'':r-'wv',-":^^tjs'*w**-;*

•'-.•vvwwwv* ,.-,«-W(

u^-WW^V^.-V--,..^

■vww^'

^ms'

i^¥'yyy\

yVV^^'VW^y^^

w/ww^vgg^

;Vsiv- -sy'«s

vgj^^UJ-'X

::^,^,*i^-m

^:^«!!«^'P^^^^^^^^;^^

>^>w.W'^^''^

;35y.««»5^ ^^y^

<wuvv;vw-^. >■

'kMi;^y

,>**

wW^WW^ • Vwi

!^v-'^^X.-«)^^^^vy^

•*^*^^^3;

M^J^'^*'^'^^.: ■ ^>WVwVWvv«^*'v.

'^"5!^y^^uw'wVv

vyw^^vv-

;w^:V^^^ww^^^\>^^

i;2f*%is5%jsittv«^^

SS5'¥^«^^^u

'^^•^si^^^

.^^^^

i^ww^^

.iU^'' ^vvWy^yyUyjiiXv^Wy^^^

The Neurones and i^upportins Klrkents
oi-' THE Brain of a ^elabhian

A Disaertation
Submltts'i to the Board of Uniuersi ty Studii
of the Johns Hopkins Uniuersi ty for the
Degree of Doctor of Fhi losophy

Gilbert Logam Houser

Baltimore
1901

cjmi^

The Neurones and Supporting ELEHENTg

OF THE tiRAlN OF A ^ELAi^HIAN

Gilbert L. Houser

Professor of Animal Morphology in the Uniuersitq of Iowa

— o —

With Plates T-VIII.

— o —

TABLE OP COMTSNTS

Section I. INTRODUCTORY ...... 3

Section II. REVIEW OP EARLIER RESEARCHES 6

1. Anatomical Work Involving Errors of In terpre t a t i on6

2. Work on the Microscopical Anatomy of the Braii 9

Section III. METHODS OP INVESTIGATION .... 15

1. Chrome-Silver Inpre^na t ion ... 15

2. The Application of Methylen-Blue . . . 17

a. The Staining Method of Nissl ... 17

b. Intra-Vitum Injection ..... 19

3. Iron Haeuatoxy 1 i n ....... 21

4. The Chloriie of Vanaiium Methol .... 22

Section IV. THE 0BL0N3ATA 23

1. General Morphology of the Oblongata ... 23

2. Review of Nerve Components ..... 26

3. Neurones of the Ventral Cornu .... 32

a. Tract-Neurones ....... 32

b. Comtii S3ur al Neuron'ss

4. fjobus Va?i and Fasciculus Cominunis

a. Neurones of the Lobus Va^i

b. Termination of Comnnunis Pibres
o. The Fasciculus Communis

5. The Viscero-Mo tor Nucleus

6. General Cutaneous Nucleus and Spina

a. The Molecular Layer

b. The Substantia Gelatinosa

c. The Deeper Neurones

d . Temi na t ion of General Cutaneou

7. The Tuberculura Acusticum

a. Molecular ani Granular Neurones

b. Purkinje Neurones

c. Termination of Acus ti co-Late ral

d. Theoretical Conclusions

8. Supporting Elements . . . .

a.Ependyma
b. NeurOi^lia

9. Summary of the Oblongata

1 V Tract

THE CEREBELLUM 63

l.TheNeuronesof Purkinje , . . . . 64

2. The Molecular Layer ....... 68

3. The Granular Layer ....... 71

a. Granular Neurones Proper . . . . .71

b. Gol?i Neurones ....... 73

4. Supporting Elements ....... 74

a. Ependyma ........ 75

b . Neuro^l i a . . . . . . . . .76

5. Architecture and Physiology of the Cerebellum . 79

6. Evolution of the Cerebellum ..... 82

7. Summary of the Cerebellum ..... 85

Section VI. THE MIDBRAIN 87

1. The Tectum Mesencephali ...... 88

a. Termination of the Opticus. S t r . Med ul . Prof un . 89

b. The Superficial Neurones . . . . .91

c. The Middle Neurones ...... 91

d. The Deeper Neurones
The Central Gray Matter

a. The Roof- Nucleus

b. The Nucleus of the Oculomo tor i us
o. The Nucleus of the Troohlearis

The Ependyna ....,,
Phylo?eny of Midbrain Structures
Summary of the Midbrain

96
102
103
1 05
106
107

Section VII, THE INTERBRAIN

1 . The Thalamus .....

a. Nucleus Strati Grisei

b. Nucleus Geniculatam

2. Epithalamus: The Nucleus Habenulae

3. Hypothalamus: The Lobi Inferiores

4. Supporting Elements ...

5. SuTimary of the Intsrbrain

110
110
111
113
114
115
IIB
118

sction VIII. THE POEEBRAIN ....

1. The Olfactory Lobe

2. The Striatum ....

a. Epistriatum

b. General Striatum

c. Nucleus Postolfactorius

3. The Nucleus Neuroporicus
4.ThePallium

a. Neurones of the Tractus Pal

b. Associative and Commissural

c. Gajal Neurones

d. General Considerations on t

5. Supporting Elements

6. Summary of the ?orebrain

lii
Neui

he Pallium

120
122
123
123
125
128
129
131
132
134
134
136
138
139

Section IX. GENERAL SOMMAPY AND CONCLUSION

Section X. LITERATURE CITED

Section XI. DESCRIOTION OP THE FIGURES

Section' I.

iNTRODOCTORf,

Th3 student of mammaliaa neurology has his attention fix-
ed on a mechanisTi of surpassing complexity. In the pursuit of
his /jork, he is continually touching problems, both morpholog-
ical and physiological, .vhich frequently transcend all his pow-
ers. But the complex nervous skein which he seeks to unravel
is merely the final -nember of a series reaching backwari through
ever simpler and simpler conlitions to the organization of the
primitive vertebrate. In other words, the mammalian brain is
the product of endless modifications wrought in the original
plan of structure by the continual adjustment of nervous mech-
anisms to the play of a shifting environment. What the archi-
tecture of the ancestral vertebrate nervous system may have
been we can never hope to know from actual observation. Port-
unately, there are simple vertebrates existing to-day which
retain many features of primitive nervous organization. To
such animals, the student of neurology must ever turn for the
solution of the problems which vex him in higher fields. One
of these simple vertebrates is represents! by a selachian of

the moderi seas, somewhat specialized in certain directions,
(3)

4

of course, bat ratninin^^, with'il, nuch of tha archnic nervous
or^-iai nation from ivhich higher braius have been griiually evolv-
ed. A study of such a simple brain as that of a selachian con-
stitutes, therefore, a necessary introduction to the more highly
differentiated nervous systems of birds and manmals.

Our kno/fleige of the nervous system is so peculiarly de-
pendent upon methods of investigation that the conceptions held
by us might almost be said to grow out of the teohnique employ-
ed. 'The development of neurological methods daring the last
few years has been indeed phenomenal, and with this advance
there has come the necessity for a re-investigation of many
nervous systems. The results obtained by the earlier observers,
while praiseworthy in themselves, siicply do not furnish the pre-
cise and complete pictures of neurones which modern comparative
neurology requires. And hence it is that the writer has attack-
ed anew the structural problems of a brain which has by no means
escaped the attention of investigators. The phylogenetic value
usually assigned to the Selachii has caused many to examine
the brain of the shark, both anatomically and microscopically.
A historical review of the latter class of researches will be
found in Section I[, 2.

The results set forth in this paper are intended to farther
exact knowledge concerning the external morphology, the internal
organisation, and the architectural relations of the selachian
neurones, while also aiding, it is hoped, in the elucidation of
certain questions of a general character. The particular sela-

5

chian selectei, Vlustelus canis of DeKay, is one in /»hich are
combined a structure fairly repre.~3entative of the Selachii as
a whole, a size convenient for work, and availability in suffi-
cient numbers to meet the rigorous demands of certain methods
of investii|atioa.

This research was begun at the Marine Biological Laboratory
of Woods HolL, where two seasons were spent in the use of an
abundance of living material. The principal study was contin-
ued at the University of Iowa as my regular duties permitted;
while certain collateral lines were followed in the neurological
laboratory of the University of Chicago. A short preliminary
notice of the most important results then in hand was published
in the Proceedings of the Iowa Academy of Sciences, Volume IV.
The research has finally been brought to completion in the
biological laboratory of the Johns Hopkins University.

It is a real pleasure to acknowledge my indebtedness to
the officials of the Marine Biological Laboratory for courte-
sies extending over a period of several years; and to Professor
Brooks, I cannot io less than express my sincere appreciation
of the numerous helpful suggestions given me during the com-
pletion of my work.

SECTION 11.

Hevikv* op &APLIKP Researches.

1. Anaton ical Wofk Involving Errors of Interpretation .
To trace the history of a miscoriceptior) is &. task upon
whicli few can enter with even the least degree of enthusiasni,
but it is necessary for us to notice briefly here certain erro-
neous views which have held sway relative to the hoaiologies of
the parts of the selachian brain. The interbrain, and along
with it the midbrain and often the cerebelluiri as well, have
been variously interpreted by the investigators of the last
three decades. Wiclucho-Maclay ('70) was the first to break
away from the established teachings of vonEser ('?7) concerning
the possible homologies of the several brain-segments. Overlool
ing, in effect, the true interbrain entirely, since he regarded
it as nothing more than a longitudinal commissure, Viclucho-
Kcaclay identified the zwischenhirn in that segment which we
know as the midbrain; the cerebellum, being next in the longi-
tudinal series, consequently stood for his mi ttelkim; while he
found his hinterhim in the small inferior lole of the true
cerebellum. Such an interpretation appears almost inexplicable

to us, but we must not allow oui-selves to forget that the homol-

(6)

7

ogies so ooni'idently traced by us to day are f^rounded on many
neurological studies, the results oi' which were not available
to the investigators of the earliei- period.

Unfortunately for coiriparative anatoniy, the conclusions of
Miclucho-Niaclay were accepted by Get^enbaur, and were incorpora-
ted by him in the second edition of his Grundzii^e ('70), and
were continued in the smaller Grundriss ('74). Appearing also,
of course, in the French translation of the former work by Vogt,
and in tlie English translation of the latter by Eell, the errors
were, through these several channels, given the widest possi-
ble distribution among investigators, with all the prestige of
Gegenbaur's authority behind them. Gegenbaur has rectified
the mistake in the latest form assumed by his text-book ('98),
but it probably 7/111 be many years before the miscliief is fully
undone.

Stied^ {'7c) devoted himself to a correction of the
erroneous conceptions promulgated by K'iclucho-N^aclay and
Gegenbaur. After carefully considering the subject-matter at
issue, he embodied his conclusions in a table in which the
homologies of the several brain-segments are properly set forth.
The work of this author had the effect, at least, of directing
the attention of anatomists once more to the fact that the
homologies of the fish-brain were really in question, resulting,
ultimately, in the true interpretation prevailing at the present
time.

Rohon ('77), although writing several years after the

8
publication of Stieda's pa^er, did not, fully accept ttie work
of that author, but took a position almost between the errors
of Miclucho-Waclay on the one hand, and the truth on the other.
While giving the cerebellun its proper recognition as a brain-
segment, he apparently annexed his regie uentricu lo tertii to
the forebrain, thus leaving the mitlbrain standing for two whole
segments. His z'viechen him, therefore, embraced the dorsal
portion of the optic lobes above, and the hypothalamus below;
while he located the mi t telhirn between and behind these two
divisions of his zwischenKirn. Such an interpretatioi' was
certainly reiiarkable for the ingenuity with which a place was
found where an error might be lodged, but it wis almost, if
not guite equaled by the general homologies drawn by Fritsch
('73), who took the whole midbrain for a secondary vorderKirn .

These several errors, curious as some of them certainly
are, might have little more than a passing interest for us to
day, were it not that they continue to reappear at intervals,
tinging the work of those making claims to a certain degree of
authoritative treatment. As an instance of this kind, it may
be noted that one of our most recent treatises on comparative
anatomy contains a figui-e of the selachian brain with the cere-
bellum designated lobe oitique, and the anterior end of the
oblongata the cervelet' .

•Roule.r; I. ' Ana tomi e Comparee des Anim&ux, Toire 2, P i i» . 1 1 5 7 .
Paris. 1R98.

2. Work on the Microsoopical Anatomii of the Brain.

Anatoniical work on the nervous system of the Selachii was
begun relatively early, but aiicrosoopical study lagf^ed somewhat
behind that on the bony fishes. One of the earliest researches
touching the onicroscopical structure of the selachian brain was
that of Leydig ('5?). This versatile investigator was engaged
in tracing the general organogeny and histology of the rays
and sharks, and so his work upon the brain was not special in
its character. Here, however, he discovered the olfactory
glomeruli, a result altogether sufficient in itself. Of course,
the nature of the microscopical methods then in use did not
permit Leyaig to see much more than the general outline of a
glomerulus and the fibres associated therewith. It remained
for later workers to trace the full significance of his discov-
ery.

The two decades following the research of Leydig witnessed
the unfolding of the germ of a special neurological technique.
Stilling had introduced the method of studying the brain by
means of sections as early as 194?, but the great advantages
to be derived from staining the sections were not realised
until 1R59, when Gerlach soaked his sections in a solution of
carmine. Later, 137?, Gerlach obtained such brilliant results
with gold chloride as to lead to many trials with this reagent,
while some of the possibilities of osmic acid wei-e also becom-
ing known. Finally, chromic acid and the bichromates had come

10
to be recognized as valuatle means for hardening; nervous tis-
sues. Fortunate that man who was permitted to contrilute to
the inauguration of a new era in the comparative study of the
nervous system, the era of microscopicel research. Two inves-
tii^ators extended such a possibility to the field of the sela-
chian brain at practically the same time, Viault publisliing
his results in 1376, and Bohon in 1877.

The research of Via.ult ('7*^) is quite broad in its scope.
It includes a review of the anatomical features of the sela-
chian nervous system; a description of the structural elements
common to all nerve-centres; topographic histology, or the
structure of the several parts of the brain and cord; and,
finally, a consideration of the homologies of the brain-segments.
The figures which accompany the paper are perfectly clear in
their execution, but they represent such a low degree of magni-
fication that they are really little more than diagrams. While
the observations recorded by this author are of the most general
character, he should receive great credit for interpreting the
brain-segments properly at a time when there was much confusion
in this respect.

In the prosecution of his research, Rohon ('77) had all
the stimulating advantages of the laboratory of Glaus, and his
work has a high order of merit. There is a section devoted to
the comparative anatoniy of the cranial nerves and the several
regions of the brain, in both the rays and the sharks. Certain
figures illustrating this portion of the work are familiar to

11

all comparative anatoicists through their reproduction in the
text-books. The histological portion oi" the research is care-
fully vyritten, and it is clear that the writer had seen all
that the technique of the period would demonstrate. We find
in his f injures, therefore, nerve-eel J s represented with some
detail of structure, nerve-fibres showing soire connection with
particular groups of nerve-cells, and fibre-tracts which take
a certain definiteness in their courses. The most noteworthy
discovery made by Rohon was the iachkerne of the midbrain; see
Section VI. But the greatest service which he has rendered
consisted in his pointing out for the first time the many struct-
ural features which the brain of the selachian has in common
with the organisation of higher vertebrate brains.

The next research which we have to notice is that of Sanders
('3*). This author seems not to have been familiar with the
great advances Just made in methods of research, (see infra);
and so we find him rejecting carmine as a staining medium and
expressing a preference for rosaniline because of the clearness
of the pictures yielded by it, while he gives a mere hint of
haematoxylin. The scale of the work is ambitious to a degree
bordering on superficial treatment. There are to be included
the anatomy and histology of the brain, spirial cord, and cranial
nerves in both the rays and the sharks. The histological descrip-
tions usually embrace the general distribution of the nerve-cells
of a given region, followed by exhaustive measurements of their
si?.es. His figures are hardly more than outlines of brain-

IP

sections, exhibitirif; very little detail. His most grievous
mistake lay in his refusal to apply the brilliant generalif^a-
tion on the pallium which Rabl-Biickhard ('8?.) had published
shortly before, rejecting it as an impossible explanation of
the selachian forebrain.

We now turn to the work of an investigator whose privelege
it has been to lay many of the stones for the foundation of
comparative neurology. Dr.^dinger has demonstrated that it is
possible to carry on research in the right way in spite of the
exhausting cares of a physician's life. His earlier work ('38)
includes the consideration of both the eir.bryonic and adult
selachian forebrain as a part of a systematic study of the
forebrain of the several groups of vertebrates. By this time
there had been given to neurological workers two of the most
important methods of investigation yet imagined, the chronie-
silver impregnation of Golgi, and the myelin stain of Weigert.
The former has led, ultinately, to the modern conception of the
neurone as a structural and physiological unit; while the latter,
including here the various modifications of the essential prin-
ciple, has grounded our knowledge of the course of nerve-f ib,res
in the cerebrospinal axis. In the research under consideration,
('S8), Edinger was the first to apply the staining method of
Weigert to the brain of the selachian. Using a counter-stain
to define the nerve-cells more clearly, his results were charac-
teri7ed by a precision not known to the earlier workers. The
chief part of the text is occupied by a description of the

13

fibre-tracts, and the drawings are evidently intended to illus-
trate this phase of the subject alone. The nerve-cells are
described as to distribution and general external morphology,
so far as they are trade visible by the method employed. The
principal aim of the research is an elucidation of fibre-tracts
rather than the investigation of nerve-cells.

In a later research ('9?), Edinger applied essentially the
saire methods to the interbrain of selachians and amphibians,
and the results have a scope similar to those just noticed for
the forebrain. In the later editions of his text-book (19C9) he
has amplified for the comparative portion of the vjork the results
of all his own studies, together with those of others, giving us
the broadest exposition of modern coniparative neurology yet
attempted by any writer.

To Sauerbeck ('9*) belongs the credit of first publishing
results from the application of chrome-silver impregnation to
the selachian brain. The paper contains a very brief descrip-
tion of those neurones and supporting elements which had been
impregnated; by far the greater number of the structures present
evidently were not demonstrated at all. The treatment is quite
unequal for the several regions, and the figures are drawn on a
small scale. While Sauerbeck must not be given credit for the
things he neither described nor portrayed, yet a first attempt
in this field is certainly to be commended.

Schaper ('95), in the course of a series of studies on the
cerebellum of vertebrates, has taken occasion to apply the

14

chroaie-silver method to the selachian cerebellun. His paper
records observeitions with a considerable degree of detail.
Reference to the results of Schaper will be made niore particu-
larly under Section V.

Inspired by the elaborate classification of nerve-cells
in general proaiulgated by Nissl in his writings, Szczawinska
( '9B) was impelled to make a study of the internal structure
of the selachian nerve-cell. His work presents the results of
his researches upon certain types of cells through the use of
methylen-blue, safranin, and haematojiylin stains. The cells
studied were from the sensory ganglia, motor cells from the
cord and oblongata, and cells of Purkinje from the cerebellum.
Szczawinska reached the conclusion that the nerve-cells of
selachians have remained on a low plane of development. In

support of this view he cites certain of his results (1) that

the cell-bodies are usually bi-polar in form; {?) that there is
but slight demarcation between the cell-body and its protoplas-
mic processes; and (3) that the chromophile substance is less
differentiated than in teleosts and hif^her vertebrates.

Section III.
I/ETHODS OP Investigation.

The technique employed in the course of this research has
covered a wide range. In fact, every process holding any promise
of value has been given a careful trial. It seems desirable to
describe here, however, only those aiethods which have contrib-
uted most largely to the final results.

1. Chrome-Si Iver Impregnation .
The production of a chrome-silver deposit in nervous ele-
ments is far more than a simple chemicsl reaction between the
potassiun, bichromate and the silver nitrate employed. There
is to be added as a prime factor, the chemistry of the nervous
tissues themselves. The substances present ir the nervous ele-
ments enter into the reaction to such a degree that the results
are either positive or negative according to the character of
those substances. In the economy of the animal, serving both
as the basis for nervous activity and produced as the result
of it, there is a constant round of metabolic change altering

the chemical composition of the nervous tissues. In my work

(15)

16
upon Viustelus, it /las soon found that strict account liad to be
taken of the physiolofe'i c&l state of the animal. An individual
fresh from the pursuit of his prey in the open sea ^ave alto-
gether different results from one which had been kept for some
time in a small aquarium, simply because the reactions of the
tissues in the two instances were quite different. And so,
before even fair results with chrome-silver impregnation could
be secured, it was necessary to make a careful study of physio-
logical conditions.

It was also found that selachian nervous elements lend
themselves but grudgingly, at best, to the reaction desired.
Numerous trials were made of the several published schemes for
securing impregnation, but particularly of the procedure indi-
cated by Golgi ('94); by Ramon y Cajal ('94); by Flechsig ('89);
by Cox ('91); and by Strong {'9^, '96). 5.very application
practicable was also made of formaldehyde as a constituent of
the reagents employed. A general critique of these processes
has already been given by me in a former paper ('97a).

The slices of perfectly fresn brain from the most active
animal procuratie were placed in the "rapid" hardening mixture
of Golgi. The pieces were always small, not over two millimeters
in thickness for, e.g., a transverse section of the forebrain.
The proportion of the hardening fluid used embraced one part
of 1% osmic acid to four parts c.5% potassium bichromate, and
this reagent was used in liberal quantities. The proper dura-
tion of hardening was influenced by the temperature of the

17

room and tlie physiological state of the animal, but an average
length oi' time was three days. The greatest clearness of impreg-
nation was secured with silver nitrate solution of 0.75^ strength.
In the preparation of serial sections, the most desirable clear-
ing agent was found in a mixture of oil bergamot, oil cedar-
wood, and melted carbolic acid crystals, equal parts. After
being hardened in chloroform, the celloidin blocks were placed
in the clearing mixture, and they were kept flooded with the
oil during cutting. The above mixture clears the block rapid-
ly, it may be used repeatedly, and it has the additional advan-
tage of allowing the preparations to be kept in it for some
time without impairing the impregnation. The sections were
cut 75 micra in thickness.

