oe may 
' 
J * . « 
® ' 
i . as! 
¢ a 
. ; 


SUBJECT AND AUTHOR INDEX 


7 eccnee RMREIAEEE ON Caden 8553 cau A tbe accom AOS 

Action of light. Changes in the rod-visual 
cells of the frog due to the................ 429 

Activity on the histological structure of nerve 
(OVERS Dd BUTS | AC) A 341 


Afferent system of the head ‘of Amblystoma. 
Correlated anatomical and physiological 
studies of the growth of the nervous sys- 
fern oD Ampnipins Ul. Phe... ............ 247 

(Albino rat). A preliminary determination 
of the part played by myelin in reducing 
the water content of the mammalian nerv- 
UMMM Gs Vette esis ite seem ae ese 443 

ALLEN, Witt1aAM F. Studies on the spinal 
cord and medulla of cyclostomes with 
special reference to the formation and ex- 
pansion of the roof plate and the flatten- 


WUE Of LHS SPINK OOKG steve. ves cssces. ese 9 
Amblystoma. I. The forebrain. Regenera- 

Pere re eyieys SOM in SON GR eee 203 
Amphibia. II. The afferent system of the 


read of Amblystoma. Correlated ana- 
tomical and physiological studies of the 
growth of the nervous system of.......... 247 
Ant, Leucotermes flavipes, Kollar. The 
brain and the frontal gland of the castes 
Ply Geer NEES 2 exces tates bans bbw A os vce « 553 
Arey, Lesuie B. Changes in the rod-visual 
cells of the frog due to the action of light.. 429 
Arey, Lestiz B. The function of the effer- 
ent fibers of the optic nerve of fishes. . 213 
Arey, Lestre B. The influence of light and 
temperature upon the migration of the 
retinal pigment of Planorbis trivolvis..... 359 
Arey, Lestre B. The movements in the 
visual cells and retinal pigment of the 
JOWOr VEriODTAtAas. << os .c<dcbuevccuwdasacs 121 


AILEY, Perctvat. Morphology of the 
roof plate of the forebrain and the lateral 
choroid plexuses in the human embryo. 79 

Battey, Perecrvar. The morphology and 
morphogenesis of the choroid plexuses 
with especial reference to the develop- 
ment of the lateral telencephalic plexus 

in Chrysemys marginata................. 507 

Brain. A study of a Plains Indian........... 403 

Brain and the frontal gland of the castes of 
the ‘white ant,’ Leucotermes flavipes, 


Vt EN SOB Ws Cres senda) Sy 553 
Brain of Amblystoma. I. The forebrain. 
BPE ers tori Mi CHOss er acoessccccescceess 203 


Brain of reptiles. Evidence of a motorpal- 
Wiis the Lore=c cee eek a Wee is eee ee ees 
Burr, H.Saxron. Regeneration in the brain 
of Amblystoma. I. The forebrain........ 203 


ASTES of the ‘white ant,’ Leucotermes 
’ flavipes, Kollar. The brain and the 


grontalizinng OF Chasse... eee cs enw ae 553 
Cells and retinal pigment of the lower verte- 
brates. The movements in the visual.... 121 


Cells of the frog due to the action of light. 
Changes in the rod-visu 

Cells. The effect of activity on the histolozi- 
Galistructure Of NELVE. 4.6.5.0. weveses 341 


Central nervous system of the woodchuck 
(Marmota_monax) during hibernation. 
Absence of chromatolytic change in the.. 391 

Change in the central nervous system of the 
woodchuck (Marmota monax) during hi- 
bernation. Absence of chromatolytic.... 391 

Case, Martin R. An oe study 
of the vagus nerve. 421 

‘Chemical sense’ in vertebrates. Regarding 
the existence of the ‘common’............ 1 

Choroid plexusesinthe humanembryo. Mor- 
phology of the roof plate of the forebrain 
andthe lateral... ... saunter os «es ec wee 79 

Choroid plexuses with especial reference to 
the a velopment of the lateral telen- 
cephalic plexus in Chrysemys marginata. 
fo morphology and morphogenesis of 


Chromataiytic change in the central nervous 
system of the woodchuck (Marmota mo- 
nax) during hibernation. Absence of..... 391 

Chrysemys marginata. The morphology and 
morphogenesis of the choroid plexuses 
with especial reference to the development 
of the lateral telencephalic plexus in...... 507 

Common chemical sense’ in vertebrates. 
Regarding the existence of the............ 