2. The Application of Methylen-Blue .
Methylen-blue holds so many possibilities as a neurolog-
ical reagent that we are doubtless but crossing the threshold
of its use to day. I have applied this aniline in every way
of which I could learn, and the most important results are set
forth below.

a. The Staining Method of Hissl. — Nissl's description of
his method ('94) called for the fixation of the tissues with
alcohol. This has proven an unsatisfactory part of the tech-
nique for my work. Better cytological preservation by far has

IB

been secured throu^,h the use of tlie chrome-oxalic mixture of
Graf ('9?). This reagent appears to have escaped the general
attention of microscopists, at least no mention is made of it
in the fifth edition of Lee (1900). The composition is here
given:

Oxalic acid, 8% aq.sol. - - - 300 c.c.

95? alcohol - - - - - - - 15 0C.C.

Chromic acid, \% aq.sol. - - - 160 c.c.

Mix in the order as named.

Quite small pieces of the brain were fixed in this fluid
for six hours, and the fixing agent was then washed out with
'?0% alcohol. Sections were made by tlie paraffin method.

The slide was taken from distilled water, and the steaming-
hot stain of Nissl was poured over the sections, five minutes.
The excess stain was rinsed away with distilled water for the
briefest possible time, and the water clinging to the slide
was absorbed with filter paper. Differentiation with the anilin-
aloohol of Nissl took but a few seconds, being stopped by flood-
ing with oil of cajeput just as soon as the sections took on a
delicate rose tint. Clearing with the oil of cajeput was aided
by holding the slide for a few mon.ents in gentle heat. Mount-
ing was done in colophoniuir dissolved in xylol. The staining
is remarkably precise, and the color has shown no tendency to
fade.

The counter-staining methods described by Held ('95),
and by ITarrington ('99), were also applied, with certain modi-
fications found necessary. The erythrosin mixture of Held

19

was gently warmed, pour-ed over- the slide for ten seconds, and
then washed away quite thoroughly with distilled water. Stain-
ing was done with either the pure stain of Nissl, or with the
same diluted with an equal volunie of ^^ acetone; the results
did not seem to differ very much. In either case, the stain
was heated and allowed to act for five minutes. Differentia-
tion with 0.1* aluo-i solution for Just a few seconds, until the
sections appeared distinctly red, was followed by a brief rins-
ing witl' water. The results given by this process have been of
value as an accessory to the pure methylen-blue stain, but they
are far from supplanting the original method.

b. Intra-Vi tuT Injeoticn. — The coloration of the nerve-
cells through intra-vitua. injection of methylen-blue was given
a most thorough trial, a large nuirber of animals being utili-
zed for this purpose. The subcutaneous injection preferred by
Meyer ('96) is not practicable for Mustelus because of the
absence of either loose areolar tissue or of lymph spaces.
The syringe was therefore inserted directly into the vascular
system. A nf solution of methylen-blue, FX brand, was injected
some four times during the course of an hour. Beginning with
a small quantity, the amount rose successively until as much
as 30 c.c. was introduced in the final injection, making some
50 c.c. in all. This whole process was governed, however,
not by fixed quantities of the reagent nor by exact periods of
time, but by the stopping of tlie heart's action and the blue-

20
ness of the animal. Half an hour after- the firal injection,
the brain was removed, cut into thin slices, and then exposed
to the air until the tint had become a brighter blue. In the
conversion of the unstable methylen-blue stain of the fresh
tissues into the insoluble form, I have not been successful
with the method recommended by &ethe i'd'z). The use of the
picrate of ammonia as a preliminary fixer has seemed to actual-
ly impair the clearness of the final preparation. I obtained
the best results with the solution given by Weyer ('96):

Distilled water ----- lOO c.c.

Affimonium molybdate - - - - - 10 granns

Hydrochloric acid - - - - - 10 drops

Heat the first two ingredients together, then add the acid.

The pieces of brain were placed in this mixture, cooled
with ice, for four hours. They were then washed with iced
water for two hours. Dehydration with cooled alcohols, and
imbedding in paraffiri were hastened as much as practicable;
in fact, it is well to have the tissues in paraffin on the
same day when the injection was begun. The preparations were
used chiefly for the study of the architectural relations between
the neurones, and so the sections were cut quite thick.

The results given by this method are characterised by
exceptioi.al clearness, due, in large measure, to the selective
coloration of certain neurones, only. The attainiient of the
desired end is far from constant, however. After experience
had shown the rule, care was always taken to apply this tech-
nique only to those animals which had been in an active condi-

21
tion, fresh from the open sea, if possible. But even with
this precautionary recognition of physiolofsical conditions, so
far as they could be readily determined, there apparently yet
remained some unknown factor which caused a negative result
in some instance where it was least expected.

S, Iron HaematoxiJ I in .

This reagent was imagined^ by Heidenhain ('9?) for refined P
cytological work, but it truly has a place in neurological in-
vestigation. It is a most excellent stain for defining the
internal structure of the nerve-cell, and also for the tracing
of nerve-fibres. F'or the latter purpose, iron haematoxylin
has proven itself preferable in this research to the stain of
Weigert, since it defines the axis-cylinder instead of the
myelin, permitting fibres to be followed through their rariiifi-
cations entirely to the terminal arborisations.

Fixation of the tissue may be done with any good fluid.
Where the tracing of axones, only, is desired, 10"? formalde-
hyde cannot be surpassed; but where the aim is purely cytolog-
ical, either the chrome-oxalic mixture of Graf or the fluid of
Plemming will give superior results. For work on axis-cylinders,
celloidin sections were made 30 micra thick; and for- the minute
study of the nerve-cell, thin sections were cut by the paraffin
method. .

The sections were brought from distilled water into the

22

mordant of 4"? iron aluii for two hours. The excess mordant was
then rinsed away with distilled water. Staining with 0.5^
aqueous haematoxylin required at least four hours for entirely
satisfactory definition. Clean tap-water was used for washing
out the uncoDibined stain, sirjce this appears to fix the lake
more firmly. The stain was differentiated with a 2% solution
of iron alurr, frequently renewed. This process was observed
with the rcicroscope, and when the desired effect had been ob-
tained, the sections were transferred to tap-water. Thorough
washing at this stage is necessary to prevent fading of the
stain, and the slight alkalinity of ordinary tap-water appears
to be a factor aiding in its preservation.

4. The ChloT-iie of Vanaiiurr Kethod.

The technique required for staining with the chloride of
vanadiuir, method of Wolters ( '90) is somewhat troublesome, but
the results, when obtained, certainly justify the means. Nerve-
cells, axis-cylinders, ependyma, and neuroglia are all defined
in one and the same section, Mo other method known to me gives
so comprehensive a picture for general study. Its sole value
lies, however, in the purely general scope of the results.

It is hardly necessary to give a description of the process
here, since its essentials are outlined in Lee (19C0, p. 410).

Section IV.

The Oblongata.

1. General Morphology of the Oblongata.

In Wustslus, the transition of the architecture of the
spinal cord into that of the oblongata is traceable with a
degree of definiteness which rarely obta.ins in ottier animals.
It is therefore possible to contribute toward the solution of
certain probleiris which ve>. the study of this highly specialised
region of the mamnialian brain. Only an introductory survey
of the entire field will be presented here, leaving the devel-
opment of details and the consideration of special questions
to the following subsections.

As the canalis centralis widens into the fourth ventricle,
the several structures of the cord lateral to it are pushed
into more and more widely divergent positions, retaining, howev-
er, essentially the same mutual relations to each other. Con-
comitant with this divergence, the dorsal ependyma becomes
broadened to form the morphological roof of the fourth ventri-
cle(Pig.P,t.c.p. ).

In approaching the oblongata, the ventral cornua are en-
croached upon more and more by commissural fibres until. Just

(?3)

24

above the level of the first spinal nerve, the n/ammalinn hypo-
glossus, these nuclei disappear altogether as continuous col-
lections of nerve-cslls. A small number remain associated
as the nucleus of the abducens, while the remaining neurones
become scattered through the formatio reticularis (Fig.S.c.n.
and t. n. )

The intermediate 2one and the dorsal cornu of the cord
are relatively small in si^e, but as the paired halves diverge
to right and left of the fourth ventricle they increase in
mass and assume characters and functions of a special order
for each region. The intermediate gray matter contributes the
lobus vagi and the viscero-motor nucleus of the oblongata;
while the dorsal cornu becomes specialized as the general cuta-
neous nucleus. We will note each of these in turn.

The lobus vagi is a longitudinal elevation in the lateral
wall of the fourth ventricle ( F'ig. 1, l.vg. ) . Anatomically, it
is one of the most striking features of the oblongata because
of the row of bead-like prominences into which its surface is
thrown. The position of the structure as seen in a transverse
section is represented in Pig.?, l.vg. The lobus vagi is the
terminal station for communis components of the Vll, IX, and
X nerves; see Subsection 4. Certain of these fibres, instead
of passirg directly to their termination here, enter- a compact
bundle and run posteriorly to the spinal cord. This tract is
known as the fasciculus communis. Its position in the oblonga-
ta will be seen in Fit-'.?, f.c.

25

The viscero-Diotor nucleus is a column of l^iri^e nerve -cells
imbedded in the lateral wall of the fourth ventricle (F'i^.?,
v.tn.n.). Different portions of this coluan are known as the
nucleus ambiguus, and the motor nuclei of the VI] and V, res-
pectively. The axones for the motor roots of the V, VII,
IX, and X nerves take their origin from these nerve-cells.
It is a curious fact, however, that but few of the axones pass
directly into their nerve-roots, but take a course first in the
fasciculus lon^itudinalis dorsalis. This is a massive, paired
tract, the two bundles lyin? side by side beneath the floor of
the fourth ventricle (["ig. ?, f . l.d. ). There are present in the
dorsal longitudinal bundle nerve-fibres from several sources.
Into this crowded highway the axones from the viscero-motor
nucleus penetrate, to finally emerge as the motor roots of
their respective nerves.

The dorsal cornu of the cord is continued into the oblon-
gata in an enlarged condition, and it becomes associated with
sensory fibres of the V, IX, and X nerves to form the general
cutaneous nucleus (Big. f', g.c. n. ). The neurones of this nucleus
provide a priniary termination for certain general cutaneous
fibres; while others turn backward to the cord as the spinal V
tract. At the posterior levels of the oblongata this system
appears dorsal to all other structural features, the pair form-
ing the rounded, crest-like margins to the fourth ventricle.
Proceeding anteriorly, the nucleus is pushed into a position
both more ventral and more lateral by the superposition of a

2«

new structure, the tuterculua acusticuir-.

The tubsrculun acustiouni is shown in F'it^s. 1 and ?, t.a.
Tt extends posteriorly from the restiformis along the lateral
margin of the fourth ventricle some three quarters of the dis-
tance to the calamus scrijjtorius, tapering as it proceeds. It
is separated from the general cutaneous nucleus below it by a
fissure which reches well toward the limitans interna. The
outer zone of the acusticum is structurally continuous wit!
the cerebellurr,; it is known as the cerebellar crest (Pig.?,
cb.cr.). The acusticum is the centre for the nerves of the
lateral line sense-organs and the internal ear. Its interpre-
tation will be considered in Subsection 7.

2. Feuie'v of Merve Components .
A proper point of view for the structure of the oblonga-
ta can best be obtained through familiarity with the problems
of the cranial nerves pertaining to this region of the brain.
Reference to the text-books of descriptive anatomy will discov-
er hardly a trace of the conceptions which dominate the modern
morphology of nerves. The discovery made by Sir Charles Bell
as to the character of the dorsal and ventral roots of the
spinal nerves was one which represented a distinct advance in
sound physiology, but the application of Bell's formula to the
cranial nerves has not been productive of sound morphology.
The effort to compare the cranial with the spinal nerves on the

27

simple basis of "sensory" and "motbr" could not avoid leading
to many dogmatic positions concerning, the real character and
ultimate distribution of many fibres. An attempt to solve such
intricate problems through so mechanical a method could hardly
be otherwise than faulty, particularly when applied to the
specialised conditions of the mammalian nerves. A thorough
study of the less modified cranial nerves of the Ichthyopsida
is a necessary preparation for sound morphological work in the
higher field.

The views of cranial nerves held by the neurologists of
to day were founded less than a decade since, but the germ of
the central idea is traceable to a somewhat earlier date. Gas-
kell in a series of publications ( 'S6, '88, '89) was making the
attempt to solve the metamerism of the head and the origin of
the vertebrate nervous system through a study of the nerves.
He took occasion to show that a spinal nerve not only embraces
the sensory and motor fibres of Bell, but that its structure,
distribution, and function, as well as the arrangement of its
central nuclei, lead to the divisibility of the nerve into two
parts. One part is somatic, innervating the external surface
of the body, and the muscles derived from the muscle-plates.
The other division is splanchnic, supplying the internal organs
and surfaces, and those muscles which Gaskell characterises as
"derived from the lateral plates of the mesoblast". Gaskell
attempted to show, further, that the cranial nerves arise from

28
centres homologous v?ith the spina] centres, likewise divisible
into somatic and splanchnic groups.

The impetus given by Gaskell has led, ultioiately, to the
modern conception that a spinal nerve embraces neurones derived
from four distinct sources, with as many different distribu-
tions. There are, then, to be distinguished in a spinal nerve:
(a) Somatic motor fibres. These neurones have their cell-
bodies situated in the ventral cornu of the cord, their axones
emerge through the ventral root, and they are distributed to
the body musculature, (b) Somatic sensory neurones, the cell-
bodies of which comprise the dorsal ganglion, and their fibres
connect peripheral end-organs with the dorsal cornu of the
cord through the medium of the dorsal root. These two divis-
ions of a spinal nerve comprise the principal number of fibres
in the two roots, (a) Viscero-motor; these fibres take their
origin from a group of cells, the paracentral nucleus of Onuf
and Collins ('9B), lying lateral to the canalis centralis;
they emerge through both the ventral and the dorsal roots,
and are distributed to the non-striated muscles of the viscera.
(i) Viscero-sensory neurones, from the viscera, through the
dorsal root, to termination in the intermediate zone of the
gray matter.

A new era in the investigation of cranial nerves was inau-
gurated by Strong ('95) when he made the application of these
principles to larval amphibians. Strong found that it is prac-
ticable to recognize certain distinct classes of fibres or

29

components of the cranial nerves, which are to be distinguished
from each other by their si?,e and histological characters, by
their central origin or connections, and by their ultinate
distribution. Kingsbury ('97) made a careful extension of these
findings to several ganoids and teleosts; while Herrick ('97,
'9B, '99) has traced the conditions in the bony fish k/enidia
with adniirable clearness. Certain conclusions reached by Johns-
ton ('93b) from his study of the ganoid brain stand apart from
the general trend of recent work, and to these we shall return
further on.

The principles developed by the researches of Strong,
Kingsbury, and Herrick, may now be applied to the cranial nerves
of Vustelus. There are to be distinguished five systems of
nerve components:

a. The Somatic Vctor System. — This system of neurones
is homologous with the ventral-cornu neurones of the spinal
cord. The fibres take origin from cells having a ventral loca-
tion in the brain, and they innervate striated somatic muscles.
The only representatives of this class are the nerves of the
eye-muscles, the 111, IV, and VT, respectively.

b. The General Cutaneous System is so called because con-
cerned with the innervation of the skin of the head, but it is
not associated with specialised peripheral sense-organs of any
kind. Its fibres are components of the V, IX, and X nerves,
and they are homologous with the somatic sensory spinal fibres.
The cell-bodies of these neurones lie ir sensory ganglia, and

30
the central termination is comparabJe to tliat of the somatic
sensoi-y spinal fibi-es. Part of the general cutaneous fibi-es
terminate in the t^.eneral cutaneous nucleus, the homologue in
the heaj of the dorsa] cornu; see Subsection ^; but others tur-n
into the spinal V tract and take a course posteriorly for ulti-
mate termination in the dorsal cornu of the cord. There can
be no question as to the identity of this system with the somat-
ic sensory system in its siiripler condition.

c. The Viscero-Votor- Siistett. — Fibres of this system
form the motor roots of the V, VII, 1)(, and X nerves. The cell-
bodies from vvhich the sxones f^rise form s column of cells later-
al to the fourth ventricle, known, according, to the level, as
the nucleus ambif-'uus, and the motor nuclei of the VII and V,
respectively. This column of cells is the cranial continua-
tion of the paracentral nucleus of the cord, and the homolot^y
is rendered complete by the distribution of the fibres to the
visceral musculature.

i. The Communii? Si/stem. — This term is used in the sense
defined by Herri ok ('99,p.?0?), and as the equivalent of the
fasciculus communis of Strong ('95) as applied to the system
in the tadpole, and of Kingsbury ('97) as used for various
fishes. Communis fibres are components of the VII, IX, and
X nerves. They are wholly sensory. They innervate visceral
and mucous surfaces, and also taste-buds and those special i7ed
sense-organs of the skin (end-buds) not referable to the later-
al line system. The fibres are characterised by their snail

31
size, and tiiey are dist.i ntfuisfiat le from other- coniijonents of
the same nerves by this feature. They terminate in the lotus
va£i. The greater number of them in J/ustelus pass directly
to their central terminatioii witliout entering the fasciculus
communis, using this term in the sense as originally applied
by Osborn ('S3) to a definite longitudinal tract.

The communis system is homologous with the viscero-motor
system, the lobus vagi in which central termi nation occurs
being the continuatior: into the brain of the lateral or inter-
mediate ?cne of gray matter in the cord. The viscero-sensory
system in its original condition is, however, of far less im-
portance than its cranial representative. The communis system
not only innervates visceral organs, but it has come into rela-
tions with taste-buds and peripheral end-buds, as well. It
therefore is a system having a very considerable magnitude.

e. The Acustieo-Lateral System is concerned exclusively
with the innervatioii of the internal ear and the organs of
the lateral line. Its fibres are components of the VIl, VIH,
and X nerves. They penetrate the tuberculum acusticum for
immediate termination there, or for a course farther forward
into the cerebellum. The significance of the latter termina-
tion will be discussed later. The components of this system
in the VIl and X are to be readily distinguished from other
fibres associated witlj them by virtue of their great diameter.

The acusti co-lateral system is not represented in the
spinal nerves, as the four preceding cranial systems are. The

3?
evolution of tliis system is bound up with tliat of the peculiar
series of sense-or-^nns which it supplies. Until the evidence
is nior-e nearly coiriplete as to the emLryolo^y and affinities of
the lateriil line and its nerves as a whole, we really are not
warranted in oiaking any positive assertion as to its phyloge-
ny. Some evidence as to the possible orit^in of its centre
will be ^iven in Subsectioi' 7,

5. Seurones of the Ventral Ccrnu.
Upon reaching' the level of the oblongata, the ventral
cornua of the cord beconie broken up as distinct collections of
nervous matter. Somatic irotor neurones froD> this source are
grouped into the nucleus for the VI nerve, but elsewhere there
are only isolated individuals lyin^ between the fibres of the
formatio reticularis. In an earlier paper touching this sub-
ject ('9?b), I left the inter[,retation of these scattered neu-
rones undecided. It is no/f certain that they correspond to
the coiiimissural cells and the tract-cells ( vorderstrangzellen ),
respectively, which vonLenhossek ('94) has described from the
spinal cord of the selachian.

a. Tract-Keurcnes. — The ventral tract-neurones are scat-
tered over the mid-ventral field of the oblongata on each side
of the raphe ventral to the dorsal longitudinal bundles (Fig.?,
t.n.). They are to be readily distinguished from all other

33

neurones of this ref:'ion by tliei r- sif.e, for they are really
giants. Their size is exceeded only by the neurones of the
midbrain roof-nucleus described in Section VI, Subsection 2.

As to sxtei-nal form, the tract-neurones have a wide range;
a representative individual is drawn in fi||.?, t.n. There are
fron, three to five dendrites, and their disposition controls
the shape of the cell-body to a very high degree. The den-
drites may be given off at opposite extremities of the cell,
in which case the outline of the cell-body is a iruch elonga-
ted one. The form is rounded or stellate -/vhen the dendrites
are spaced at equal intervals. A dendrite is always a massive
process, very wide at its base, tapering quite gradually, and
reaching far out into the surrounding nervous matter. It gives
origin to but fev? branches.

The axone always arises from the body of the cell. It
runs for a short distance in the transverse plane and then
turns into a longitudinal bundle of fibres on the same side
of the oblongata, or even on the opposite side.

The internal organisation of a tract neurons is shown in
Pig. 44. The cytoplasm is voluminous in quantity, investing
the nucleus witb a thick layer on every side. The outline of
the nucleus is regular, and in form may be circular or oval.
There are always a remarkally soiall number of coarse chroiiiatin
granules, the chromatic material being distributed in the
form of a delicate reticulum. Subsidiary nucleoli are rarely
present.

34

The cytoplasm is, as already noted, great as to actual
quantity. Some neurones when stained by the metiiod oi' Nissl
absorb the methylen-blue equally throughout all parts of the
cytoplasm, and hence they appear almost homogeneous; the sig-
nificance of this fact is noted below. Other neurones in the
same section exhibit a large quantity of tigroid substance;
the neurone represented in Fig. 44 is of this type. In the
vicinity of the nucleus the ti groid-bodies are usually trian-
gular in outline, and some of them have a very considerable
size. Tigroids are found far out in the dendrites, assuming
here a lenticular or even a linear form, their long axes paral-
lel «ith the course of the dendrite. A finely granular axone-
hillock lies at the origin oi the axone. In rectangular cells,
as the one figured, the hillock ir,ay be spread so widely as to
assume a disk-like form. The tigroids tend to become somewhat
smaller in the region of the hillock.

The tract-neurones are not demonstrated readily with either
chronie-silver impregnation or the intra-vitum injection of
methylen-blue. The homogeneous coloration assumed by some of
them with Nissl staining has already been recorded. The charac-
ter of such micro-chemical reactions indicates clearly that
many of these neurones are not physiologically active, at least
not all of the time. The axones from some of theni niay enter
the motor root of one of the anterior spinal nerves, and, being
in an active condition, give indications of it in a well-marked
store of tigroid substance. Other neurones, however, chaining

35
together hit^her and lower levels of the oblor^ata, have come
to have their functiors largely usurped ty the develojiiient of
more speciali?.ed tracts. Such neurones are, therel'ore, degra-
ded to a far lower plane of metabolic activity, and they respond
but feebly to those of our stains which depend upon the presence
of definite chemical constituents of the protoplasn.

b, Corrfiissural Heurones. — The commissural neurones are
readily distinguishable from the tract-neurones by their small-
er si7.e (Fig. ?.,c.n. ) . Ooii.missural neurones have a wide dis-
tribution. They are scattered between the tract-neurones in
the vicinity of the median raphe, and they also are to be found
in all parts of the lateral region as far dorsal as the base
of the general cutaneous nucleus.