Cocuiut, G. E. Correlated anatomical and 
physiological studies of the growth of the 
nervous system of Amphibia. II. The 
afferent system of the head of Amblvs- 
co) See eee. eee 247 

Content of the mammalian nervous system 
(albino rat). preliminary determi- 
nation of the part played by myelin in 
reducing the water: <<. ccebemess «6 0< +. sea 443 

Cord and medulla of ecvclostomes with spe- 
cial reference to the formation and expan- 
sion of the roof plate and the flattening of 
the spinal cord. Studies on the spinal... 9 

Cranial nerves. The structure of the third, 
fourth, fifth, sixth, ninth, eleventh and 
CWPRLCH ; Sane s0 Sian Soe tess «dc ctee en 541 


Crozier, W. J. Regarding the existence of 
the ‘common chemical sense’ in verte- 
DYRCAG sinc vn on na eas kane cuatcs 1 

Crozier, W. J. The taste of acids........... 453 


Cyclostomes with special reference to the for- 
mation and expansion of the roof plate 
and the flattening of the spinal cord. 
ee on the spinal cord and medulla 5 
lic: ccs Sele ee Pains pare ine « iss wager 


EVELOPMENT of the dorsal ventricular 
TIOHO 1 Hurhlede PEDO. oo. c0 ganna ee 481 


Development of the lateral telencephalic 
plexus in Chrysemys marginata. The 
morphology and morphogenesis of the 
— plexuses with especial reference to 


Donatoson, H. H. A preliminary determi- 
nation of the part plaved by myelin in 
reducing the water content of the mam- 


malian nervous system (al. qnO Tat) vn.2es 443 
Dorsal ventricular ridge in turtles. Ihe de- 
WelopMent OF GHG... 2c6e.0e +s bee acme oun ee 481 


605 


606 


FFECT of activity on the histological 
EK structure of nerve cells. The........... 341 
Efferent fibers of the optic nerve of fishes. 


The function of the: ... 007 cu.c ceases 213 
Embryo. Morvhology of the roof plate of 
the forebrain and the lateral choroid plex- 
TSCA IN HOG NUMAN <6; oveenaslvenie ara g aaa 79 
IBERS of the optic nerve of fishes. The 
function of the efferent. ...............- 213 
Fishes. The function of the efferent fibers 
Ol bNeiantic Nerve) Of. «ic sccemcceee dens 213 


Flattening of the spinal cord. Studies on the 
spinal cord and medulla of cyclostomes 
with special reference to the formation 
and expansion of the roof plateandthe.. 9 

Flavives, Kollar. The brain and the frontal 
gland of the castes of the ‘white ant,’ Leu- 
COLORINIOS cen ge icine 1 Uhce cera uO eee 553 

Forebrain and the Jateral choroid plexuses in 
the human embryo. Morphology of the 


TOOMplAteOleChO>;.0. icc Taree 79 
Forebrain of reptiles. Evidence of a motor 
pallitimbinithes.. eke cease ce sees 475 
Forebrain. Regeneration in the brain of 
Amblystoma-. cl. Phexeeresseeceee ne. 203 
Frog due to the action of light. Changes in 
the rod-visual cells of the............. ... 429 


Frontal gland of the castes of the ‘white ant,’ 
Leucotermes flavipes, Kollar. The brain 
Piste Mulole co eee oc deo coder raodacoro kc 553 


LAND lf the castes of the ‘white ant,’ 
Leucotermes flavipes, Kollar. The 
braimvand theiirontaltes seeceeeeeceees 553 


EAD of Amblystoma. Correlated ana- 
tomical and physiological studies of 
the nervous system of Amphibia. II. 
The afferent system of the............... 247 
Hibernation. Absence of chromatolytic 
change in the central nervous system of 
the woodchuck (Marmota monax) during 291 
Hiton, Wruutam A. The nervous system of 


DY. CHOZONIGS 44... sae eee eee 453 

| goes brain. A study of a Plains....... 403 
OHNSTON, J. B. Evidence of a motor 

pallium in the forebrain of reptiles...... 475 
Jounston, J. B. The development of the 

dorsal ventricular ridge in turtles......... 481 
EEGAN, J.J. Astudy of a Plains Indian 


DVRINM Petes oc tinea ts ean eee 403 


Kocu,Somner L. Thestructure of the third, 
fourth, fifth, sixth, ninth, eleventh and 


twelfthveranial) nerves): 2/2 .4..)5..0.05.00. 541 
Kocuer, R. A. The effect of activity on the 
histological structure of nerve cells....... 341 


| Bary and temperature upon the migra- 
tion of the retinal pigment of Planorbis 


trivolis. The influence of............... 359 
Light. Changes in the rod-visual cells of the 
frog due to the action of........... Meroe 429 


Leucotermes flavipes, Kollar. The brain and 
the frontal gland of the castes of the 
Swi tevanty weer arch tlie a rns 553 


AMMALIAN nervous system (albino 
rat). A preliminary determination of 
the part played by myelin in reducing 

the water content of the................. 443 

Marginata. The morphology and morphogen- 
esis of the choroid plexuses with especial 
reference to the development of the lateral 

telencephalic plexuses in Chrysemys..... 507 


' Nerve of fishes. 


INDEX 


(Marmota monax) during hibernation. Ab- 
scence of chromatolytic change in the cen- 
tral nervous system of the woodchuck.... 391 
Medulla of eyclostomes with special reference 
to the formation and expansion of the roof 
plate and the-flattening of the spinal cord. 
Studies on the spinal cord and.......... 
Migration of the retinal pigment of Planorbis 
trivolvis. The influence of light and 
temperature upon the.................... 359 
Monax) during hibernation. Absence of 
chromatolytie change in the central neryv- 
ous system of the woodchuck (Marmota.. 391 
Morphogenesis of the choroid plexuses with es- 
pecial reference to the development of the 
lateral telencephalic plexus in Chrysemys 
marginata. The morphology and........ 507 
Morphology and morphogenesis of the choroid 
plexuses with especial reference to the de- 
velopment of the lateral telencephalic 
plexusin Chrysemys marginata. The.... 507 
Motor pallium in the forebrain of reptiles. 


Hividence Of'8 ct... Ronse eee 475 
Movements in the visual cells and retinal pig- 
ment of the lower vertebrates. The...... 121 


Myelin in reducing the water content of the 
mammalian nervous system (albino rat). 
A preliminary determination of the part 
played) DY-.iscewaksx oe ae eee 443 

Myers, J. A., Rasmussen, A. T. and. Ab- 
sence of chromatolytic changein the cen- 
tral nervous system of the woodchuck 
(Marmota monax) during hibernation.... 391 


ERVE. An experimental study of the 
VAP USI erceeiiem issn ©) aa ea a ee 421 


Nerve cells. The effect of activity on the his- 
toloricalistructureloh.. +. 052. see eee 341 
The function of the efferent 
fhibersomtheropticus. cree cee nan 213 
Nerves. The structure of the third, fourth, 
fifth, sixth, ninth, eleventh and twelfth 
Cranial? op ees kee eee eee 541 
Nervous system (alhino rat). A preliminary 
determination of the part played by mye- 
lin in reducing the water content of the 
Mammalian) cece eee eee eee Eee 443 
Nervous system of Amphibia. II. The af- 
ferent system of the head of Amblystoma. 
Correlated anatomical and physiological 
studies of the growth of the.............. 247 
Nervous system of pyenogonids The....... 453 
Nervous system of the woodchuck (Marmota 
monax) during hibernation. Absence of 
chromatolytic change in the central...... 391 


PTIC nerve of fishes. The function of 
the efferent fibers of the................ 213 


ALLIUM in the forebrain of reptiles. 
Bividence lof a motores. ccc cine ee 


Pigment of Planorbistrivolvis. The influence 
of light and temperature upon the migra- 


tion/of theiretinalass.aca-e-6 pare eee 359 
Pigment of the lower vertebrates. The move- 

ments in the visual cells and retinal...... 121 
Plains Indian brain. A study of a........... 403 


Planorhis trivolyis. The influence of light 
and temperature upon the migration of 
theiretinal pigment ole sss ote 359 

Plate and the flattening of the spinal cord. 
Studies on the spinal cord and medulla 
of cyclostomes with special reference to 
the formation and expansion of the roof.. 9 

Plate of the forebrain and the lateral choroid 
plexuses in the human embryo. Mor- 
phology, of the roof 32. cles e<c=2)islolesietenre 79 


Plexus in Chrysemys marginata. The mor- 
phology and morphogenesis of the choroid 
plexuses with especial reference to the de- 
velopment of the lateral telencephalic... . 