The external morphology of a commissural neurone is repre-
sented in Fig.3,c.n. The cell-body is relatively small in
proportion to the extension of the dendrites and the axone.
The form of the cell is usually an elongated oval or an irreg-
ular triangle. The dendrites are quite often only two in num-
ber, arising from the ends of the cell. The course taken by
the dendrites does not seem to be affected in the least by the
tracts of fibres of the formatio i-eticularis, since it fre-
quently takes them obliquely through a bundle of arcuate fi-
bres. The dendrites are stout at their bases. They give off
only a few short branches. The principal axis of the dendrite
pursues a rather even course for a comparatively long distance.

36
taperiiif' gradually to the end. Its surface bears but i'ew gem-
aiules.

The axone m&y arise from tlie body of the cell, or, when
the long a>is of the cell is horizontal, from one of the den-
drites. It takes a course directly for the opposite side of
the oblongata, giving off collaterals near its point of origin
and then remaining free from theni. The function of such an
axone is, doubtless, commissural.

The internal structure of a commissural neurone from the
right side of the oblongata is drawn in ?ig.45. The cytoplasn
is always far less voluminous in proportion to the si^e of the
nucleus than in the tract-neurones. The nucleus is almost in-
variatly oval in outline, with its major axis disposed the
long way of the cell. The nucleolus is not conspicuous. The
chromatic material is distributed in elongated, irregularly
formed strands which lie throughout all parts of the nucleus.
The several masses may be connected with each other, but only
the faintest suggestion of it is indicated with the highest
magnification.

The cytOf.lasm contains a considerable quantity of tigr-oid
substance. A few broadly triangular tigroids are scattered
round the nucleus, sometimes in the condition of nuclear caps.
The greater part of the tigroid substance is to be found at
the wide expanse where the cell-body merges into a dendrite.
Here the form of tigroid is a greatly elongated triangle. This
is replaced farther along the dendrite by spindle-shaped or

37
linear masses. A small axone-hil lock of finely Granular inatter
lies just witl.in the orit^in of ttie axone. The tit^U-oid -bodies
near it are snialler and more irret^'.ular than elsewhere.

The commissural neurones, in contrast with the tract-
neurones, always resj^ond to the stains applied; this indicates
a uniformly active physiological condition. Commissural neu-
rones doubtless have a not unimportant place in the economy of
this region of the brain, into which so many iaipulses sweep
through afferent nerve-fibres to great vital centres. The
chaining together of the opposite halves is precisely what we
should expect to find under such conditions, and the mechanism
is provided by the comi/iissural neurones Just described.

4. iobus Vagi and Faseiovlus Communis .

The Iobus vagi is the terminal station for the communis
fibres, components of the VII, IX, and X nerves. Its general
morphology has been outlined in Subsection 1.

The structure of the Iobus vagi embraces a narrow 2one
next to the limitsns interna occupied chiefly by nerve-fibres;
and a deeper part in which there are both neurones and termi-
nating fibres. The course of the nerve-fibres will be consid-
ered later.

a. neurones of the Lcbuf> Vagi. — The constituent neurones
of the Iobus vagi are many in number and closely crowded. They

39
are comparatively small in sif.e, and are referable to type II
of Golgi. The external morpholot^y of a neuror.e is shown in
Pi£.4.

The shape of the cell -body ran^'es frou^ almost triangular
to broadly oval. The axone may emerge directly from the cell-
body, or it may spring from one of the larger dendrites. It
pursues an irregular course away from the limitans interna
into the deeper levels of nervous matter. Before proceeding
far, it breaks up into a widely-spread arborization such as is
characteristic for neurones of this type. I believe from the
course of the axones that they constitute a means for trans-
mitting impressions to the viscero-motor nucleus; see Subsec-
tion F5.

The dendrites are three or four in number. They are quite
stout near their origin, they taper gradually, and they do not
become very fine at their terminations. Their morphology is
simple, since there sre only one or two branchings at most.
Their lengths may be as great as that of the axone, and so the
dendrites from all of the neurones here interlace to form a
veritable Jungle. The surface of a dendrite bears a few gem-
mules, together with certain small knobs and elevations of
various shapes.

The internal organization of a neurone from the right
lobus vagi is shown in Pig. 4^. The nucleus is always large in
proportion to the bulk of the cell. In some instances there
is only a thin film of cytO[.lasn enclosing it at certain points.

39

The nucleus is a more or- less perfect oval, holding:' one or
more nucleoli. The chromatin is distributed in a few thin,
branching strands which give the appearance of joining in soine
parts of the nucleus. The chromatir never stains intensely in
these neurones, and so the nucleus is represented by light col-
oration in the figure.

The cytoplasir lies principally in the broad areas where
the cell-body merges into the dendrites. Its tigroid substance
is never collected into large masses. In the region of the
nucleus, the tigroid material is chiefly in the condition of
medium-sized granules, with a few small triangles intercalated.
The dendrites have fusiform or linear tigroids scattered at
irregular intervals. A very small sxone-hill ock gives origin
to the axone. In the specimen figured, this lies at the side
of the most massive dendrite, but this location is not the rule
for these -neurones.

b. Tefmination of Communis Fibres. — Ey far the greater
number of the coaimunis fibres from the VII, IX, and X nerves
pass directly to their termination in the lobus vagi, a few,
only, entering the fasciculus cooimunis, described below.

Fibres penetrate the lobus vagi for ultimate termination
chiefly from the dorsal side ( Fig. ?, c. f . ) . They reacli this
position by a sweeping curve which carries them to an ever
higher level as they run inward from the exterior-. These in-
coming fibres constitute a thin stratum next the limitans inter-

40
na, the neurones lyint' just beneatli. The final arborization
occurs near the body of some neurone. It is of a narrowly
branching type, with fine, bristle-like twi^s terminating the
branches (Fit;. 4,c. f . ) .

c. The Fasciculus Communis. — This remarkable tract was
given the narce it now bears by Osborn ('85) in recognition of
the coniaion relationship of several cranial nerves to it. Va-
rious authors had noticed the fasciculus communis previous to
this titoe, but they had failed to grasp its significance. ?tie-
da ( '7Sa, p.4?9) had noted the presence of such a bundle in the
spinal cord of the selachian; Ahile Bohon ('77, p. 4*) described
it from the selachian brain under the name fasciculus longitu-
iinalis lateralis. The latter writer conjectured that it might
pertain to the tegmental system.

In f/ustelus, the fasciculus communis is a very sharply
defined tract extending posteriorly from the VII nerve into
the spinal cord, where it lies close beside the gray matter
dorso-lateral to the canalis centralis. Its position in the
oblongata is shown in Fig.?, f.c. During a part of its course,
the viscero-motor nucleus is traversed by it. Some of the
communis fibres, instead of pursuing a course to the lobus
vagi for termination there, turn downward into the fasciculus
communis. Thence they are carried posterioi-ly into the spinal
cord for their ultimate distribution.

From the account given by Strong ('9-5) of the fasciculus

41
communis in amphibians; and by Herriok ('99) for the bony fish,
I concUne that this tract is developed in Niustelus to a con-
spicuously less degree. The sit'nificance of this smaller siee
for the brain of the selachian is not clear. The communis
fibres are essentially viscero-sensory , but there have been
annexed to the systeni certain external sense-organs, such as
taste-buds and end-buds. There are no observations indicating
marked differences between the visceral connections of this
system in >/,ustelus as compared with teleosts and amphibians.
In fact, the archaic value and deep-seated physiological impor-
tance of such connections in all vertebrates would lead us to
infer a considerable similarity in related groups. We thus
appear to be thro.vn upon the more recent additions to the sys-
tem for the explanation. It seems to me that a comparative
investigation of end-buds and taste-buds will contribute much
toward the solution of questions pertaining to their central
tracts.

-5. The V iscero-Votof Kucleus.
The adoption in this paper of the name viscero-motor nu-
cleus expresses the need for a general term which shall includ(
all members of the morphologically continuous colunn of cells
giving origin to the motor fibres of the V, VI], IX, and X
nerves. This nucleus is the continuation into the oblongata
of the paracentral nucleus of Onuf and Collins ('9B), and the

42

components having, their ori^'ifi here innet-vtite viscera..

The nucleus is composed of quite large neurones. The only
larger ones in the oblongata are the gigantic tract-neurones
described in Subsection 3, a. The cells are arranged in a com-
pact cluster, as seen in transverse section, and this is tra-
versed during a part of its course by the fasciculus communis.
Pig.?, v.m.n., illustrates the disposition of the neurones
and their characteristic forms.

The cell-body has its form influenced by the number of
its dendrites, ranging from triangular to stellate. The den-
drites are always several in number. They are massive proc-
esses, often arising through such wide bases that it is diffi-
cult to say A'here the line of demarcation between dendrite
and cell-body should be drawn. The dendrites branch freely,
and the closeness with which the neurones are arranged gives,
therefore, a most complicated tangle of interlacing branches.

The internal organization of a viscero-motor neurone is
represented in Fig. 47. The nucleus is central, or only slight-
ly eccentric, and it has an evenly rounded contour. The nucle-
olus is large; there is rarely a subsidiary nucleolus. The
nuclear reticulum has a coarse mesh which exhibits great clots
or lumps of chromatin at the points of intersection.

The cytoplasm contains tigroid-bodies of various sizes.
The largest masses lie in the field of the nucleus. The ones
next to the nuclear membrane may assume the form of nuclear
caps. At the periphery of the cell, the prevailing form of

43

titifoid is much elonc^.ated. It lies parallel with the marP'in of
the cell. Tigroid masses are continued far along the dendrites,
even into the tertiary branches. They are disposed parallel
with the course of the dendrites.

In that side of the cell from which the axone arises,
the tigroids have a special arrangement. The axone-hillock
(Fig. 47, ax. h. ) is an oval mass of finely granular material,
and the tigroid-bodies are packed rather densely round it.
The form of tigroid is also less elongated here, merely an
irregular lump.

The axone arises directly from the body of the cell. While
it is destined to ultimately be a component of either the V,
VII, IX, or X nerves, it generally takes quite an indirect
course. The usual way is through the mediun of the fasciculus
longitudinalis dorsalis. An axone passes into this bundle
to finally emerge at some other level. In Pig. 47, the axone
passes dorsal to the fasciculus communis. There is a bundle
of fibres of considerable size lying here, composed chiefly
of axones which have emerged from the dorsal longitudinal bundle
to enter the nerve-root. Other axones, but these are few in
number, may enter the nerve directly. Still others pass into
a bundle of arcuate fibres and doubtless enter the nerve on
the opposite side. These several paths to the nerves are shown
in Pig.?, v.m.f.

The viscero-motor neurones have an internal structure
which is conspicuously motor in character, the large amount

of tigroid substance representing much expenditure of energy
here. The large size of the neurone is evidently the correl-
ative of the iuiportiince which the system assumes in the inner-
vation of great vital organs. Such innervation requires a
nexus with viscero-sensory neurones, and the means have already
been suggested in the preceding subsection, the interlacing
of the axones from the lobus vagi with the dendrites of the
viscero-motor nucleus. 'This gives a complete reflex circuit
for visceral innervation.

6. Gen.eral Cutaieous h'ucleus and Spinal V Iract.
The literature of the portion of the oblongata included
here is in some confusion.. The older writers did not have
methods of investigation which would demonstrate the presence
in this region of several important groups of neurones. They
did see that it contains nunerous bundles of nerve-fibres,
however, some of which run to the cerebellum, and so peduncu-
lus cerebelli appeared to be a perfectly satisfactory designa-
tion. Kingsbury )'97) extended the term spinal 7th tract to
both the nuclei of the region and the trigeminal fibre-tract
proper; such use of a term originally intended to designate
a definite group of nerve-fibres is certainly to be avoided
as leading to confusion. Johnston ('99b) appears to include
a part of this region under his tuberculwr acusticum. Since
there appears to be need for a precise terci which shall desig-

45

nate the several nuclei of the general cutaneous system, I
therefore propose that the whole be called the general cutane-
ous nucleus.

The general cutaneous nucleus is the continuation into
the oblongata of the dorsal cornu of the spinal cord, and it
carries its associated tracts with it. Its position will be
seen frorr Fig.2,g.c.n. In structure, it is indeed complex,
embracing as it does, three groups of neurones, large numbers
of nerve-fibres intercrossing in several directions, the spinal
V tract, and many supporting elements. The intrinsic neurones
will be considered first.

a. The Molecular Layer. — The molecular layer appears in
a transverse section as a dorsal cap to the other constituents
of the nucleus (Fig. ?,m.l. ). It is seen to be continuous with
the cerebellar crest of the tuberculum acusticum.

The neurones of the molecular layer are of two varieties,
the molecular neurones and the neurones of Purkirje. Both of
these varieties are identical with those described in Subsec-
tion 7 for the tuberculum acusticum; their morphology will
therefore not be given here.

As to distribution, the minute molecular neurones are
found scattered through the whole thickness of the molecular
layer. The Purkinje neurones, on the contrary, lie only in
the deeper part of the layer, sending their great dendrites
into the upper levels. The possible phylogenetic significance
of the presence of the molecular layer in this part of the

46
brain will be treated in the theoretical considerations of
Subsection 7.

b. The Subs tan tiu Gelat inosa. — The gelatinous substance
of Rolando is continued frorri the spinal cord into the oblonga-
ta. It takes on such an intimate relation to the terminating
nerve-fibres here that Barker ('99) has proposed to call it the
nuclei tractus scinali^ nervi trigemini . The Folandic sub-
Stance occupies a position hi^h up in the dorsal part of the
general cutaneous nucleus, just beneath the cap of the rr.olec-
ular layer, and itself forming an investment for the bundles
of fibres and larger neurones of the central mass. The complex-
ity of its structure, and the difficulty of staining it with
the usual reagents are facts well known to microscopists. The
nerve-cells of Rolando's substance in the spinal cord have
been studied by vonLenhossek ('94) from man, the pig, and the
mouse; while Ramon y Cajal ( '96) has given us an elaborate
description of the oblongata of the mouse. It gives me great
pleasure to be able to extend our conceptions of this peculiar
formation to the field of the brain of selachians.

In Mustelus, the structure of the substantia gelatinosa
foreshadows to a remarkable degree the organisation which has
been described for the higher vertebrates. The neurones are
associated in groups; a representative collection is drawn in
Pig. 5. The axones and dendrites branch profusely and interlace
so closely that the simulacrum of a fine network is given. In

47

this tangle, certain nerve-fibres of the general cutaneous
system terminate. Such a fibre is shown in F'ig.5, g.c. f . ; its
arborization may be distinguished from the tangle of neurones
in which it lies by the somewhat greater size of its terminal
tw i gs .

The neurones of the substantia gelatinosa, considered
morphologically, are of three varieties, all of which are rep-
resented in fig. 5. The one that is present in greatest num-
bers, giving character to the formation as such, is of quite
minute size, and is an extreme example of Golgi's second type
(Pig. 5, a. The cell-body is very small, polygonal in form, and
there are a few short dendrites. The axone ramifies immedi-
ately into an extremely complex series of branches, constitu-
ting the principal member of the tangle mentioned in the pre-
ceding paragraph. These axones are strikingly conspicuous
structures when successfully impregnated with chrome-silver.

A second form of neurone is considerably larger in size,
its dendrites enter into the tangle referred to, instead of
its axone. From a polygonal cell-body, some three or four
dendrites radiate indifferently in all directions. The den-
drite is of fine calibre, and it branches repeatedly into ever
finer twigs. The branching of this system is far less profuse,
however, than that of the axone of the first variety described.
The difference will be readily seen by reference to F'ig.5,b.
The axone takes a course out of the gelatinous substance into
the deeper parts of the nucleus. It gives off a number of

48

long collatet-als duririfs ttiat part of its course lyin^ in the
substantia gelatinosa.

The two kinds of neurones just described evidently con-
stitute the physical basis for the central reception of gener-
al cutaneous impressions. The first variety, with its short
axone, is so concerned entirely, it would seem; while the sec-
ond one described, having its axone proceeding to deeper re-
gions, is probably involved to an equal degree in both the
reception of impressions and in their distribution to deeper
levels.

Still a third type of neurone remains to be described
(F'ig.-5,c.). From an oval cell-body, a few sparsely branching
dendrites arise which lead far out into the surrounding field.
The axone takes a course into the deeper nervous matter, giv-
ing off only a few collaterals. This form of neurone is found
outside the margins of the groups, and apparently is purely
associative in function.

c. The Deever Heurones. — The deeper parts of the gener-
al cutaneous nucleus are occupied by fibres having courses in
several different directions, and by the bundles of the spiral
V tract. Neurones are scattered at intervals between the nerve-
fibres, with a somewhat more closely crowded area just dorsal
to the lobus vagi. Neuroglia is especially abundant in all
parts of the nucleus, providing a support for the intricate
ma?e of nervous structures.

49

The external irorpholoi^y of a neurone from the middle re-
gion of the nucleus is shown in Pi^^.^. Such a neurone is con-
siderably larger in size than any neurone from the substantia
gelatinosa. Prom an elongated-oval or triangular cell-body,
three or four dendrites proceed straight outward into the tan-
gle of nerve-fibres and neuroglia. A dendrite is a rather
stout process. It has but few branches, and these are of small
size. Its surface bears scattering gemmules and minute knobs.

The axone takes its origin directly from the cell-body.
It runs in a medio-ventral direction.

Two neurones stained with methylen-tlue are represented
in Pig.4S. These are from the median collection just dorsal
to the lotus vagi. The nucleus of such a neurone is propor-
tionately quite large. It is also eccentric in its position,
consequently cells are often found in which the nucleus appears
to be in direct contact with the cytoplasmic pellicula at some
point. The nucleoli are of some prominence. The chromatic
material is also conspicuous. It is distributed in branching
strands of some thickness in places, and the several strands
may have slight connections.

The cytoplasm is often practically absent at the side of
the nucleus. The dendrites are so thick, relatively, at their
bases that the greater part of the cytoplasm appears collected
in them. The tigroid substance is most abundant in the vicin-
ity of the nucleus. The individual masses here are usually
irregular in form and quits small in size. In the basal parts

50

of the dendrites, the tit^roid-bodies become elongated. These
do not reach far alon^ the dendrites, however.

i. Termination of General Cutaneous fibres. — The nerve-
fibres of the general cutaneous system are distributed to the
skin of the head without the intervention of special] ^•ed nerve-
endings. This sensory system has components in the V, IX, and
X nerves. There are two principal central stations for these
fibres: the general cutaneous nucleus, and the dorsal cornu of
the spinal cord through the medium of the spiral V tract. The
two termini are really not essentially different, however,
since the general cutaneous nucleus and the dorsal cornu are
morphologically continuous structures.

Fibres which have their termination directly in the gen-
eral cutaneous nucleus may pass into their arborizations in
relation to each of several groups of neurones. What is prob-
ably the chief mode of termination is shown in Fig.5,g.c.f.,
where the fibre is seen entering the tangle of the substantia
gelatinosa. At no other point in !/ustelus is there such a
bewildering maze of nervous processes as that presented by the
neurones here. 3o far as structural features may be interpre-
ted, it would seem that the substantia gelatinosa is well adapt-
ed for the reception of the most delicate sensory impressions.
The large number of arborizations found here indicates the real
importance of the group.

Terminations are also to be traced in connection with the

51

deeper neurones of the nucleus. In Pi£.'^ there is shown an
incominf^ ner-ve-fibre, g.c.f., breaking up into its arborization
near one of these neurones. The arborization is of the broad-
ly digitate variety, spreading the disturbance over some slight
area. The scattered distribution of the neurones at this point
is doubtless a correlative of this fact.

Finally, the molecular layer which caps the whole nucleus
contains many fibres of quite minute size. This level, with
its small branching neurones, and the dendrites froni deeper
70nes, may serve to distribute impressions superficially.

Turning to the spinal V tract, this is a series of bundles
of fibres which run posteriorly to the spinal cord. The bun-
dles are scattered through the deeper part of the general cuta-
neous nucleus; refer to Fig. 48, sp.V. General cutaneous fibres
may enter the nucleus directly fron^ the nerve of the same side,
or through the mediuir of the arcuate bundles from the opposite
side. They may terminate at once in the nucleus, as described
above; or they may turn into the spinal V tract for termination
farther posteriorly. Some of those present in the tract are
doubtless branches of fibres which have termirated in part in
the nucleus. The spinal V tract, then, is a means for carry-
ing great numbers of sensory nerve-fibres from cranial nerves
to the dorsal cornu of the spin^sl cord, giving them a second
and far wider hold.

The manifold central terminations of the general cutane-
ous system must have a significance, if we can but interpret

52
the facts. The systeni is, prin.arily, a tactile apparatus for
the heaj. Any one who has watched Wustelus exploring with his
snout every corner of a new aquariuir cannot doutt that tactile
impressions from this region must have a large place in the
life of the animal. With practically no other check upon his
visual sensations than can be derived from poking his nose
into things, widely spread central terminations of cutaneous
fibres is no more than should be expected.

7. 7 he TubefouluTT. Acusticum.

The tuberculuir acusticum is the trigeminal lobe of Viault
('7'^:), Hohon ('77), ynd Sanders CS^); Kingsbury ('97) has
extended the teriii oerebel lar crest to the entire structure;
while Johnston ('98b) appears to include under his tubercului
aouaticuT. all those structures in the oblongata which are homol-
ogous with the dorsal cornu of the spinal cord. The term tri-
geminal lobe has been so variously used that it should be drop-
ped from our nomenclature. The earlier writers on the brain
of the selachi3.n so designated the tuberculuir acusticum, sup-
posing, erroneously so, that the great trigeminal complex,
which emerges just beneath its anterior end, must have the
origin of its nerve-roots here.

The tuberculum acusticum is the prinary terminal station
for the acustico-lateral system. Fibres from the lateral line
sense-organs and the internal ear are components of the VIl,

53

VII I , and X nerves.

The structure of the tubei-culuir, acusticurr, embraces a medio-
central core in which bundles of nerve-fibres predominate; and
an outer ir.vesting cap, the cerebellar crest ( Fitj. ?,cb.cr. ) .
The latter, in its outer levels, is morphologically continu-
ous with the molecular layer of the cerebellum, so that the
term TiOleoular lauep is very properly applied here. The neu-
rones of the tuberculum acusticum lie chiefly in the cerebel-
lar crest, with just a few, also, in the central core. Three
varieties are to be distinguished: Purkinje neurones, granular
neurones, and molecular neurones. I have so designated them
because each is the equivalent in every way of the element
bearing such a name from the cerebellujr.

a. Molecular and Granular Heurones. — The molecular and
the granular neurones do not call for any extended description
here, since they are directly comparable with the cerebellar
representatives described at length in Section V.

The molecular neurones are the most numerous of the three
varieties peculiar to the tuberculum acusticum. They are found
scattered throughout the molecular layer, most abundantly in
the deeper levels. Bach has the characteristically small cell-
body with slender dendrites radiating from it.

The granular neurones are far less numerous. They lie
just within the base of the molecular layer, but there are
not enough of them to form a distinct zone. Stained with

54

metbylen-blue, ttisy exhibit the typical dense nuclear struct-
ure and the small amount of cytoplasm.

b. Purhinje neurones. — The phylogenetic interest attach-
ing to the presence in the oblongata of these eleir.ents calls
for more special notice. The Purkinje neurones of the tuber-
culusr acusticum are not disposed in any regular order. They
lie scattered between the deeper molecular neurones, and they
are also found among the granular neurones. Their dendrites
are always directed toward toe limitans externa, whatever the
situation of the cell-body may be, although the exact course
may be at a considerable angle to the perpindicular. The axone
always passes inward. Refer to Big.?, t.a.

The cell-body is slightly oval or even perfectly circular
in outline (Fig. 7). The particular form assumed appears to be
correlated with the position of the dendrites. The gnarled
character of a dendrite is quite evident. There are three or
four of them, terminating in blunt tips. A dendrite bears few
branches, or none at all, and its surface is studded with numer-
ous thorny gemmules of various sizes, some of them quite large.

The axone arises from the internal border of the cell.
It takes a ventral course through the intervening nervous tissue
into a bundle of arcuate fibres. Its ultimate distribution
has not been definitely followed.

The internal structure presents a large nucleus with deep-
ly staining chromatin in the form of a reticulutr; and cytoplasm

55
somewhat less in quantity than in the cerebellar neurone, and
holding less tigroid substance.

Proni the above description of the morpholok^y of the neu-
rone of Purkinje from the tuberculum acusticuni, it will be
evident that it is strikingly like its earlier-known represen-
tative in the cerebellum. Compare, also, Pi^.7 with Fit^.l^.
The neurone from the tuberculum acusticum (Bif:.7) is merely
slightly smaller in size, with a dendritic top which branches
somewhat less. The presence of conspicuous gemmules is char-
acteristic of both, 'f.e shall take occasion to point cut fur-
ther on that the cerebellar neurone is to be regarded as the
more fully developed derivative of this element from the ob-
longata.

0. Terrdnation of Acustico-Lateral Fibres. — The fibres
of the acustico-lateral system enter the tuberculum acusticum
through the medium of the thin neck joining the acusticum with
the general cutaneous nucleus below. This fact is expressed
in Pig.?, a.l.f. It appears, so far as they are traceable,
that these fibres have passed over from the opposite side in
the arcuate bundles. Penetrating the core of the acusticum,
the fibres are at first associated in groups, but these soon
become dissolved and the individual members spread through the
outer levels. The fine terminal branches pass through the
thorny dendritic tops of the neurones of Purkinje, and the
final arborisations are to be frequently observed near the

56

bodies of the Purkinje cells. Such a termination is shown in
Piff.?, a.l.f. The nerve-f ibre divides at some distance t'roir.
the cell into several twites, and from the portions of those
nearest the cell-boiy several small terminal twi£;S arise which
form an arborization quite near to the cell. There is given
in this way a direct riexus between the acustico-lateral fibres
and the most conspicuous neurone of the tubercalum acasticum.
Other fibres of iiinute siJ^e evidently terminate in the molec-
ular layer between the small molecular aeurones. E^ven here
tnere would be an indirect connection with the Purkinje neu-
rones, for these send their dendrites into the molecular layer.
Fibres are also to be seen in numbers which do not termi-
nate in the tuberculum scusticun at all. Some of these reciain
in the central core, continuing anteriorly; while ethers are
branches of fibres which have taken a course into the acusti-
cum for partial termination there. All of these pass upward
to the cerebellum; they will be duly considered in Section V.

i. Theoretical Coiolusions . — It has already been pointed
out that acustico-lateral components are not present as such
in spinal nerves. Waiving the question of the origin of the
system as a whole, it will net be out of place to consider
here certain problems relative to its primary centre, the tu-
berculum acusticum.

The position which the tuberculum acusticum occupies in
the oblongata is inde'id significant, superposed, as it is.

57
on th5 crani^il r^presantati.v's of the dorsal oornu, the ggnernl
cutansoiis aucleus. This fact caa only be taken to mean that
the tuberculum aoustioim is phylogeneticalLy the younger of
the two straotares in question. In Subsection ^ it was shown
that the general cutaneous nucleus has the archaic structural
features of the dorsal cornu cappei with a molecular layer
continuous /?ith the cerebellar crest of the acusticum; and that
there are present here both the molecular cells and the neu-
rones of Purkinje. ^!oa' this structural continiiity iiay signi-
fy that the acusticun has been derived fron the general cuta-
neous nucleus in the phylogeny of the vertebrate nervous sys-
tem. Of course it flill be necessary to have a thorough study
of the embryology of this region in the lower vertebrates before
such a conclusion can receive unqualified support. Then, too,
the rise of the tuberculum acusticum has been but one feature
in the evolution of the lateral line system as a whole, and
so tie may confidently expect some assistance from the further
study of the affinities of its sense-organs.

There is somewhat more solid ground for the belief that
the tuberculum acusticum has itself given origin to a very
important region of the brain, the cerebellum. The full ev-
idence on this point ^vill be presented in Section \', Subsec-
tion *5.

58
5. Supporting Kleme'itu,

a. ^peidi^mj. — " The spenlyma of ths obloni^ata constitutes
the menabrana It-nitms i^iterna throut^Jiout, of course, bat the
degree of iif fereatiation covers a /»iJe raa^.e. The simplest
condition is found in th^ tela choroidea, for tne cells here
io not give origin to ependyoial fibres. This might be callei
the epithelial phase of develop^nent.

The first step in the differentiation of epeniyma occurs
in the dorso-lateral region of the oblongata, embracing the
tuberculuni acusticum anl the adjacent part of the general cuta-
neous nucleus. The ependymal cells are rather narrow here,
giving a larger number per surface area. F'rom the tip of each
cell, a very sleniar epeniymal fibre pushes its way through
the nervous fnatter to the limitans externa. The fibre does
not branch, and its general coarse is a straight one, although
the details of its path are very irregular indeed, sharp turns
of small size occurring along its whole length. While the
dia-neter of the fibre as a whole is quite slender, its size
is increased at irregular intervals by bea3s and snail knobs.
Consult H'ig.S.

A slightly more advanced condition is to noted for the
ependyma of the ventral region. The ependymal cells in the
floor of the fourth ventricle are larger than those described
above, and their fibres are also much stouter. The general
course of these fibres is shown in Kg.S.ep.; and the details
of two fibres are given in Fig. 9. It is to be noted that the

fi9
fibre extanls eatirely to the limitaas exterrin, thnt it benra
no branches, th'it its course is frre frooi sharp tur-is, airJ that
its diameter is variei by nodes ani thickenin>5S of various
shapes ani sizes, Groups of these fibres forn radiating bands
through the fornaatio reticularis, and their evident purpose is
to provide stays for the vast number of longitudinal and arcu-
ate nerve-fibres .vhich characterize this part of the oblongata.

The epenlyma attains its highest differentiation in the
lobus vagi; see Pig. IT. The fibres do not reach the surface,
and they give origin to a most complicatei series of branches
at the level occupied by the neurones of the lobus. The prox-
imal part of the fibre is quite stout and some.vhat irregular
in its course. The distal part becomes dissolve! in branches,
the terminal t^igs of which are minute. The interlacing branch-
es forn a support for the neurones anl nerve-fibres proper to
this region.

b. Xeur-ogUa. — The neuroglia of the oblongata is, like
the epenlyma, present, in great diversity of form, but the va-
rious elements are referable to two general classes. One, the
well-known astrocyte, is distribatei in every part of the ob-
longata where thare are ner/e-fibres in numbers — in the roots
of nerves, in special tracts, in centres where many fibres have
their teruination, and in the formatio reticularis. The special
form of the astrocyte depenis, to some extent at least, upon
the densit/ with which the nerve-fibres are arrangel. At points

«hsr3 there has been little pressure, the rounJei form of tne
cell-boiy iiay be retainei (Pig.3,n^.). In certain closely
packed tracts, the cell is squeeze! into lenticular forrn. In
centres .vhere many oiin'ite fibres interlace, there may be har-i-
ly any body to the cell at alL, the matter bein^ chiefly in
the radiating processes (Pig. 11). In every instance, the proc-
esses exhibit tortuous irregularities due to their insinuation
between the nerve-fibres among vhich they lie.

Another type of neurogliar cell appears to be character-
istic of the ventral part of the oblongata, occurring in some
numbers in the formatio reticularis (Pig. 3). The cell-body is
large, elongate!, .viti great processes eoierging from its ends
and a fe« smaller ones from its sides. The large processes
take a course in the radius of the oblongata. They give ori-
gin to secondary and tertiary branches, the finer ones of ^vhich
lie at right angles to the coarse of the principal branch.
The flhole system does not spread vfidely, but has the appearance
of compression in one plane.

The astrocyte is evidently an important supporting ele-
ment for the iniividual nerve-fibres between which its proc-
osses twine. The second class of neurogliar cell described
seems to be aiaptei to the supporting of bundles of nerve-
fibres. The arcuate fibres of the formatio reticularis pass
in groups through the spreading brush of such neurogliar proc-
esses and are thereby given support.

61

.9. Summary of the Oblongata.

The strnctaril features of tha oblont^itd of Wustslus may
be horaologizei with those of the spinal ooni to an unusual
legree of cleanass.

The ventral cornu is representei by the nucleus of the
VT nerve, and by the scatterei commissural and tract-neurones
of the formatio reticularis.

The lobas vagi is, morphologically, the centre of the
viscero-sensory system, but there has been annexed to it a
complex of peripheral sense-organs. It has come to be, there-
fore, the centre for the entire communis system, components
of the VII, IX, and X nerves. Sndings are found related to
neurones of the Golgi TI type. These send tneir axones toward
the viscero-motor nucleus. Fewer fibres of the system enter
the fasciculus communis than in either the teleosts or the
amphibians.

The viscero-motor nucleus is the representative of the
paracentral nucleus. It is a morphologically continuous col-
umn of cells giving origin to the motor fibres of the V, VII,
rx, and X nerves innervating viscera. The axones do not enter
their nerve-roots directly, as a rule, but through the medium
of the fasciculus longitudinalis dorsalis. Some appear to
cross to the opposite side. Dendrites from the nucleus come
into relation ^ith axones of the lobus vagi.

The general cutaneous nucleus is the homologue of the
dorsal cornu. It is tne primary centra for the general cuta-

62
neons components of the V, IX, anJ X necves. The isubstantii
gelitinosa contains ^,roups of small Golgi 11 neurones forrning
dense, felt-like tangles; many terninil arborizations are found
in these. Other terminations occur near larger neurones lying
in the deeper parts of the nucleus. Many fibres of the system
do not terminate in the nucleus, but are carried posteriorly
as the spinal V tract for ultimate termination in the spinal
CO r i .

The tuberculum acusticum is a phylogenetioally young struct-
ure, not deriv/able from the cord directly. It is the primary
centre for the acustico-latsral system, components of the VII,
VIII, and X ner^^es. There are present, neurones of the molec-
ular, granular, and Purkinje types, identical ^ith those of
the cerebellum. Terminations occur in the molecular layer,
and near the neurones of Purkinje. The Purkinje neurones send
their axones ventrally in the arcuate bundles. The cerebellar
crest is morphologically continuous -vith the cerebellum. There
is ground for believing that further investigation «ill derive
the acusticum from dorsal cornu structures.

Supporting elements are present in the oblongata in great
numbers. There is traceable a wide range of developmental
forms of both neuroglia an3 epenlyma.

Section V,
The Cerebellom.

The cer.^bsllaii of Mustelas is of large size in conaparison
rtith the adjiicsnt brain-ssgmsats. In the aJalt animal, it is
of sufficient longitaiinal extension to overhang the larger
part of the midbrain in front, ani also rnuuh of the oblongata
behini (?ig.l,cb.). The base is only a third as great, howev-
or, indicating the smaller prototype from /»hich the organ has
been evolvel. In fact, this cerebellar occupies an interne-
iiate position in the phylogenetic scale, standing midflay be-
tween the simple plate-like cerebellar of the cyclostOTie, the
dipnoan, or amphibian, and the solid mass flith radiating lam-
inae characteristic of the mammal. It is essentially a great
bulbous dilatation of the dorsal side of the neural tube, the
;»all of ^hich has been thrown into folds as the process of
growth thrust the vesicle against the unyielding and more slow-
ly expanding cranium. So, instead of a solid central mass of
nerve-fibres coverei with layers of gray matter, we find in
Mustelus a hollow organ, the fourth ventricle extending free-
ly into it and ramifying through its several folds. The folds,

therefore, are simply doublings in the cerebellar wall, and

(63)

64

are fania-nea tally liffersnt from the solii Laminae familiar to
us ia higher vertebrates.

The arrangement of the strictupal elements of the cere-
bellum is readily folLo^ei, for in no other part of the brain
are iistinctions more clearly marked. Superficially, there is
the molecular layer ( Big. 18, m. 1 . ); the granular layer lies
next to the ventricle, or is separated from it only by the
basal fibres (Pig. 12, g.l. ) ; between the t/fo, there is the layer
of Purkinje (ftig.l?,p.l. ).

The nerve-fibres of the organ are found in two well-defin-
ed groups, the basal fibres and the median fibres. The basal
fibres (Fig. 18,b. f . ) enter the posterior region from the ob-
longata and soon become dissolved in the outer layers. These
fibres lie next to the ventricle, and they supplant the gran-
ular layer entirely during their course, or at least displace
it quite largely. The median fibres are disposed in scattered
bundles which lie in the outer part of the granular layer just
beneath the cells of Purkinje. The bundles take two general
directions: transverse, (cut across in Plg.l2,m. fC ); and longi-
tudinal, extending parallel witn the str.ictural zones (Pig. 18,
m.f.).

1. The Neurones of Purkinje.
The neurones of Purkinje are the largest, and by far the
most impressive strictiral elements of the cerebellum. Their

65
celL-bodies are iispose-i in a thin stratun interoalatei bet»feen
the molecular and the granalar layers. At tnose points vihere
th3 granular layer is absent, the Parkinje cells come into
iirect contact with the layer of basal nerv^e-fibres, instead..
As to superficial distribition, the neurones follow a well-
dofined rule (Pig. 12, p. 1. ) . They are most numeroas on the
sides of a cerebellar fold, often several cells in depth here,
and not infrequently having their cell-bodies in contact. At
the summit of the fold, there are wider spaces between the
individual cells; and for a small space at the bottom of a
fold, the cells are absent altogether. The posterior fold of
the cerebellum has a part of its area without Purkinje neurones.

Pig. 13 illustrates the features of external morphology
characteristic of a neurone of Purkinje. The cell-body is
situated, as already noted, at the base of the molecular layer.
The dendrites, therefore, grow upward into the molecular lay-
er, and to their presence here is due, in part, the marked
striation perpindicular to the surface which is so character-
istic of this superficial zone.

The axone always arises from the base of the cell-body,
and it pursues a horizontal course for some distance. It then
turns downward through the granular layer to enter the system
of fibres leaving the cerebellum. Where the granular layer is
absent, the axone may be traced into the layer of basal fibres
for a still greater horizontal course. The axone is remark-
able for the fact that it does not give off collateral branches.

The cell-boJy app^nrs to have it.? form detamined quit?
largely by the lu-nber and iisposLtioa of the denlrites arising
from it. It may be roanded, oval, or even triangular in out-
line. The lon^^er axis of the cell, vfhile usually perpiniica-
lar to the surface of the fold, may be oblique or still farther
tilted over from the orientation characteristic of it for the
higher vertebrates. 6ven the size of the cell is influenced
by its position. The dianeter is greatest in those cells lying
at the part of the fold where the side curves abruptly into
the summit. On the sides of the fold, the size is remarkably
less, and the dendrites pass outward at a wider angle, often
causing the cells here to assume a horizontally elongated form.
The cell studied by Szczawinska ( '96) evidently vfas of this
latter type.

The internal structure of the Purkinje cell is demonstra-
ted less readily with methylen-bliie than is the case for most
nerve-cells of Mustelus. In successful preparations, however,
the cytoplasm is found to hold tigroid masses of triangular or
spindle-shaped form. These bodies are neither large in size
nor many in number. There are always several of the greatest
diameter arranged near to and concentric with the nuclear mem-
brane. The thick bases of the dendrites have small, narrow,
lenticular tigroids distributed sparingly as far as the level
of tne first great branches. The character and distribution
of the tigroid-bodies will be seen by reference to fiig.49.

The nucleus of the cell lies in a basal position, but it

67

is al.tHyg bDriarial by sonae litbls thickness of cytoplasm. The
forii is oval or ciroalnr, /lith an alfnost evin contour. Thef?
is one large nucleolus, *ith a subsiilary nucleolus in some
calls. The nuclear reticuluTi has a rather fine mesh; lar^e
granules of chromatin occur at the points of intersection.

The denirites -arise from the peripheral siJe of the cell-
body, and they pass at once into the molecular layer, taking a
direct course for the limitans externa. The size of a denlrite
is maintained without narkel diminution, the tip, in fact,
being so large and blunt as to give a club-like appearance to
the 'Whole process. The mode of branching is far simpler than
that of the corresponding mammalian neurone. While the den-
drite of the latter ramifies to an extraordinary degree, it is
rare to fini branching carriei beyond the tertiary divisions
in Mustelus; compare my Fig. 13 with Plates 14 and lo of Starr's
Atlas ('9^). Prom its greater simplicity, the selachian neu-
rone of Purkinje corresponds, in a broad way, to the embryonic
condition of the higher form. The branches are spread in one
plane, like a plant trainel on a wall, and this plane lies
transversely to the cerebellar folds. The course of the branch-
es in this plane is less wide than for mammals, causing the
whole top to present a markedly more compact appearance. The
surface of a denlrite has the conspicuous roughness charfticter-
istic of this neurone wherever found. The gemraules, to which
this roughened surface is due, are of several different sizes,
ranging through mere points through slender spines anJ stouter

68
thDf'is to knob;? and mush room-like excrescences which rise far
^bove the general level. The extent of receptive surface of
the entire deniritic series is thus increasei to a very con-
si:ierable degree.

The part «hich the neurone of Purkinje takes in the econ-
omy of Mustelus will be discussei in Subsection 5.

2. The Holeoular Layef.

This is the external layer of the cerebellum (^ig. 1?, m.l. ),
It conforms throughout to the contour of the several folds
into which the cerebellar wall is thrown. Its tnickness is
carried flith considerable uniformity, but in most of the re-
gions the depth of matter is only about half as much as that
of the granular layer within.

The molecular layer is characterized by the predominance
of fibrous elements and the fewness of nerve-cells. In con-
trast with the mass of densely-packed, conspicuous cells of the
granular layer, the outer region presented hardly more than a
minutely punctate appearance to the earlier investigators who
had recourse to nothing more differential than the general
stains in use at that time. Hence moleaular seemed an appro-
priate descriptive term for this layer.

Through the use of modern methods, we find the molecular
layer to have a few proper neurones of small size, but the
great mass of substance consists of fibrous material, between

69

tha parts of which the ner^e-celLs are intercalatei. But, it
is certainly a fact worthy of njentiin, the fibrous constit-
aents of the layer do aot have their origin there, penetrating
it, rather, froin deeper levels. The dendrites of the Purkinje
neurones, and the processes from the neurogliar cells comprise
one great class of constituents. These take a course perpin-
dicalar to the surface and cause the vertical striation /»hich
is so luarked a feature of the region. Then, too, the neurones
of the granular layer send their axones outvjard into the rnolec-
ular layer for a T-shapei division, each thus giving origin to
a pair of fibres. These branches take a course across the
sagittal plane of the cereb9llu-n, parallel at once with the
limitans externa and the lateral surface of a fold. To the
presence of these fibres, cut across in such numbers in a
sagittal section of the cerebellun, the characteristically
punctate appearance of the molecular layer is chiefly due.
Finally, it should also be noted, there are numerous termina-
tions in the molecular layer of nerve-fibres which have entered
the cerebellum from some other region of the nervous system.
Such terminating fibres branch so as to distribute the endings
over a considerable superficial are^.

It will thus be seen that the molecular layer is really
a tangle of nervous tissue, a series of paths where many asso-
ciations may be formed. This topic will be discussed more at
length in the fifth subsection.

The neurones proper to the molecular layer are, as already

70

not-a^i, but fe* in number, an;.i thoss prssant are scittered be-
tifssn ths nerve-fibres. The externnl iiorpholo^y of one of
these neurones is representel in Fig. 14. Prorn a small, almost
perfectly spherical cell-body, three or four ienJrites raiiate,
branching in what is an approach to a dichotomous nanner. The
dendrites are not thick at their bases and they become lost
to view before proceeding far, owing to the fineness of the
terninal t.*ig3. The surface of a dendrite is almost perfectly
smooth, there being only the faintest indication of gemmules,
but there are small varicosities at irregular intervals which
proluce slight variations in the thickness.

The internal organization of the cell presents a large,
rounded nucleus .vhich is enveloped by a stratum of cytoplasm
of no great thickness (fig, 50). The cytoplasm contains ti-
groid substance distributed in granules of the most minute size,
mere points even when highly magnified. The chromatin is dis-
tributed along a linin reticulum of such fine mesh that the
nucleus is thereby often made to appear almost perfectly black.
There is a single nucleolus.

In mammalian neurology, the molecular layer is known to
have two varieties of cells: (1) Stellate cells, the process-
es of which lie freely in the layer; and (8) basket cells,
somewhat larger in size, occupying the deeper levels of the
layer, anJ having their axones associated with each other in
such a way as to form plexuses or baskets aroun3 the cell-
bodies of the Purkinje neurones. The latter type of cell is.

71
of coursa, the mar's spscialized of the t«o. It is therefore
interestiag to note th'it it is not representaJ in Mustelus,
bat thit all of the Tiolecul^ir cells correspond, rather, to
the stellate cells of higher vertebrates. Such a result is,
however, to be expected in a brain of lower phylogenetic value.

5. The Granular Layer.

The granular layer lies interaal to both the inolecular
layer and tne neurones of Purkinje. It is sorDe^vhat irregular
in its distribution. At many places it is t«ice as thick as
the molecular layer, notably at the summits of the great folds;
while it te-ids to decrease in extent as the bottom of a fold
is reached. A sagittal section of the entire cerebellum shows
a few regions where the granular layer is absent altogetb.er
(F'ig.l8,g.l.).

With nuclear stains, the granular layer appears to con-
sist of a vast number of densely packed, rounded nuclei, from
which fact the names granular- and nuclear have been appliel
as descriptive terms. It is only through the application of
metallic impregnation that the real character of the elements
anl the relations between them have been determined. When thus
demonstrated, there are to be recognized neurones of two dis-
tinct varieties, the granular, proper, and the Golgi neurones.

a. The Granular neurones Proper. — Nearly all of the

72

neurones comprising the granular l-iyer ars inclnied in this
class. The distinctive fe>itares of such a neurone (F'i^.l'^)
embrdce a roundel cell-body having a few short dendrites, and
an axone vrhich ascends through the levels of the granular lay-
er above its point of origin to the molecular layer, where it
divides in a T-like manner.

The size of the cell-body varies slightly, but it is al-
vtays smaller than a neurone of Purkinje. As to shape, there
is considerable diversity. The derivative form appears to be
a rounded one, but this has been subjectei to much modification
by the origin of the dendrites so that triangular, rectangular,
and varioas polygonal outlines are given.

The internal structure of the cell, (F'i?.51), consists
chiefly of the nucleus, only the faintest halo of cytoplasm
being visible at any point; even the bases of the dendrites
can hardly be recognized .vith purely cytological stains. Dem-
onstrated '.Tith either methylen-blue or iron haematoxylin, the
nucleus is found to contain a few very coarse and irregular
chromatin granules strung on fine interlacing threads of linin.
The presence of a nucleolus is doubtful, at least it is dif-
ficult to distinguish one from the masses of chromatin. Fig.
ol also shows how closely these neurones are packed.

The dendrites are three or four, only, in number, aris-
ing from the cell-body at approximately equidistant points
(?ig.ln). A dendrite is short anl nearly always relatively
stout. Its course involves sinuous curves. Tt branches bat

73

r-iraly until the tip is reached, whsrj it dissolves into a
brash or a series of hooks spreading, over a relatively consid-
erable area. There are no ^emmules.

The axone is exceedingly slender. It may arise directly
from the cell-body, but it usually takes its origin from one
of the dendrites, either near the base or at some distance
from the cell; Pig. It illustrates the two modes of formation.
The course of the axone is invariably peripheral, pushing through
the intervening thickness of the granular layer into the molec-
ular layer. At the boundary between the t^fo layers, the course
changes abruptly to a horizontal one for a short distance, and
hence the entire course of an axone can rarely be traced in
one section. In the molecular layer, the axone divides into
two branches which, with the original stem, form a T-shaped
figure (F'ig.f ). The two branches parsje a course parallel
with the surface of the cerebellum and the sides of the fold
in which they run. It is thus seen that they pass through the
dendritic tops of the neurones of Purkinje, comparable to tel-
ephone wires passing through the tops of the trees along a
highway.

b. Golji Veurones. — A few neurones of the granular layer
have an altogether different character from the ones just iescri-
bei. These lie in the upper levels of the layer. Such a neu-
rone is shown in ''ig.l?. The cell-body is always a little
larger than that of the typical granule neurone, and its form

74

is nore roandei, A faw club-liks dendrites fiJiate from it
for a short distance, branchim? but sparsely. The size of a
dendrite is increased at intervals by slif5ht s/»elLinf?s.

The axone passes downward into the deeper levels of the
granular layer, instead of ap.vard. It gives off collateral
branches soon after its origin, and it breaks ap into a num-
ber of fine ter-ninal twigs before any great distance has been
traversed.

This neurone is homologous vvith the variety described by
Golgi ('94) frorn the human cerebellum, and by him made a repr^
sentative of his second tyoe of nerve-cell. The branching of
the axone is far less profuse, hoi^ever, than in the mammal.

We thus see that the granular layer of the cerebellum of
Mustelus is marked by the presence of the same varieties of
neurones /fhich characterize this layer in higher vertebrates.
The morphology is somewhat more simple in the selachian, as
should be expected, but it is a suggestive fact that the
structural plan of higher forms is here outlined in its essen-
tial features.

4. Supporting Elements.
The supporting elements of the cerebellum are referable
to both the ependymal and the neurogliar series, the former
being limited to the granular layer, and the latter to the
molecular layer. The structure in each instance appears to

75

be particalarly adaptei to the supportinf? of the nerv^oas mech-
anisms peculiar to these regions.

0. The Ependyma. — Tne epeniyna of the cerebellim presents
the usual palisaia of closely crowded cells forming the membra-
na limitans interna (Fig. 13). The cell-body is irregularly
pyramidal in form, the sides rarely tapering evenly to the
apex but exhibiting more or less bold cur/atures of outline.

From the apex of the cell-body, a process, the ependymal
fibre, arises, and this pursues a course through the struct-
ures of the granular layer to near the outer limit of that
zone. I have not found a single instance where one of these
fibres passel beyond the granular into the molecular layer,
a fact of some possible phylogenetic value, indicating the
more ancient character of the internal region.

The ependyma-f ibre is relatively stout, but its diameter
is far from uniform. There are fibres which have portions of
the length four or five times the thickness of the intervening
parts; see Pig. 18, b. Occasionally, knobs and other rounded
thickenings are found, particalarly at angles 'where the course
of the fibre changes abruptly. The trenJ of the fibre is never
conspicuously irregular, only such slight turns and windings
occurring as might be expected where obstructions are present
during the period of growth. The general trend is directly
towarl the surface of the cerebellar fold, the several fibres
lying more or less nearly parallel with each other.

76

The iegrae of br-inching exhibits v»iis diversity. Certain
fibres (Pig. 18, a) have some half-dozen principal branches of
various lengths, none of them very long nor diverging widely
from the main stem; besiles the principal branches, there are
shorter twigs distribated sparingly along both the main stem
and its ramifications. Other fibres (Pig. 18, c) are so beset
with a multitude of small twigs, quite irregular in their branch-
ing, that the whole series is given much the appearance of a
long and cylindrical brash. Betwsen these two extremes, there
are every intermediate condition of branching forms.

The ependymal fibres lie quite near to each other, and the
many processes from them constitute an interlacing tangle, the
profusion and extreme delicacy of which cannot be adequately
represented in any drawing. In order to appreciate the sig-
nificance of this dense supporting framework, the vast number
of the granular neurones must be recalled. Here we fini the
means by which these nervous elements are given that support
which is one of the primary conditions for their activity.

b. h'aupoglia. — The neuroglia provides a support for the
outer structures of a cerebellar fold, just as the ependyma
functions within. The characteristics of neuro^liar elements
are shown in F'ig.l9,ast and bg.f. The cell-body lies at the
juncture of the molecular and granular layers, between the
cells of Purkinje. Two conditions are to be distinguished.
One variety (Pig. 19,ast) has a cell-body of quite irregular

77

Dutlina from /»hich many processss r-iiiate. These processes
raraly branch. Some of them may proceed to a distance equal
to several times the greater axis of the cell, but most of the
branches are far shorter. Thsy are placed so closely along
the margin of the cell that the whole has something the effect
of a halo. These cells are clearly homologous with the astro-
cytes of higher vertebrates.

The other type of neuroglia cell (Pig.l9,bg. f . ) is refer-
able to the category of Bergmann's fibres of the mammalian
cerebellira. The cell-body has fewer processes than the astro-
cyte, but it gives origin to one stout fibre which takes a
peripheral course, without branching, through the entire thick-
ness of the molecular layer. The path of the fibre is not one
directly toward the surface, for it runs almost parallel with
the layer of Purkinje neurones for a short distance; it then
turns upward, terminating at the limitans externa in a conic-
al expansion. During the proximal part of its length, it bears
fine processes similar to those emerging from the cell-body.
The distal part of the fibre has a remarkably vast number of
fine processes. These may remain separate from each other,
but at intervals they become matted together so closely that
the whole has the appearance of felt or even that of a solii.

The account given by Schaper ('93) of what he took to be
Bergmann's fibres does not correspond with my findings in sev-
eral particulars. This author does not mention the processes
of the cell-body, nor do his figures show them. It may be that

78

his prspar-itions were insufficiently impregnated. Weii^ht is
given to sach a possibility from the fact thnt he dii not find
astrocytes at all. It is certainly of some phylogenstic inter-
est that Bergmann's fibres appear to be derived from astrocytes.
I have found numerous instances where transitional forms are
recognizable, linking the extremes of the simple astrocyte
flith the one provided with a long process, the fibre of Berg-
mann.

The position occupied by the cell-body of a Bergmann's
fibre is also significant. Kolliker ('9^,0.3^-3) states that
at birth Bergmann's fibres in mammals lie at the boundary of
the granular and molecular layers, and that during growth they
normally migrate into the granular layer. Now the permanent
position of Bergmann's fibre in Mustelus corresponds to the
embryonic state in the mammal. An additional comparison may
be instituted regarding the form of the fibre, the adult ele-
ment of Mustelus remaining simple, while the mammalian fibre
becomes much branched. That the condition in Vustelus corres-
ponds to the embryonic stage of development in higher verte-
brates is, of course, no more than should be expected.

The physiological interpretation of the cerebellar neu-
roglia is not difficult. The numerous processes from both the
astrocytes and Bergmann's cells provide a delicate suspensory
apparatus for the neurones of Purkinje. The great Bergmann's
fibres, reaching upward as they do through the molecular layer,
may be likened to so many telegraph poles; and the matted tan-

79
gles exteniinp; later^illy from these ara the cross-bars for the
wirss. The sires to be supported, following out our compar-
ison, ara the horizontal axones of the granular neurones, which,
as *fe have already noted, extend through the molecular layer
in large numbers.

5. Arshi tea twe and Phijsioloqu of the Cerebellum.

It is now a veil established fact that the principal func-
tion of the cerabsllJTi is to preside over the equilibration of
the body. Moreover, a fairly direct connection is traceable
between the morphology of this segment of the brain and the
nature of the muscalar activities -vhich are characteristic of
the animal. To the student of comparative neurology, there-
fore, the cerebellum holds problems of a special order, and
is ever potential with no mean interest.

We have seen that the cerebellim of Wustelus retains an
external form of a far lower order than that found in the bird
or mammal. The internal organization is not so inferior, howev-
er, as we might, from this fact, be led to infer. The archi-
tectural features of the lower and the higher types are so
similar that the contrast is really one of degree and not one
of kind. In both instances, the same sorts of neurones are
present, grouped in a similar way, and related to each other
physiologically in essentially the same connections. This
remarkable identity of features can only be interpreted to mean

so

that ths cerebellum bec^ime establishe'i in its ors^inization
quite early in the history of vertebrates.

Undoubtedly, the key to the significaace of cerebellar
structure and physiology is to be looked for in the neurones
of PurkiTje. These striking cells, with their characteristic
tree-like tops, apparently have all of the other structural
elements present arranged contributory to them. Imbedded in
the midst of the nervous matter, supported by the interlacing
processes of the neuroglia there, a Purkinje neurone sends its
great branching dendrites outward into a veritable maze of
possible physiological connections, for such the molecular
layer really proves to be. We have seen that the granular
neurones contribute their axones to a series of fibres passing
horizontally through a row of spreading Purkinje dendrites;
that the neurones of the molecular layer are themselves radi-
ately connecting paths; and, finally, that the incoming nerve-
fibres, those arising outside the cerebellum, end, some in the
granular layer, others taking a longer course into the molec-
ular layer. These several elements evidently have no other
purpose than the bringing of all incoming impressions, through
one path or another, to bear upon the neurones of Purkinje.
The great spreading top of a Purkinje cell is obviously a de-
vice for providing a large receptive surface for such impres-
sions, thile the many thorn-like gemmules with which the den-
drites are stadded serve as a greater extension still of that
surface, or, perhaps, make one which is more readily impressed.

81
The axone of the cell takes a more or less direct coarse oat
of the cereballurn, carry inf5 the resaltant of the nervous in-
teractioTS to the proper point for altimate distribution. If
it is the purpose of the whole series of cerebellar elements
to provide a central mechanisn of equilibration, then the neu-
rone of Purkinje is certainly the centre of that mechanism.
The various nerve-fibres sweeping into the cerebellum may ter-
minate in several ways and at diverse levels of the organ, but
everywhere there are devices for connectinf; them physiologic-
ally with the neurones of Purkinje, which receive all and pre-
side over all.

The researches of Lee ('92, '93, '94, '98) upon equilibration
in fishes have shown with what nice discrimination these aquat-
ic animals balance themselves, devoid, as they are, of many
sources of impressions possessed by terrestrial vertebrates.
Doubtless a large number of equilibrial impressions are de-
rived from the visual mechanism of the fish. Axones from the
roof-nucleus of the midbrain pass backward into the cerebellum;
and it will be shown in Section VI that there is a most inti-
mate association between the roof-nucleus and optic termina-
tions. Another source of equilibrial impressions is to be
found in the fins and body musculature, entering the cerebellum
through the tractus cerebello-spinalis. But the work of Les
clearly demonstrates the overshadowing importance of the pis-
cine ear as a peripheral organ of equilibration. In fact, his
latest research ('9B) makes it clear that the great ear of

82
Mustelas is not a true auiitory organ at all, but thit its
several parts are to be interpretei from the standpoint of the
equilibrium sense. The large nerve-tracts froro the ear to the
cerebellum are, therefore, iefinitely sii^nif icant, bein(5 the
connecting fibres between the central mechanism and its most
important peripheral organ. It also appears that not only
does the ontogeny of the ear show its relationship to the lat-
eral line organs, but the functions in the two instances are
comparable as well.

An analysis of the habits of Viustelus will, it appears to
aie, go far towari explaining the disproportionately great devel-
opment of both the ear ^ni the cerebellum which we find the
animal to have. This shark, although a comparatively small
representative of the group, is, withal, a restless hunter of
the seas, ever urged onward by an appetite which, apparently,
has no bounds. Continually suspended in a fluid medium, and
compelled to balance itself at every turn, the animal requires
a precise mechanism of equilibration. This is to be found, in
the main, in both an ear and a cerebellum developed to a degree
out of all proportion to the scale occupied by the creature
as a whole.

6. Evolution of the Cerebel lum.
6vidence of the origin of one brain-structare from another
is usually of tne most meager value, but it appears as though

S3
it v»9r8 no'."» possible to wsavs to^sthsr a fe« scHtterjJ threais
ia the 3\/olntiori of the cerebellum. Schaper ('94) has callel
attention to certain facts in the srnbryolofiy of the teleostean
cerebellum which nay be taken as the starting point. The cere-
bellurn arises in ontogeny, not as a brain-segment having the
fall value of the others, bat as a paired thickening in the
parietal walls of the neural tube at the anterior end of the
oblongata. These thickenings grow upward and mest, each other
in the nedian line. Schaper has since extended his studies
to all classes of vertebrates, and he finds ( '99) that the an-
teriDr limit of the bilaterally symmetrical anlage may be def-
initely fixed, coinciding with the boundary between the pri-
mary mid- and hiadbriin vesicles. The posterior limit of the
future cerebellum is by no means clear, however, for it seems
to merge backward into the oblongata. These are certainly
significant facts.

It was shown in Section IV, Subsection 7, that the molec-
ular layer of the cerebella-n is continuous with the cerebellar
crest of the tuberculum acusticum, maintaining essentially the
same morphological characters throughout. It was also shown
that there are present in the taberculam acusticum neurones
which are identical with those of the cerebellum — molecular
neurones, granular neurones, and Purkinje neurones; the last
two varieties are not grouped into definite strata, however.
The presence of granular neurones in the acusticum is worthy
of remark, but the greatest weight must be attached to the

34

finiiaf^ of Purkinje neurones here. Tnese neurones nan lonii.
supposed to charicterize the ceraballani alone, anJ their stri-
kingly peculiar appearance makes them readily identified. The
neurones of Purkinje from the acusticum of Mustelus agree with
those from the cerebellum as to size, general form, shape and
character of dendritic top, and even in the presence of the
spiny gemmules so characteristic of this nervous element. Ther.;
can be no question as to their morphological identity in 'viuste-
lus. Johnston ('98b) has found Purkinje neurones in the acus-
ticum of Acipenser, somewhat smaller and simpler than those of
the cerebellum, but undoubtedly equivalent. Similar results
will probably be obtained in all of the simpler vertebrates.

It may be concluded, in the light of these embryological
and structural facts, that the cerebellum has arisen in the
phylogeny of vertebrates as a fused outgrowth of the pair of
tubercula acustica. The acusticum is the primary end-station,
as ie have seen, for the nerves of the ear and the lateral
line organs. The cerebellum has been differentiated from the
primary ending as a special centre for presiding over equili-
bration. Parallel with the increasing development of the equi-
librial sense in vertebrates, the cerebellum has Gradually
acquired associations with other than the original source, so
that the fibres entering the organ have ever been growing more
numerous, and the bulk of the fibrous centre consequently more
massive.

95
7. Summary of the Cerebel I im.

The cersbellarn of Must^lus is of relatively lari^e size,
and, although remainiaf? in the primitively hollo« condition,
its wall has the same essential plan of structure as the organ
in higher vertebrates.

Neurones of Purkinje form a zone, crowded at certain points,
between the nolecular and the granular layers. Such a neurone
has the stractural features characteristic of its mammalian
representative, but with a simpler branching of dendrites. Its
office is to receive incoming equilibratory impressions.

The neurones of the molecular layer are few in number.
They are all of one variety.

The neurones of the granular layer are of Vto kinds: Golgi
neurones and granular neurones. Tne former lie in the upper
strata of the layer. The granular neurones are strikingly
like those of higher vertebrates; their axis-cylinders branch
in a T-shaped manner in the molecular layer, mediating between
many incoming fibres and the dendrites of the neurones of Pur-
kinje.

EJpendyma is developed into a profusely branching fibre
which extends through the granular layer, only.

Neuroglia provides a supporting framework for the neurones
of Purkinje and the molecular layer. Both astrocytes and Berg-
mann's fibres may be recognized, but there are transitional
forms indicating that the latter have been derived from the
former.

86

The large size of both the ear and the cerebellurn of Mas-
telus is to be explained by the swimming habits of the animal.
The structures have an equilibr^tory value.

The cerebellum has apparently arisen in the phylogeny of
vertebrates as a fusel outgrowth of the pair of tubercula acus-
tica, a specialization of that part of the oblongata forming
the original terminal station for the acustico-lateral system.

Section VI.
The Midbrain.

The miibr-iin of Mustslas has retained to a marked degree
the featipes characteristic of this cerebral vesicle in its
primitive condition. Its ventricle has been so little encroach-
ed apon by nervous matter that the name aqueiuat of Syl'jius
■^oali not be applied as a descriptive term were it not made
necessary by jsage in higher vertebrates. The aqueduct is
produced laterally into a pair of spacious recesses, each of
which occupies the interior of an optic lobe, while the general
cavity communicates freely through narrower extensions with
the third ventricle in front and the fourth ventricle behind
(Fig. l,mb. ). The optic lobes are two thin-walled bodies of
dome-like form, separated from each other by a conspicuous
median furrow, and the pair taken together are somewhat broad-
er than the base of the midbrain on which they rest. The ante-
rior divisions of the cerebellum lie upon and partially con-
ceal the optic lobes from the dorsal aspect. The base of the
miJbrain is a direct continuation anteriorly of the great fibre-
systems of the oblongata and metencephalon, showing but slight
diminution in size. -^ .

88

The microscopicil anatomy of the raiibraia is really quite
coTiplicatei, owin^ to the diversity of the tenninations and
connections which are established here. The base is a cro/»d-
ed hi'§h<«ay between the different parts of the sncephalon. The
fibres of the optic nerve have their termination in the dorsal
midbrain, the tectam -nesencephali, with a nexus of intrinsic
neurones and fibres of more distant origin associated with the'
Finally, the aqueduct of Sylvias is bordered by nervous matter
which is phylogenetically distinct from either the base or the
tectam, and this has undergone special development at certain
points as the nuclei of important motor neurones.

1. The Teotum Vesenoephal i .

The tectum mesencephali lies as an investing cap upon the
central gray matter which surrounds the aqueduct of Sylvius,
Its situation here is fraught with significance, for it repre-
sents an addition to the more ancient nervous structires made
necessary by the development of lateral eyes in the vertebrate
phylum. The tectum embraces the central expansions and asso-
ciated connections of the nerve-fibres having their origin as
axones of the retinal neurones. Certain neurones of the tec-
tum may, also, send their axones outward into the optic nerve.
The whole complex is, in fi-ie, the primitive visual centre.

The tectum of Mustelus does not approach the extreme de-
gree of differentiation which ftamo'n y Cajal ('8?), '91) has de-

89
scribed from the optic lobes of biris, but it is pricticable
to jistiaguish three zones of structural elements: the layers
of the superficial, the millLe, and the deeper neurones, res-
pectively, (Pig.?l,3.n. ,m. n. ,d. n. ). VanGehuchten ('94) has
described three zones frona the optic lobe of the teleost, but
these are not exactly equivalent to the layers noted here,
his conohe moleanlaire appears to include both my superficial
and middle layers; his couohe granul ease corresponds in posi-
tion to any deeper layer; /vhile his ooaohe ies oell-iles eoeniy-
maires is an inner zone .vhich apparently does not include the
central gray matter.

a. TerTiiiation. of the Opticus. S tr . '.feiul I are Pro fund um. — A
great bundle of fibres may be traced dorso-posteriorly from each
angle of the chiasma to the optic lobe, flhere it becomes dis-
solved through spreading over the surface. A section sho-vs
that the fibres in the outer zone lie parallel with the sur-
face during some little part of the length of their course, and
that they then pass do.vn/iard into deeper parts of the tectum.
Fine branches are given off in the regions occupied by the
neurones of the middle and deeper layers, and many terminations
appear to occup here. A certain number of fibres, however,
pass to ever deeper levels, tending to become collected into
bunlles, and they finally blend into tne stratum medullare
prof undum.

The stratum meiullare profundum is a conspicuous feature

90

of the deepest part of the tectum mesencephali . At the crown
of the arch which the optic lobe presents in section, the fi-
bres are seen as gre^it horizontal bundles interrupted at al-
most perfectly regular intervals by small groups of vertical
fibres. The two lateral halves are united by a strong commis-
sure just dorsal to the aqueduct of Sylvius. Pig. ?0, s. m. p. ,
represents these features in a transverse section.

Traced laterally, the fibres of the stratam meduUare pro-
funda™ are found to pass into the base of the midbrain in tvo
somewhat clearly marked divisions. The outer division takes
a large number to the lateral surface (Pig. ?0,e.l. ); while a
markedly smaller group passes downward nearer the median line
for a ventral decussation (Pig. 80, i. 1. ). Tne inner division
is composed of those fibres lying nearer to the central gray
niatter from the outset. Some of these have but a short course
downward and out^rari, but the great mass of fibres continues
near the median line as a series of intercrossing bundles which
are destined to decussate ventral to the aqueduct of Sylvius
(Pig.20.e.m.,i.m. ).

The fibre-system composing the stratam medullare ppofun-
dura must be indeed important in the economy of the brain. Into
this system we have traced fibres from the cord and oblongata,
fibres from the optic nerve, and from the neurones of the tec-
tum mesencephali itself, 'fie will subsequently have occasion
to point out that a great tract sweeps into it from the inter-
brain as a relay in the olfactory apparatus. Pibres are also

91

present here from certain of the cranial nerves. All of these
fibre-systems from iiverse sources are to become relatei to
the remarkable mechanism of Reissner's fibre. It is certainly
evident that there are here every means for intercommunication
between different parts of the nervous system, a switch-boari,
so to speak, of extraordinary possibilities; but to this sub-
ject »e shall return further on.

b. The Suoerfisial Vewones. — The outermost layer of
the tecturn mesencephali is characterized by the predominance
of nerve-fibres and by the feeble development of its neurones.
Receiving as it does the fibres of the optic nerve for their
first expansion, we should hardly expect a hi^h degree of ner-
vous activity here. There are a few scattering nerve-cells
present, however, (Pig. 81, s. a. ) . These are quite minute in
size, lenticular in form, and are disposed with their long
axes tangential to the surface of the brain, as though squeezed
into crevices betv»een the mass of nerve-fibres. The internal
organization of the cell presents no features which would mark
it as having any degree of importance physiologically (Fig. 5?,
s. n. ).

c. The Hiidle Veurones. — The neurones which lie in the
middle layer of the tectum are characterized by their larger
size and vastly greater numbers as compared with the outer
region. In fact, the number is so great that in a section.

98

the n3r\^3-col L;=5 oft^ii ^ippenr superp^saJ one over -inother, ani
th9 processv'?.-? :D=ike a varitnble t'^nAle of ifiterl^icini^, branches
(?ig.21.in.n. ).

The nearoaas of the level are to be distinguished from
those of the deeper layer, on the other hand, both by their
compactness and by their mode of branching, v?hich is of the
radiatini^ type. The several processes of a nerve-cell spread
oat anl branch freely in all directions, bat they do not ex-
teni far away from their points of ori>|ifi. A representative
forn is dra^vn in Pig. 22. There are always several dendrites,
and these may arise either from the outer end or from the side
of the polygonal oell-body. The branching begins quite near
the origin, so that the size diminishes rapidly from the base
outward. The dendrites of the outer extremity have their finest
twigs penetrating the superficial layer of nerve-fibres. The
dendrites arising from the sides of a cell interlace more large-
ly with those of other neurones. The surface of a dendrite
always bears an abundance of simple gemraules.

The axone pursues an irregular coarse, often spreading
over a considerable horizontal area, but the general treni is
ever toward the center of the brain. It gives off a great
profusion of collaterals as it proceeds, an3 the final terni-
nation is found at no great Ustance from the cell-body. There
is here, then, an illustration of a cell of the Golgi II type.
The several ramifying axones lie in the region occupied by the
cells of the deeper layer.

93

Th8 nucleus is a rouaiei, centr^lLy locHtei boiy, filling
the larger part of th? mass of the cell. The cytoplasii holis
a few tii^roid-boiies of small size, an eviieace of a loi« oP'ier
of metabolio activity in this type of cell. Pig. 52, m. n, , exhib-
its what letails of internal organization are \^isible under
the highest amplification.

The part which the neurones of this layer take in the
physiology of the optic ter:nination may be inferrei witn some
degree of certainty. The dense tangle of dendrites just be-
neath the incoming optic fibres constitute a large surface for
purposes of reception. The spreading axones in the layer be-
neath afford, with the neurones there, a physiological nexus
of some superficial extent, possible paths of association, if
ne choose to apply the term here. 'We will returi to this sub-
ject in the following subsection.

i. The Deeper Veurones.-- This layer is thicker than
both the preceding ones put together. The neurones are less
closely crowded than those of the middle layer, and they lie
in groups between the bundles of optic fibres passing downward
to the stratum meduUare profundum (Pig. 21,d. n. ).

The neurones which give character to this layer are long-
drawn-out, the dendrites extending nearly to the external sur-
face of the brain, and the axone reaching well toward the lim-
itans interna. Pig. 83 shows a typical neurone considerably
enlarged. The cell-boiy is spindle-shapei or oval in outline.

94

F'roTi its outer anJ, a single stout liealrite proceels straight
to(»ard the periphery. This dendrite branches bat sparsely,
and no branches are given off for some distance beyond the
point of its origin. The several branches pass oatward through
the niddle layer of neurones, and the most delicate twigs can
be traced into the superficial layer of nerve-fibres. The
'Thole deniritic series presents a top both tall and narro/r,
in strong contrast to the for-n assurnei by the neurone of tne
middle layer. Seramules are scattered over the branches, but
they are conspicuous neitaer for their size nor their numbers.

There may be other minor dendrites arising from near the
base of the cell, as sho/»n in Pig. 23; but the axone always takes
origin from a point opposite the principal dendrite. The axone
is directei toward tne limitans interna. It may be traced for
some distance without any very marked dimination of size. Wany
collateral branches are given off in its course, and the ulti-
mate destination of the principal stem appears to be a bundle
of fibres contributing to the stratum meiuUare profundum.

The body of the cell is so largely filled with the nucle-
us that often only the thinnest investing film of cytoplasm
is exhibited. Six cells from this layer have their structural
features shown in Pig.o3, and one of these has the thinnest
enveloping cytoplasm found in this class. The nucleus of a
cell is regularly oval in contour, holding one prominent nucle-
olus. The chromatic network is delicate and of small mesh.

The cytoplasm contains masses of tigroid substance which.

95
in proportion to the -ictuHl magnit.iie of the cell, are rela-
tively lar^e ia size, 'f^hen the nucleus lies to«arl one enJ of
the cell, the lari^est ti^roids are found in the free extreuity.
The prevailing form of tigroid mass is triangular, and t.ie
several masses have a rather open arrangement.

F'rom the form of the neurone just describe I, we should
infer that it is adapted to the part of a conducting mediuL
bettveen the superficial and the deeper levels of the tectum
lesencephali. The internal structure also indicates that it
is characterized by no slight amount of metabolic activity.
Evidently, then, there is here a nervous element of consider-
able importance in the economy of the midbrain.

VanGehuchten ( '94) nas described the deeper neurones from
the optic lobe of the trout (his conohe gra^rj.leuse) as being
flithout dendrites, and as sending their axones peripherally
into the superficial levels. The character of this type of
neurone in Mustelus certainly corresponds the more nearly with
the conditions fDund in higher vertebrates, although much sim-
pler of course, so that, in this respect at least, the sela-
chian is seen to be in the direct phylogenetic line, while the
teleost is divergent.

In the inner zone of the deeper layer of the tectum, at the
level occupied by the numerous collaterals from the proper
axones, there are to be found at intervals neurones of another
character (Pig. 21). Such neurones are irregularly stellate as
to general forn, the dendrites anl the axone radiating widely

96
frorn a polygoa^l call-body. The office of this neurone may be
to bring the collateral branches of the other ani more numer-
ous nery^oas elements into relationship. Pig. 54 shows a group
of five of these cells, an unusually large number to find in so
limitei a fieli.

2. The Central r;rau Hatter.
The aqueiuct of Sylvius is surrouniel by a layer or nerve-
cells and nerve-fibres representing the most archaic part of the
brain, and quite properly termed the central gray matter (Pigs.
20 and 21, c.g.ii.). The greater part of this nervous matter
has been supplanted or overshado/fel functionally by more recent
aiditions upon its outer surface, but certain nail defined
groups of neurones have retained their pristine importance
through the character of their ultimate connections. These
are, respectively, the roof-nucleus, and the nuclei of the
oculomotorius and the trochlearis.

a. The Roof-Huol eu3 . -- This is a collection of very large
neurones lying in the roof of the aqueduct of Sylvius between
the ependyma anl the dorsal commissure of the stratum medullare
profundum. The group has a considerable longitudinal exten-
sion, reaching from the anterior end of the optic lobes to near
the juncture with the cerebellu-n. It is broken into two later-
al halves by the median line. Pig,2'),r.n. exhibits the entire

97

nucleus; flhils Pi^.5o represent?; the dlstpibution of the cells
more in letail ->n the right siie.

These neurones are the largest of any in the nervous or-
ganization of Mustelas. The size attained by the cell-body
may be as great as 60 micra in transverse diameter by 100 mi-
cra lengthwise. The group, therefore, presents a very stri-
king picture in the field of the aiicroscope. The forois assuni-
ei by the celL-bodies are someAfhat diverse, ranging from ovyil,
through irregular outlines to a considerably elongated condi-
tion. The longer axis lies parallel, or nearly so, with the
limitans interna.

In a thick section, the dendrites appear as tv»o or three
stout processes vrhich push their /vay into the nervous matter
dorsal to tne group and are soon lost to view. Those cells
lying very close to the median plane send their axones to the
opposite side; but the remaining members of the group, compri-
sing nearly all of the cells, have their axones extending away
from the mid-line. The axones from the several cells of the
same side, together v»ith the crossed axones, run laterally for
a greater or less distance, tarn anteriorly, and become associ-
ated into bundles which constitute a fairly well-marked tract
(Pigs. 20 and 55, r.n.t.). This tract extends forwarl to the
anterior limit of the midbrain, where it unites with its fello*
from the opposite side, and tne united group of fibres emerges
from the midbrain roof to penetrate the aqueduct of Sylvius
as the fibre of Seissner. The ultimate destination will be

98
traced ia a subsequent paragraph.

Those cells of the roof-nucleus lying in the posterior
region have a differsnt ternination for their axones from the
one just described. In tnis instance, the axones pass poste-
riorly, instead of anteriorly, and they take a course into the
cerebellaii. The significance of this fact has been consider-
ed in Section V, Subsection 5.

The cell-nucleus is a large, evenly rounded body, almost
invariably eccentric in its position, sometimes, even, lying
in i»hat appears to be a special protrusion of the general cell-
mass. The chromatin is distributed in the forte of a reticulum
of rather fine mesh which holds coarser granules at intervals.
The nucleolus is evenly rounded and of conspicuous size. Many
of the cells have t.vo or even more nucleoli.

The cytoplasm, stained .vith methylen-blue, exhibits a
minutely punctate appearance even unler the highest amplifica-
tion, due, chiefly, to the minute size and diffuse distribu-
tion of the tigroid substance. The tigroids are quite densely
packed in the peripheral regions of the cell. In the field of
the nucleus, the prevailing forn of granule is rounded; farther
a/?ay, the shape is a more elongated one, the long axis being
tangential to the margin of the cell. Pig. 56 sho<«s the details
of cell-organization, Nissl staining.

When these cells are stained with iron haematoxylin, they
exhibit what appears to be the equivalent of the perinuclear
reticulum of Golgi (1900). We fini the internal part of the

99

cytoplasm exhibiting deeply stained, massive bodies, branching,
and anastomosing, ^ith each other through more slenier connec-
tions. The several individual masses are disposed in such a
way as to give the appearance of enclosing the nucleus as with
an open net.vork. The bodies fade away as they enter the den-
drites, and there is no appearance of their having communica-
tion with the exterior such as has been described by Holmgren
('99a, '99b). Pig. 57 represents the appearance of this series
of structures.

It is doubtless necessary to avait further researches in
many distantly related fields before we attempt to pass final
judgment as to the significance of tne perinuclear reticulum,
bat the hypothesis noted by Golgi (1900) is one flhich certain-
ly deserves our consideration. The appearance presented by
the network may be caused, not by solii bodies at all, but by
a series of coramunicating canaliculi filled with a fluid which
is deeply colore! by certain stains. Such a reticular canal-
system woull probably take no part in the irritable life of
the cell as such, but would function on a lower plane of pure-
ly vegetative character.

The neurones of the roof-nucleus come into intimate rela-
tions witn the nerve-fibres of the stratum medullare profun-
dum. The dorsal decussation between the opposite halves of
the stratum carries a strong bundle of fibres across the me-
dian plane immeiiately above this group of neurones (?'ig.55,
dc.s.m.o, ) There are to be found here numerous instances of

100

nerv3-t'ibres smerging frotn the general bunils and tenoiaating
in arborizations nenr the bodies of the nerve-cell^. Pig.'i?,
ar. , represents t*»o such arborizations near the same cell.
The character of the teruination is exceeiingly interesting.
The axones are founJ to present a reticalo-vesicular structure
throughout their whole length, the protoplasm apparently con-
sisting of vesicles of several degrees of size united by a
reticulum. No^ as the axone approaches its ternination, the
reticulation becomes naore pronounced, and tne final arboriza-
tion is seen to be essentially an expansion of the same thing.
The ending is simply a widely-spread, digitate reticulum. There
are many thorn-like branches frox all of the strands, and nu-
meroiis anastomoses occur tetfleen the principal ones. Consult
H'ig.57,ar.

The remarkable group of nerve-cells comprising the roof-
nucleus has been variously interpreted, but its true relations
were not discovered until quite recently, ftohon ('77) first
described this collection of cells as the iaohkeme, a name
applied, of course, from the position occupied by its paired
members in the roof of the aqueduct of Sylvius. It was later
recognized in the brains of various fishes, Burckhardt ( '9?)
identifying it as the midbrain trigeminal nucleus. It has
remained for Sargent (1900) to show that not only is the roof-
nucleus present in all vertebrates, but that it is part of a
most interesting mechanism, the fibre of fteissner ('*^0).
Reissner's fibre is a rod-like body lying in the central canal

101
of the spinal corl, exteniini^ t'or^ari. It h'ii come to be ne-
glected entirely in recent years o/»inf5 to the prevalent view
that it naerely represents a coagulation of the cerebrospinal
fluid. Sargent (1900) demonstratei that such a vie^r is erro-
neous, that Reissner's fibre is a real structure, with a per-
fectly definite character and distribution, and, f urthemore,
that it is found in all classes of vertebrates, always extend-
ing from the postsrior end of the canalis centralis to the an-
terior end of the aqueduct of Sylvius. He nas also sho^a (1901)
that the fibre represents, in the main, the closely fused ax-
ones of the cells of the roof-nucleus. The great number of
neurones comprising the anterior field of the roof-nucleus
send their axones into the aqueduct of Sylvius, as noted a-
bove, whence they pass backward as Heissner's fibre through
the extent of the fourth ventricle and the central canal of
the spinal cord. Pine processes are given off to the nervous
matter of the cord as the fibre proceeds.

An interpretation of tne roof-nucleus and of Seissner's
fibre arising from it may now be attempted. In the next to the
last paragraph, a nexus was traced between the fibres of the stra-
tum medullare profundum and the neurones of the roof-nucleus.
There are here, it is evident, the elements of a tract through
which quite direct connections may be established between the
somatic motor neurones of the spinal cord and certain sensory
impressions, visual and olfactory, at least. The classes of
impressions noted are carried through the stratum melullare

10?

profunium to the tarninal arboriz-itions which »e hive ieacri-
bad Qsnr the cells of the roof-aacleas. The aeurones of the
roof-nucleus transmit such impressions through their axones,
Seissner'3 fibre, directly to motor neurones at the several
levels of the spinal oori. The apparatjs of the roof-nucleus
and Reissner's fibre, regaried as a thing apart, is advantage-
ous to V.ustelas because it is a path without relay, a short
arc for motor reflexes betvfesn the eye and the olfactory organ,
on the one hand, and the body musculature on the other. Tne
giant size of the neurones of the roof-nucleus is doubtless the
correlative of not only their long axones but of their importance
in the econony of the aninal, as /veil. In ascending the scale of
the vertebrate series, however, it .vill be found that this mech-
anism ever takes on a progressively less and less value, owing
to the development of other means for attaining the same end.

b. The Huoleus of the Oculomotor ius . — The nucleus of
the III nerve lies ventral to the aqueduct of Sylvius (Pig.
53,n.iri). This collection of neurones is sharply marked off
from all surrouniing nervous elements by the large size of its
cells and by the greater intensity of stain ^ith lethylen-blue.
The group extends antero-posteriorly for some distance.

The prevailing form of cell is oval, with two or three
marked extensions of outline produced by the broaUy triangu-
lar bases of the lenirites (Fig. 59). The axone is more slen-

103
der by f-ir th'in the ienlrites; it is diractel away from the
limitans intenn, taking the coarse of a s/ieepinfJ curve.

The nucleus of the cell is evenly roanJed in forn, rather
large as to proportionate size, and it has a central location.
There is but a single nucleolus. The chromatin is iistribited
in a delicate network, the few interlacing strands visible
being of great tenuity.

The cytoplasm is remarkable for the large size of its
masses of tigroid substance, these bodies being visible as
distinct things even under low amplification. The form of a
tigroid is almost invariably triangular. The base of the tri-
angle lies toward the nucleus, the apex pointing towari one of
the dendrites. Those masses lying in contact with the nuclear
membrane are somewhat broader, taking the form of a so-called
nuclear cap. The size of mass decreases toward the periphery
of the cell, those lying in the bases of the dendrites assu-
ming a slender form. Reference may be made to Pig. 59.

The striking size attained by the masses of tigroid sub-
stance here is doubtless associated with the purely motor func-
tion of the III nevvB, the fibres of which are the axones of
these particular neurones.

0. The Vuoleus of the Troahlearia. — Hohon ('77) fell into
a curious error with regaris the nucleus of the TV nerve. As is
well known, the root of the trochlearis passes backward and
crosses over to the opposite side, appearing dorsally in the

104

furrow beti«33n the miibPiin anJ the cer.3b3llntn. hohon evident-
ly sought for the nucleus of the nerv^e near its superficial
origin, for he identified as such a group of cells on the bor-
der of the cerebellum.

The group of neurones constituting the nucleus of the IV
nerve lies posterior and slightly ventral to the nucleus of
the Til nerve (?ig.53,n.IV) . The anterior end of the trochlear
nucleus overlaps the posterior end of the oculonotor nucleus
for a short distance.

The cells of this collection are, as compared v»ith the
cells of the oculomotor nucleus, decidedly smaller in size,
and the general outline is more nearly triangular. Tne nucle-
us of the cell is relatively larger in proportion to the amount
of cytoplasm. The chromatic network is so delicate as to be
but faintly visible even under high magnification.

The masses of tigroid substance are few in number, rel-
atively large in size, and wholly irregular as to form. There
is a perinuclear zone of cytoplasm entirely free from tigroids.
it often appears as though many of tne tigroid masses are act-
ually clinging to the limiting pellicula of the cell. Pig.'^O
exhibits a condition typical for the cells of this group.

I really am unable to offer any explanation concerning
the marked differences to be observed between the tigroid-
bodies of the trochlear and the oculomotor neurones, respect-
ively. The contrast in both the form anJ the arrangement of
the tigroids is quite evident. The distinction is a real one.

105

not iue to variation in the action of reai?,ents, for I have
several instances showing the marked contrast in one and the
saoie section. A structural difference here is a noteworthy
fact, since the tv»o neurones are both of the somatic motor
type, entirely equivalent morphologically.

5. The Epenit^ma.

The general character of the ependynoa of the midbrain is
represented in Pig.?l, ep. , while Pig. 61 illustrates the de-
tails of cell-organization. The outline of a representative
cell is somewhat lance-like, tne length four times the breadth,
the pointed extremity touching the ventricle and the greatest
breadth at a point further removed. The interior of the cell
is occupied almost entirely by the nucleus. It is really dif-
ficult to detect any cytoplasm at all except in a small area
at the base of the ependymal fibre. The observer has the im-
pression forced upon him that most of the cytoplasm during
the course of growth has passed over into the cell-process.
The strjcture of the nucleus is reticulated to an almost ex-
treme degree. Some strands of the reticulirn are relatively
coarse, but many of them are so tenuous as to lie almost beyond
the capacity of the microscope.

Tne ependymal fibre takes a course which is almost straight
during the first part of its length, but toward the outer limit
of the central gray matter crooked turns occur, and branches

106

are givea off. Ths diamet-ar of the process remains unifor'n
throut^hoat except for slight swelliai^s v?hich occur at inter-
vals. The fibre terminates at the periphery of the brain.

The fact that the ependynal fibre becomes irregular and
branches only after it leaves the central gray matter for the
ne-ver additions outside may have a phylogenetic significance,
indicating that at one time the process had no farther course
than the outer limit of this most ancient nerve-substance.
But it is also entirely possible that the phenomena in ques-
tion are /without such deep significance, having been caused
by the greater number of obstacles in the path of the fibre
as it dre^ through the outer levels.

4. Phylogeny of Miibrain Structures.

The midbrain has ever been a stable part of the neural
tube. Marked out early in ontogeny from the other brain-seg-
ments, the midbrain of Mustelus retains many features of organ-
ization which are really primitive in character.

The central gray matter is the most archaic of the mid-
brain structures, and the newer additions of the outer levels
are derivable from it. The central gray matter is to be com-
pared, both as to general functions and morphological topog-
raphy, to the gray matter of the spinal cori before the latter
has developed its specialized cornua. In a broad way, the
ventral region of each is motor, and the dorsal part a series

107

of sensory centres. The homology is most readily tr^iceible
in the ventr-il region. The neurones of the ILL and IV ner«/e3
are true somatic nootor neurones, corresponding entirely to
those of the ventral cornua of the cori.

The dorsal region of the midbrain has become more and
more specialized as an optic termination. At a phylogenetically
early period, the optic fibres grew backward from their orig-
inal relatioTS to establish terminations here. Probably the
most primitive connection is the one .vith the giant neurones
constituting the apparatus of the roof-nucleus and Heissner's
fibre. Through this means, the optic neurones were chained
directly to the somatic motor neurones innervating the body
musculature. Later, the midbrain roof became thickened by the
wandering oat'vari of neurones from the central gray matter,
and by the development of new optic terminations associated
with them. Thus has arisen the tectum mesencephali, an end-
station which has remained important in the vertebrate series
as a visual centre until secondary connections were establish-
ed with the pallium. Hence the tectum is of great magnitude
in the lower vertebrates, where the pallium is weak, but be-
comes dwarfed in the mammalian brain in which there is an over-
shadowing development of pallial connections.

5. Summarif of the Miibrain.
The tectum mesencephali of Mustelus receives practically

103

all of the optic fibres, apparently only collateral branches
being given to the interbrain. Three structural zones of the
tectum are to be recognized: the superficial, the middle, and
the deeper zones, respectively. The superficial layer has
chiefly fibres, »ith a favv minate, tangentially elongated neu-
rones. The middle layer is composed of a densely crovided tan-
gle of neurones of the Golgi II type, the axones of which spread
laterally. The deeper layer has neurones which send long den-
drites iTto the outer levels, while their axones penetrate tne
stratum medullare profundum. Optic terminations occur in all
of these layers. The structure of the deeper layer places the
selachian more nearly in the direct phylogenetic line than the
teleost.

Tne stratum medullare profundum receives optic fibres,
axones from the tectum, fibres of the olfactory mechanism from
relays in the thalamus, as well as fibres from posterior re-
gions.

The central gray matter has become differentiated at cer-
tain points to form the roof-nucleus, and the nuclei of the
III and IV ner/es, respectively.

The roof-nucleus is a collection of very large neurones
lying dorsal to the aqueduct of Sylvius, the axones of which
form, ultimately, the fibre of Reissner. Terminations of fi-
bres from the stratum medullare profundum occur near the neu-
rones of the roof-nucleus. The roof-nucleus and Reissner's
fibre constitute a direct path for motor reflexes between certain

109
senses and the body musculatare, involving the somatic motor
neurones of the spinal oori. The senses thus me.liatei are,
prinarily, the olfactory and the visual, but the acustico-
lateral and the general cutaneous systems may be represented
also.

The neurones forming the nuclei of the III and IV nerves
have the structure of the somatic motor neurones pertaining
to the spinal cord and the oblongata.

Spendymal fibres extend through the entire thickness of
the midbrain wall, branching but feebly.

Section VII.
The Intshbrai'J.

The research of Edingsr, Das Zwisahenhifn, ('98), presents
an account of tne fibre-tracts of the interbrain of Selachii
and Amphibia as ienaonstratel by the Wei^ert method. It remains
for lie to add to the results of that »?ork a description of the
morphology of the neurones proper to the interbrain of Mustelus

1. The Thalamus.
The thalami of a selachian are so small in proportion
to the other parts of the brain (Pig.l.th.) that certain of
the older anatomists were thereby caused to overlook the inter-
brain entirely; see Section II. The small size of the thal-
amus is, I find, the expression of a low degree of organiza-
tion. Several investigations have made it clear that the thal-
amus of a mammal has several well-defined thalamic nuclei. Tne
thalamus of Mustelus, however, has remained in a condition of
such primitive simplicity that it is not practicable to insti-
tute very strict comparisons between its neurones and those
which are characteristic of higher forms. Before such compar-
isons can be of much value, a study must be made of thalami

(110)

Ill

having interms-liate degrees of development.

It has seemed to me advisable to distinguish but two col-
lections of neurones in the thalamus of Wustelus. One group
represents a differentiation of the ancient central gray matter;
this I have designated the nucleus strati grisei. The other
collection is certainly the one from which the several genic-
ulate nuclei of higher vertebrates have been derived; this I
have called the nucleus geniculatum.

a. The nucleus Sti'ati Grisei. — As has just been mention-
ed, this collection of neurones represents a differentiation
of the primitive central gray matter. The nucleus strati gri-
sei has retained its original situation next the third ventri-
cle (Big.84,n. s. g. ). It forms a broad zone just within the
limitans interna, comprising something like one-fourth the
thickness of the entire thalamus.

The neurones of this group are the largest of the thal-
amus. The cell-bo iy has a polygonal form, the several diam-
eters not greatly unequal. The dendrites radiate freely in
all directions, but they are not very long. The nucleus of
the cell has an sccentric position, causing the cytoplasm to
appear massed on the side from which the chief dendrites arise.
The chromatic substance is disposed in a ten thin strands having
thickened nodes. The entire amount of chromatin is not great,
and 30 the nucleus presents a lightly stained appearance. The
tigroii substance is limited almost entirely to that part of the

112
cytoplas'n havi'i^ the Greatest mass. Some of the ti>J,roi'i3 are
altogether irregular in form and are relatively quite lart5e.
Pig. 62 exhibits t/io neurones as they lie in place.

The nucleus strati grisei is the terminal station for
those axones of the tractus strio-thalamicus having their or-
igin in the general striaturn. Tnese sweep into the nucleus
in bundles, and their terminations are to be noted bet/»e3n the
constituent neurones (Pig. 24, f . s. t. ) .

The neurones of the nucleus strati grisei are, primarily,
a relay in the olf acto-motor chain. The tractus strio-thala-
micus terminating here is one of the links of that chain, as we
shall point out in detail under Section VIII. The axones from
the cells of the nucleus strati grisei pass backward into the
base of the midbrain as the tractus thalamo-tectalis, and then
sweep upwari into the tectum to lie in the stratum medullare
profandum. Here they are associated with other sensory nerve-
fibres, as already noted in Section VI, and the entire group
becomes related to the remarkable motor conducting path provi-
ded by the cells of the roof-nucleus and the fibre of Reissner.

It is certainly not worth while, with the knowledge which
we have at present, to attempt an extensive comparison of the
nucleus strati grisei with the specialized thalamic nuclei of
higher vertebrates. It seems fairly safe, however, to regard
the nucleus rotandus and the nucleus magnocellularis as de-
scendents of this simple collection of cells founi in the thal-
amus of selachians.

113

b. The Nucleus Genidul atwn. — This nucleus is imbeidei
in the substance of the thalamus lateral to the nucleus striti
grisei. It is separated from the neurones of that group and
from the external surface of the brain by bundles of fibre-
tracts. In transverse section, this collection of neurones
appears as a broad band curving parallel with the li-nitans
externa (Pig. 84, n.gen.).

The size of a neurone from the nucleus geniculatum is
considerably less than that of one from the nucleus strati
grisei, and the form is of the elongated instead of the radia-
ting type. Fibres from the opticus leave that nerve to form a
terminal zone on the periphery of the thalamus, and the den-
drites of these neurones extend out'fari into this zone, while
their axones take a course inward. The neurone, therefore,
comes to be drawn out in a direction approximately at right
angles to the limitans externa (Fig. 84, n.gen.). The cell-
body is rendered somewhat elongated by the processes taking
origin at its extremities. The interior of the cell is almost
wholly occupied by the nucleus, leaving but a scanty amount of
cytoplasm in the bases of the cell-processes (F'ig.63). The
chromatin is in the condition of a fine reticulum. The tigroid
substance is necessarily small in amount where there is so
little cytoplasm, embracing only a few scattering granules.

Structurally considered, the neurones of the nucleus genic-
ulatum are little specialized, remaining in an embryonic condi-
tion, so to speak. Their functional value is also of a low

114
order. Tney are cert-iinly of far less importance as an optic
termination in the selachian than are their specialized repre-
sentatives in the mamnnal. In y.ustelas, only collateral branch-
es are, evidently, sent to the nucleus ^eniculatum, the great
mass of optic fibres sweeping backward to the midbrain for ter-
mination in the tectjm. 'flith the progressive evolution of
higher vertebrates, the thalamic termination of the opticus
appears to have become more and more important, leading to the
corresponding differentiation of geniculate nuclei. Hand in
hand witn the growing importance of the interbrain as a prima-
ry optic centre, however, there has been a related decline of
the midbrain roof. And hence it is that the optic lobes of
the fish appear so lisproportionately large in comparison with
the homologous parts of the mammalian brain.

2. Epi thalamus: The Nuclei Eabenulae.

The nuclei habenulae, or ganglia habenulae of authors,
rise considerably above the level of the thalami (Pig.l, n.h.),
the pair meeting each other to form a conspicuously arched
bridge across the third ventricle at its posterior end. The
epiphysis springs from the middle of the arch. The left nucle-
us is a little larger than the right one, its margin extending
slightly more anteriorly.

The nucleus habenulae is an important relay-centre, and so
its strjctjre exhibits many nerve-fibres taking various direc-

lln
tions, between ^hich thera are neurones ani supporting elenoents.
The tractus olfacto-habenularis (Section Vin,l,c) teminates
here, and the neurones of the nucleus give origin to the trac-
tus habenulo-peduncularis, (the bundle of Meyaert, and the fas-
cicaltis retporeflexus, of authors).

A representative neurone is shown in Fig. 85. The size of
the entire element is rather large. The cell-body tends to
retain a rounded form, although diverted from this condition
more or less by the thickened bases of the dendrites. The
dendrites are some three or four in number, gnarled and irreg-
ular processes, branching only a few times, and extending far
outward in every direction from their points of origin. The
surface of a dendrite is roughened by nodal thickenings, knobs,
and a few gemmules. The axone arises directly from the cell-
body in all of the instances observed. Its course is tracea-
ble for only a short distance in a transverse section, since
it soon turns posteriorly into the tractus habenulo-peduncula-
ris. This important tract takes the usual course toward the
base of the miibrain. Its termination occurs there in the
nucleus interpeduncularis. The significance of the tract is to
be interpreted as a part of the olfacto-motor complex, discussed
nore particularly in Section VIII.

5. Eli DO thalamus: The Lobl Infer iorea.
The hypothalamus is very large in \(ustelus, projecting

116

far back bene=ith the diilbrain; consult Pii?.l. Its large size
is merely the expression of the unusual importance which is
assumed by this part of the interbrain in selachians. Intrin-
sic neurones, ani fibres from vrithout ara to be noted in num-
bers in both the infmdibulum and the lobi inferiores.

The wall of the infundibulum exhibits neurones separated
from each other by considerable intervals {Bi^.2^.). The cell-
body is polygonal or elongated-ov^l in form. The dendrites
are tei in number. They spread ^videly, rarely branch, and
pursue a nearly straight course.

Tfie Lobi I^zferiores are the most conspicuous feat ires of
the hypothalamus, a pair of great bulbous outpushings of the
lateral wall of the infundibalum (?'ig.l, l.i.). These lobes
are the seat of a crowded group of neurones, a fact which is
doubtless the ontogenetic cause of their large size.

Herrick ( '92) has described several distinct nuclei from
the lobus inferior (hypoarium) of the teleost, but I have found
it impracticable to distinguish cell-groups in Mustelus. The
neurones are disposed in a layer next to the limitans interna,
the cell-bodies forming a closely-packed zone involving some-
thing like the inner fourth of the thickness of the wall (Pig.
27, i.z. ). The dendrites are directed outward, forming, to-
gether with the nerve-fibres here, a fine tangle which presents
the appearance of a molecular layer with general stains.

The form of a neurone is quite unlike that of any other

117

foand in the anterior divisions of the briin. It is not dis-
similar to a fideiy spreading bush, the cell-body bein^ the
short stem, and the dendrites the top (Pi^.23), The denJrites
are thick at their bases, they give origin to only a few branch-
es, they taper gradually, and their tips usually reach almost
to the limitans externa. The surface of a dendrite exhibits a
multitude of spiny gemmules of various sizes.

The course taken by the axone depends upon the position
of the neurone. A neurone lying in the r-oof gives off its
axone from the base of the cell, and the axone passes ventral-
ly, branching profusely (Pig.?7). A neurone from the side-
wall (F'ig.SB) invariably has its axone emerging from the side
of the cell, taking a course toward the limitans externa for
a short distance, then branching in a T-shaped manner. The
fibres thus formed run parallel with the surface of the brain,
one tarning into the ventral part of the lobus, the other pur-
suing an arcuate course out of the hypothalamus (Pig. 27, f.b.).
Such a fibre is marked by varicosities at intervals, and it
bears collateral branches (Pig. 83).

The internal structire of two neurones from the lobus
inferior is given in Pig. 54. Tne figure also shows how close-
ly these neurones are packed. The nucleus is only fairly large,
and it is surrounded by a thick layer of cytoplasm on all sides,
the chromatic substance is scanty in amount; it is distriluted
in a few thin strands. The tigroid nasses are not numerous,
and most of them are ouite sn.all. witl; Just a few lart^e ones

IIB

iistributei at irretSular intervals in the peripharil region of
the cell.

4. Support lig ElemeitH .

Associatei with tne many ani crowiei nerva-traots charac-
terising the structure of the interbrain, there is to be notei
a corresponiing de\^elopinent of both neuroglia and epeniyma.

A neurogliar eleinent has a few stout processes raiiatin^
in every direction from the cell-body, often pursuing a marked-
ly tortuous course. They raiiify in an exceedingly complicated
manner, the finest t^igs interlacing to form a dense mat; see
Pig. 89. Neuroglia is found in the nuclei of the interbrain,
serving to support both the neurones and the terminal fibres
occurring there.

Spendyma is found in all parts of the interbrain. The
epeniymal fibre al-rays extenls throughout the entire thickness
of nervous matter, from the ventricle to the limitans externa.
The fibre branches mid/»ay in its course, the several limbs often
diverging considerably. The entire fibre-system bears a multi-
tude of fine mossy processes. These features are sho/»n in Fig. 30.

-5. Summary of the Interbrain.
The thalair.us is small in size and has remained on a low
plane of differentiation. It is practicable to distinguish

119

but tvto thalamic nuclei. One, the nucleic striti i^riaei, has
become isfined from the centr-il ^^r-iy mitt=»r for the reception
of f ibre-teminations, chiefly those of the trictus strio-
thalamicas from the striat'in; the axones of the nucle'is give
origin to the trictus thilano tectnlis. The other thnlamic
nucleus, the nucleus geniculatum, receives collateral branches
from optic fibres. It is wholly inferior to the tectum mesen-
cephali as an optic termination, but it represents the special-
ize! genicilate nuclei of higher forms.

The two nuclei habenulae do not exhibit great disparity
in size. An olfactory tract, the tractus olfacto-habenularis,
terminates here; the neurones of the nuclei give origin to the
tractus habenalo-peduncilaris.

The lobi inferiores are the seat of a crowded group of
neurones, and the lobes are thereby given a large size. A
neurone has a widely spreading dendritic top. Its axone branch-
es in a T-shaped manner.

Supporting elements are strongly developed The neuroglia
is remarkable for the mat-like interlacing of its branches.
Spendymal fibres often ramify to a striking degree; they ex-
tend through the .Thole thickness of the neural tube.

Section VIII.
The Porebrain.

Tha key to the naierstanJin^, of ths vertebrate forabrain
^as given by Rabl-Riickhari ('83) when he fornulatel the theory
of the membranous pallium for the teleost. The extension of
the generalization to other groups has been productive of re-
sults flhich fall into place in an almost schematic way. A
quite remarkable series is given by the fishes. This series
begins with a coniition in the teleost /rhere the entire roof
of the forebrain reiiains non-nervous, and culminate.^ in the
iipnoid flith a pallium having the essential morphological char-
acters pertaining to all brains of a higher order.

The forebrain of Mustelus appears superficially to be
somewhat livergent from the direct line of the series, o<?ing
to the absence of conspicuous external evidence of its bilat-
erality. The tfo striata are fused together into a solii ba-
sal mass; thera is merely a broad and shallow? furrow on the
ventral side to mark their plane of contact. The pallium,
while a thickened nervous plate as in far higher types of brains.
is perfectly continious between the right ani left sides, anJ it

is also confluent in the median plane with the striata below.

(1?0)

121

Thass p3culiarly corapactBi chHr.acteristics have been the source

of no little controN^ersy. 3tujnick'-t has atteraptei to sho»» as
one of the iicta of a serie:5 of papers ( '94a, '94b, '95, '38) that
the selachian forebrain stands entirely apart from that of all
other vertebrates. His conclusions were drawn from stuiies on
Petromyzon, and they appear to rest apon a misconception as to
the exte-it of the tela choroidea superior. The fallacies of his
views have already been pointed out by Burckhardt ('94c), and
by Rabl-Riickhari ( '94).

The lateral ventricles and the olfactory lobes of Mustelas
really anticipate the two-lobei condition of higher forebrains
in their essential characters. The lateral ventricles (F'ig.l)
are derived from the outer angles of the third ventricle, and
they pass the lamina ter-ninalis some distance apart from each
other. Their courses lie nearly parallel throughout. E^ach
gives off two diverticula, the olfactory ventricle and a dor-
sal branch, respectively. The dorsal diverticulum, (big. 31,
p. v.), is a short and narrow vertical cavity for the pallial
eminence; see Subsection 4. This stricture, while peculiar to
the brains of certain selachians, is nevertheless of some com-
parative value, as will be evident further on.

The olfactory lobes are massive prolongations of the lat-
eral angles of the forebrain. The primitive character of an ol-
factory lobe is exhibited by the sharply defined bulbus and
tractus, and by the presence of tne olfactory ventricle (?ig.
31, ol.v.). This is a slender cavity derived from the outer

122
siJe of ths latar^l ventricle at the Isvjl of the pnllial iiver-
ticalan.

The recessus neuropopicus of Burckharit ('9451) is of such
interest ^s to dssecy^e especial notice. It appears superfi-
cially as a ^ell-uarked depression in the median plane a little
anterior to the pallial eminences (F'ig.l, np.). It is render-
ed ev^en more noticeable during life by the penetration of blood-
vessels here, owing to which fact Rohon ('77) called it the
foramen nutritiuem. This structure is the vestige of a neuro-
pore. Its ontogeny was first described from the s^anoid brain
by vonKupffer ('90) as the lobus olfaotorlus impar. Rabl-
Riickhard describe! its development in the selachian embryo
('93); and Burckhardt ('943) has identified it in many other
vertebrates. The recessus neuroporicus of Mustelus retains
the character of an open passage in the adult animal (F'ig.31,
np.), serving as a channel for blood- and lymph-vessels. It is
accompanied by a pair of fibre-tracts, right and left, which
take this primitive path from dorsum to base of the brain; see
Pig. 31, m.s.t.

2, The Olfactory Lobe.
It is my purpose to devote a future paper to the entire
olfactory apparatus of Mustelus, and so a detailed description
of the olfactory lobe will not be ^iven here. For the under-
standing of what follows, it will be sufficient to state that.

123

in ths bulbus, th3 olfactory neurones of the first order are
chninei to those of the second orier through the asuil tanf^^les
kno/<n as the olfactory glomer'ill; and that the neurones of the
second orier, mitral cells, send their axones through the trac-
tus to terninal stations described further on. Other neurones
of the olfactory lobe, having a purely accessory value, need
not detain us at present.

S. The S tria tUTi.
The striatuTi is primarily a part of the olfactory appara-
tus, and its morphology must be interpreted with this fact in
mind. T'^o groups of neurones have become defined from the
general mass of the striatum for the especial reception of ol-
factory impressions. One of these lies next to the lateral
ventricle, the epistriatiim; while the other one is peripheral
in its location, the nucleus postolfactorius. The structure of
the principal mass, or general striatum, will be considered
after that of the epistriatum.

a. Eoistriatu-n. — The epistriatum is a group of neurones
lying ventral and lateral to the lateral ventricle ( ^15.31, estr. )
It comprises something like the inner fiftn of the entire thick-
ness of the striatum. While not so sharply ielimitei from the
outer levels as ^dinger ('96) has described for the reptilian
brain, the epistriatum of Mustelus is readily distinguished

124

from the general striatum by tha smaller size 'inJ more closely
cro**iei iisposition of its nearoaes. The neurone.? are not ar-
ranged in any definite order; they are of the Golgi TI type,
sealing their axones ventro-laterally into the striatum; refer
to Pig. 3?.

The celL-body of such a neurone rant^es from triangular to
polygonal in outline. There are three or four dendrites of on-
ly moderate length. These seldom branch except near their ba-
ses. They bear a very fev* geramules and irregular knobs. The
dendrites may extend indifferently in all directions, or they
may lie tangential to the limit-:ins interna.

The axone may arise from the cell-body, but in many in-
stances obser/ed it emerges from the thick base of a dendrite.
It passes into the region of the general striatum, bearing
short collateral branches along its whole length, and ultimate-
ly dividing into a widely spreading arborization between the
neurones of the outer levels. Pig. 33 shows the external feat-
ures of one of these neurones.

The internal structure of the neurone is shown in F'ig.*??.
The nucleus is seen to fill the greater part of the cell-body.
It presents an oval form, and has chromatin disposed in a few
relatively large masses, each consisting of a central clot
with radiating streamers. There is a single nucleolus. The
cytoplasm lies chiefly in the bases of the larger dendrites.
Its tigroid-bodies are small in size, few in number, and irreg-
ular both in form and distribution. Judf^ing from the struct-

1?5
are prasentel by this nsuroae, it is not charicterizei by a
high degree of metabolic activity.

The epistriatum is one of the nuclei for the termination
of olfactory neurones of the second order. Pibres also end
here /rhich have ascendei in the tractus strio-thalamicus, and
have crossed over in the anterior commissure. Some of the
terminations occur near the bodies of the cells, but many of
the fibres are to be traced to a narrow zone next the lateral
ventricle into which the dendrites of the neurones penetrate
(?ig.?3, ol.f.). Here the fibres give off branches /»hich run
parallel wit-i the limitans interna. The significance of these
terminations will be considered under the heading of the general
striatum.

Comparing the description and figures given by Johnston
( '93a) for the ganoid brain, it would seem that the epistria-
tum is the more sharply marked in Mustelus. Although found
in an animal ranking lower in the zoological series, the fact
is doubtless a correlative of the more powerful olfactory or-
gans which characterize the selachian organization.

b. General Striatum. -- Tne median zone common to the
pair of striata contains but few neurones. It is occupied
chiefly by interlacing fibres of small calibre. The greater
number of these are doubtless commissural, but in the more
dorsal region there appear to be connections with the pallium.

The ^reat mass of the striatum has neurones scattered

1?6

through it /»ith3ut any orier of arr-ini^efiien t (Pi^.r?l, str.).
These neurones are nsvec closely crowiei, there usually being
«ide inter\/als betteen the cell-boiies. Their ■dendrites are
very lons^, however, so that an interlacing plexus is given
throughout the ^hole field. Pig. 34 will illustrate these feat-
ures.

An individaal neurone is sho^a in Pig.3o drawn to a smal-
l3r scale than the other neurones of the striatuii because its
processes spread so wiJely. The cell-body Tiay be rounded,
oval, or polygonal in outline. There are numerous long and
slender branching dendrites. Several dendrites often arise
from a co-nmon thick stena .Thioh mi^ht almost be considered a
part of the cell-body. A dendrite is noteworthy for the many
little crooks which appear in its course, and also for the
peculiarly spine-like genrimules which beset it. The latter
feature is evidently characteristic of these neurones in fish-
es, as Van Gehuchten ('94) has noted their presence in the
trout, and Johnston ('99a) in the sturgeon.

The cell-nucleus has its chromatin disposed in slender,
branching threads, and in minute granules distributed some-
what diffisely (Fig. ^6). The cytoplasm forms a thick invest-
ment to the nucleus on all sides. Its tigroid masses are mar-
kedly larger than those found in the neurones of the epistri-
atum. The largest ones lie in the bases of the dendrites, and
these have a triangular form. Other masses of smaller size
an3 of more irret^ular shape are intercalated between the lar-

127

gsr ones. Tii?roiis extsnl into the leniritas for a short ii.s-
tance, only.

The axons takes its ori^,in directly from the cell-body,
so far as observed (*ig.3n). It gives off collaterals during
the first part of its length. The several axones descend to
the base of the forebrain and ran posteriorly in the tractus
strio-thalaiiicas for ultirnate termination in the thalamus; see
Section VII. This conspicuous fibre-tract was described by
Rohon ('77) as the peiunoul'is cerebri; by Sanders ('3^) as the
arus cerebri; and by idinger in his earlier ^ork ('88) as the
basale vorierhi rnbundel .

The general striatum has the type of structure and the
associations «hich are characteristic of motor centres. The
large neurones, with their long and videly spreading denJrites,
enter into a nexus with the axones derived from a purely senso-
ry centre, the epistriatum. Tne latter receives, chiefly,
impressio-is of the olfactory order. The neurones of the stri-
atum are indirectly affected through the axones of the epistri-
atum, carrying, in tarn, the nervous disturbance to the inter-
brain, whence a relay carries it to the great roof-nucleus of
the midbrain for direct connection with the neurones of the
body musculature. The enormous olfactory organs of Mustelus
indicate how important the olfactory sense must be in the econ-
omy of tne animal, and observation demonstrates the large place
taken by olfactory impressions in the location of food. The
striatui^i is evidently the centre for motor reflexes, of which

128

the sensory neurones chaining the olfactory ort'in to it con-
stitute one arrr, and the neurones intervening betvfeen it -inj
the body nusculature comprise the other.

0. Vuoleus Postal faotor'iis . -- The nucleus postolf acto-
rius is a collection of neurones at the ventro-lateral sur-
face of the striatum near its anterior end (B'ii^.?!, n.po.).
This nucleus is sharply demarcated from the general striatum
by the closeness .vith flhich its neurones are arranged; Pi^.?''.
shofls the morphology of fi\?e neurones as they lie in place.
The peripheral zone is occupied by fibres, and also by the
dendritic tips of the neurones themselves. The cell-bodies
lie in a densely matted tangle of nervous processes just in-
ternal to the peripheral layer. A cell-body is somewhat lar-
ger than one from the epistriatum. The form ranges from py-
ramidal to elongated oval. The dendrites are rather fe/» in
number, relatively short, branching dichotomously once or twic<^
They are usually stout processes, gnarlei, rough, and irreg-
ular, with bit few true gemmules. Some neurones send their
dendrites radiately in all directions; others take a more or
less tangential course; still others are extended between the
zone of nerve-fibres at the periphery and the striatum within.
A conspicuously felt-like tangle is thus given by all of these
interlacing dendrites.

The axone arises from one of the dendrites, so far as ob-
served. It passes into the tractus olf acto-habenularis, cours-

129
ia^, posteriorly aloni* the ventral border of the forebriin for
ultimate termination in the nucleus habenulae of the inter-
brain; see Section VII.

The internal strictjre of a postolfactory neurone is not
of a prono'incei motor type. The nucleus is al^vays so lar^e
that it touches the periphery of the cell at one or more points,
leaving the cytoplasm almost entirely in the bases of the den-
drites. The chromatin lies in a fine reticulum. There may be
two nucleoli. The tigroid substance is disposed in triangular
or irregular bodies of smaller size than those noted for the
striatum. These features are shoAin in Pig.*7.

The nucleus postolf actorius receives many olfactory fibres
of the seconi order. A fibre entering for termination here is
seen in B'ig.36,ol.f. This nucleus evidently holds a very dif-
ferent place in the olfactory apparatus from the epistriatum.
The latter appears to be a sensory gateway to the striatum,
while the postolfactory neurones enter into direct relations
^itn the posterior regions of the brain.

•9. The nucleus fl europoricus ,
A group of neurones situated in the region of the recessus
neuroporicus gives origin to a special bundle of the tractus
strio-thal amicus, the medianbu^del of Sdinger ('98). I shall
designate this group the nucleus neuroporicus. The neurones
of the nucleus neuroporicus lie chiefly behind tne external

130
openiafi of ths neuropora, just beneath the surface of «hat is
really the most anterior part of the pallium. Other neurones
of the same group are found at deeper levels, in the striatun,,
of course. The entire (group is distributed, therefore, without
regard to the anatomical boundaries vthich ^a seek to dra» be-
t/»een the pallium and the striatam.

A sagittal section of the neuroporic nucleus stained by
the 3olgi method exhibits a bewilderinfs nervous tangle. The
components of this plexus are the processes of the intrinsic
neurones, and the terminations of the olfactory fibres *hich
s/»eep over from the tractus. The nucleus is seen, then, to be
an additional olfactory centre.

Two neuroporic neurones are sho/»n in Pig. 37, exclusive of
the maze of nervous processes in flhich they lie. The cell-
body is large, the largest, in fact, of any found in the fore-
brain. Its form is quite distinctly polygonal, the number of
sides being determined by the number of dendrites. Tne den-
drites are very thick at their bases, and they are some three
to five in number. Each one soon breaks up into large branch-
es. The length of a dendrite is relatively short. Its sur-
face exhibits thickenings and minute knobs sparingly distrib-
uted.

The axone may arise from the cell-tody, but in most in-
stances its origin is traceable to the thick base of one of
the dendrites. The several axones from each lateral half of
the nucleus run anteriorly to the level of the neuropore, where

131

they car</e upon themselves to follow the course of that chan-
nel, forming the meiianbiindel of ^dinger ('3-3); see Pi*?.?!,
•n.s.t. The melian bundles, right and left, are thus carried
posteriorly and centrally beside the recessus neuroporicus
to/farJ the base of tne brain. Here each blends '^ith the part
of the traotus strio-thalamicus taking origin from the stria-
tum of that side, for'ning its central portion.

The coarse taken by the median bundle, curving forward as
it does to lie beside the vestige of the neurDpore, is certain-
ly fraught witn significance. This is probably an ancient
route which was at one time quite direct, but the phylogenet-
ic enlargement of the forebrain has carried the recessus neu-
roporicus and the seat of the neurones ever farther and farther
apart.

The neurones of the nucleus neuroporicus probably function
as an olfacto-motor centre, supplementary to the chief one
represented by the general striatum. The relations of this
nucleus to the pallium will be discussed under the following
subsection.

4. The Pallium.
The neurones of the pallium lie chiefly in what are called
in this paper the pallial emiienoes; see Big-l and t<'ig.31, p.e.
I have introduced the definitive term pallial eminence for the
hemispherical roof of the dorsal diverticulum of each lateral

13?

ventricle (Fi(?.'=!l, p. v.). The enrlier Hiritoaiists had observed
the pressace of these elevations in certain selachian fore-
brains, but their place in the organization of the palliarn has
not been recognized heretofore. A pallial eminence is the
seat of a crowded group of neurones, and the elevated condi-
tion of the mass has probably arisen from the disproportion-
ately rapid growth which occurs here.

The superficial zone of a pallial eminence is occupied
almost exclusively by axis-cylinders, the significance of which
will be noticed presently. The neurones lie in a crowded aggre-
gate just vit^in. Pig. 39 shows a representative group of these
neurones. They are seen to embrace several varieties of form
and size, all disposed without arrangement into definite layers.

a. neurones of the Tr-aotus Pallii. — The external feat-
ures of the largest variety of neurone of the pallial eminence
will be seen by referring to F'ig.39. This type assumes a vari-
ety of guises, but it is characterized by its larger size and
by the fact that its axone enters the tractus pallii. The
cell-body of such a neurone ranges from broadly oval to dis-
tinctly polygonal in forn. The dendrites are long and widely
spreading processes, arising through thick basal masses at
three or four almost equidistant points. A dendrite branches
once or twice, dichotomously as a rule, and the terminal
lengths become quite slenier. Its surface is slif^htly rough-
ened by minate ^emmules. These are far less conspicuous, how-

133
evsr, thin ths gemmulss obser/eJ in the neurones of the strii-
tu-n.

The internal stracture of one of these neurones is shown
in Fig. '^3, p.t.n. The or^aniz^ition is markedly motor in type.
There is an abaniance of cytoplasm investing the nucleus on all
sides. The nucleus is a sharply defined, central body, /lilh a
single nucleolus, and chromatin which is disposed in a condi-
tion strongly suggestive of a reticulu-a, although not actually
appearing so. The tigroid-bodies are quite large in the re-
gion of the nucleus, the largest, in fact, of any tigroids
found in the forebrain. Vore slender tigroids extend into the
dendrites for some distance. In the cell figured, the axone
exhibits a small axone-hillock of oval form.

Tae axone arises from the cell-body, (Pig. *i8, p.t.n. ), or
from the thickened base of a large dendrite, (Pig, 39). It pur-
sues a straight course, although one with many local sinuosi-
ties. The several axones of this class come to lie in the
superficial zone, and they are gathered into a longitudinal
bundle which lies as a broad cap on the neurone-aggregate of
the pallial eminence (Pig. 31, tr.p.). This is the beginning
of the tracLiS pallii. The pallial tract runs backward to the
great posterior carve of the forebrain, where it dips ventrally
to the base and enters the interbrain. Here it decussates and
passes onward towari the oblongata. The tractus pallii is the
mantelbiindel of "Ldinger ('38).

134

b, AssocLatLoe ani Commlnsaral 1 eurones. -- A second type
of neurone founi in the pallial eminence is distinguished from
the preceding one by its associative or cotniiissaral value, and
it is also readily recognizable by its soialLer size (Fig. 40).
The cell-body tends to retain a globular form. It gives ori-
gin to tfo or three dendrites which radiate frorn it at equidis-
tant points. The dendrites are only moderately stout, and
they are never of great length. They branch but a few times.
Their surface is sparsely studded with small gemmules. The
axone nay terminate on the same side of the brain; or it may
enter the pallial commissure, (Fig.31,p.c. ), for decussation,
terminating in the opposite half. The neurone shown in F'ig.
40 is an example of the commissural class.

The internal structure of a neurone of either the associa-
tive or the commissural type is drawn in Fig. 63, as.n. The
nucleus is a subspherical mass, so large that it occupies prac-
tically the whole of the cell-body. The only cytoplasm dis-
tinguishable is that which composes the bases of the dendrites.
The nucleus stains quite deeply, owing to the dense reticulum
of chromatin which pervades it. The cytoplasm has only a few
tigroids of small size, an evidence of a feeble degree of activ-
ity for this type of neurone.

a. Cajal Veuroiea. — Still a third type of neurone is
clearly the representative of the Cajal cell which Ramon y
Cajal ('91) described from the cerebral cortex of the rabbit.

13FS
and flhich Retzias ('93), and Vsritti ('97) have iesi^natel aft-r
the naras of ths discoverar. Cajal neurones have been observed
in all of the higher vertebrates by various investigators, and
their identification in the brain of Mustelus makes it prob-
able that they are common to all groups.

A Cajal neurone is dra«n in Pig. 41. This neurone occu-
pies a superficial position, just beneath the stratum of fi-
bres noted above; refer to Pig. 31. The cell-body is an irreg-
ularly elongate-oval, with its major axis horizontal. Prom its
opposite extremities, thick dendrites arise which run more or
less nearly parallel with the limitans externa. A dendrite
does not have marked turns in its coarse. It gives off branch-
es along its upper margin which ascend toward the surface of
the brain, while a few branches are derived from its under side
and penetrate to deeper levels. The dendritic surface bears
small gemmules and larger bosses in considerable numbers.

The axone is a direct continuation of one of the dendrites.
Its point of origin is indicated by the surface becoming smooth
and the course more irregular. The axone continues to hold a
course tangential to the limitans externa, and it terminates
through a few small branches at no great distance from its
point of origin. This type of neurone is evidently purely
associative, linking together areas not widely separated.

Strict irally, the Cajal neurone is somewhat peculiar (Fig.
69). The tigroid substance is collectei into a few lumps of
relatively large size. These are disposed irregularly in the

136
gre=it masses of cytoplasm /rhioh lie lat?r=il to the nucleus.
The nucleus is eccentric in position, leaving but a thin pel-
licle of cytoplasm on one of its si'ies. The chromatin is group-
ed into a fef stranis having conspicuously thickened nodes at
intervals.

i. Geieral Oonsi ierat ions on the Pallium.-- The pallium
Of the selachian is really a primitive representative of the
stem giving origin to the pallia of amphibians, reptiles, birds,
and mammals. The continuity of its t'.vo halves, contrasted .vith
the two-lobed condition of the higher type, has caused many
/writers to rank the selachian pallium as a divergent branch.
The correctness of tne vie'? set forth in this paper v»ill appear
«hen the significance of the relations /fhich exist between
certain neurones of Mustelus have been considered.

It has been shown that the ventral border of a lateral
ventricle is the seat of one olfactory termination, the epi-
striatum. The epistriatua tends to become extended around the
siie of the ventricle towarl the pallium. Axones from the
marginal neurones of this olfactory centre enter the pallium,
thus serving to link that region indirectly with the olfactory
organ. Besides this source of impressions, the olfactory centre
of the nucleus neuroporicus radiates an influence to the pal-
lium, of which it is really a part. The pallium of Mustelus,
therefore, is seen to have a general association with the olfac-
tory mechanism.

137

It has been shown by 5diriger ( '9^) that the special pal-
llal olfactory coniucting path characteristic of higher verte-
brates appear? for the first time in the brain of the reptile.
Both "iliin^er ani Herrick have also shown that the first sense
to thus enter the field of consciousness is the olfactory one.
The palliirn of the selachian really anticipates the reptilian
olfactory connection, although, of course, in a much simpler
way. In Vmstelus, we therefore find quite a primitive condi-
tion represented.

3o far as we may be permitted to interpret morphological
facts, the palltjm of Mustelus would appear to be a long-dis-
tance motor centre of the olfactory apparatus. Other olfacto-
motor centres there certainly are in abundance. Both the gener-
al striatum and the nucleus neuroporicus contribute motor neu-
rones to the tractus strio-thalamlcus; while the nucleas post-
olfactorius sends a tract to the nucleus habenulae. These
centres obtain connection with posterior regions only through
relays in both the interbrain and the midbrain. The pallium,
on the other hand, sends its tract directly through the base
of the interbrain toward the nuclei of the great nerves of the
oblongata. The tractus pallii, therefore, gives the olfactory
sense an additional hold on the nervous system, a series of
connections which cannot obtain, of course, in those fishes
with membranous pallia.

The phylo^enetic development of the pallium of selachians
would thus appear to be the outcome of the great dependence

138
which thsss fishes place upoti the olf-ictory app^ritus in the
search for fool. In higher vertebrates, as the olfactory sense
becomes linkei to the palliam through stronger bonds, and as
other senses make pallial connections, one after another, this
part of the brain takes on functions of an ever hii?her value.
In Mustelus, there is merely an anticipation of pallial possi-
bilities.

■'). SuDDor tiig Elements.

Both neurogliar and ependymal elements are present in the
forebrain of Viustelus.

Ependyma is found r^idiating from every part of the later-
al ventricle, and its characters are nearly uniform for the
several regions. The ependymal cells have their nuclei situa-
ted at slightly different levels, and so a broad zone next to
the ventricle presents chiefly nuclei. The shape of the cell-
body is influenced by the position of its nucleus, of course,
bit the broadened part is usually not directly in contact with
the ventricle. The ependymal fibre runs straight outward from
the cell-body, as a rule, and it reaches entirely to the lira-
itans externa. It does not branch, its size remains nearly
uniform, and its course is only slightly irregular. It bears
a greater or less profusion of delicate, mossy twigs of short
length. The inner portions of two ependymal elements are shown
in Pi^.42.

199

Neuroglia is founl in the severul parts of the forebr^iin
whera nervs-cell.s are grouped in numbers. While the specific
forms assumei by aeurogliar elements exhibit consiierable diver-
sity, all are referable to but one type. Proro an irregular cell-
boiy, numerous fine processes extend in all directions for a
short distance, branching profusely as they proceed. The ulti-
mate t^igs are of quit? minute size, and hence it is that the
«hole presents a characteristically mossy appearance. Fig. 43
illustrates the features of a representative specimen.

6. Summary of the ^orebrain.

A fairly well-defined epistriatum is present, receiving
olfactory and other terminations. Its neurones are of the Golgi
II type, sending their axones into the striatum, in the main.
The axones from marginal zones enter the pallium.

Neurones vrith widely spreading dendrites are arranged in
open order in the striatum. Their axones enter the tractus
strio-thalamicus for termination in the thalamus. The stria-
tum appears to be an olf acto-motor centre.

The nucleus postolf actorius is a densely crowded group of
neurones. Olfactory fibres terminate here, and the derivative
axones form the tractus olfacto-habenularis.

A group of large neurones in the vicinity of the recessus
neuroporicus gives origin to the median portion of the tractus
strio-thalamicus. This paired tract accompanies the neuropore

140

for some iistaac^. The nucleus neuroporicua is a thiri olfac-
tory centre.

The pallium has its t-^o halves fused in the median plane.
Its most important region is the pair of pallial eminences.
Sach of these has a special extension of a lateral ventricle.
The neurones of a pallial eminence ars not arranc^ed in layers,
but it is practicable to recognize three distinct forms:
(1) The neurones of the tractus pallii, the largest variety,
their axones comprising the tractus pallii. (2) Commissural
ani associative neurones, the axones of /?hich are distributed
in the pallium itself, a special decussation of commissural
fibres occurring in the median plane. (3) Cajal neurones,
lying tangential to the limitans externa.

The pallium of Mustelas is regarded by the writer as a
primitive representative of the stem giving origin to the pallia
of amphibians, reptiles, birds, and mammals. It anticipates the
olfactory connections of the reptilian brain. The tractus pal-
lii is interpreted as giving the olfactory sense a hold on the
nervous system in addition to that provided by the epistriatum
and striatum, and the neuroporic and postolf actory nuclei.
The phylogenetic development of the pallium of selachians is
belie/ed to have been the outcome of the dreat dependence placed
upon the olfactory sense by these animals.

Section IX.
General Summary and Conclusion.

Special problems relative to the several parts of the
brain have already been iiscussei in the foregoing pages; flhile
summaries have been plaoei at the close of each of the sections.
There yet remain for consicieration a fe« topics of more general
scope gro/»ing out of the study as a /?hole. Before proceeding
to these questions, ho/vever, it will be desirable to review
briefly the most important results which have been obtained.

1. General Summary.

The Oblongata of Mustelus has become only slightly diver-
gent from the structural plan of the primitive neural tube.

The ventral cornu of the spinal cord is oontinued into
the oblongata as the nucleus of the VI nerve, and as the scat-
tered commissural and tract-neurones of the formatio reticularis.

The viscero-sensory system has had annexed to it a complex

of peripheral sense-organs. The communis system -.vhich results

from this union is represented by components of the VII, IX,

and X nerves. The lobus vagi is the centre of the system.

(141)

142

Kewer fibres enter the fascicalus corDinunis th'io in either the
teleosts or the amphibi=ia3.

The viscero-Tiotor nacleus gives origin to the motor fi-
bres of the V, VII, IX, and X nerves. The axones enter their
nerves chiefly through the medium of the fasciculus longituii-
nalis dorsalis. The nucleus receives impressions radiated from
the lobus vagi.

The general cutaneous nucleus is the homologue of the dor-
sal cornu of the spinal cord. General cutaneous fibres, compo-
nents of the V, IX, and X nerves, terminate in both the substan-
tia gelatinosa and the deeper