Plexuses in the human embiyo. Morphology 
of the roof plate of the forebrain and the 
lateral ehorGidoe-esk see cl sawe es seen tc vee 

Pléxuses with especial reference to the devel- 
opment of the lateral telencephalic plexus 
in Chrysemys marginata. The morphol- 
ogy and morphogenesis of the choroid... . 

Pyenogonids. The nervous system of........ 


ASMUSSEN, A. T. and Myers, J. A. 
Absence of chromatolytic change in the 
central nervous system of the wood- 

chuck (Marmota monax) during hiber- 

LCT ae cao c ter as le deen Ae > son's 

Rat). A preliminary determination of the 
part played by myelin in reducing the 
water content of the mammalian nervous 


BYBLCII GIO yt de ee actor ssc ea dee a's 44 


Regeneration in the brain of Amblystoma. I. 
PNM OVEDURI 20; te oeena sn ota aes acs. eb 
Reptiles. Hvidence of a motor pallium in the 


TOLEDO ny AURA NS eae eo Sie aes eed « 47 


Retinal pigment of Planorbis trivolvis. 
influence of light and temperature upon 


Lhe mictation OF CHO): says cbseee dba scek : 


Retinal pigment of the lower vertebrates. 
The movements in the visual cells and... 
Ridge in turtles. The development of the 
DKSHLVERULICIIAT...vavebaenns esse cs... 5 
Rod-visual cells of the frog due to the action 
of light. (Changes in the...............5. 
Roof plate and the flattening of the spinal 
cord. Studies on the spinal cord and 
medulla of cyclostomes with special ref- 
erence to the formation and expansion of 


ENSE’ in vertebrates. Regarding the 
existence of the ‘common chemical..... 

Spinal cord and medulla of cyclostomes with 

special reference to the formation and ex- 

pansion of the roof plate and the flattening 

of the spinal cord. Studies on the....... 


INDEX 


507 


79 


7 
453 


79 


607 


during hibernation. Absence of chromato- 
lytic change in the central nervous....... 


co OlROGE:  VEHO Meme. =o. o.sdguces 453 


Telencephalic plexus in Chrysemys margi- 
nata. The morphology and morphogene- 
sis of the choroid plexuses with especi«l 
reference to the development of the lateral 507 
Temperature upon the migration of the retinal 
pigment of Planorbis -trivolvis. The in- 
Hance of light aiid) sesms. ss... .0.-08 
THompson, CAROLINE Beritne. The brain 
and the frontal gland of the castes of the 
‘white ant,’ Leucotermes flavipes, Kollar 553 
Trivolvis. The influence of light and tem- 
perature upon the migration of the reti- 
nal pigment of Planorbis................. 359 
Turtles. The development of the dorsal ven- 
tricular ridge in............ yo pes 


AGUS nerve. An experimental study 
OF CRA ila coda: « «tc vc vlc c ale bate 421 


Ventricular ridge in turtles. The develop- 
ment of the dorsal. /5Vi sme. «2... .c008 
Vertebrates. Regarding the existence of the 
‘common chemical sense’ in 
Vertebrates. The movements in the visual 
cells and retinal pigment of the lower.... 
Visual cells and retinal pigment of the lower 
vertebrates. The movements in the..... 


ATER content of the mammalian nerv- 
ous system (albino rat). A prelimi- 
nary determination of the part played 

by myelin in reducing the;.............. 

Woodchuck (Marmota ‘monax) during hiber- 
nation. Absence of chromatolytic change 
in the central nervous system of the...... 391 


~~ 


en = 


NUTT TET TE 


© 
©O 
ep) 
ae! 
Ls) 
LL 


UAT TT 
5 WHS 


ee ee 
Cate ee ee) 
+ +_2_¢ 2 2 0 2 
A 070 4525", 
63-49 e* 
ee ee ee 

9555 °=*. 


8 
ae 


Ais 
resets 
tse 8) 


5 
a @ 


> 
’ 

i 
6 
% 


> 


ee eae 


Ca) Ce? 9 3 
Feis tage 8168 © 
le tes tats" 


° 

4 

eg ta % ft 
sth stat ele: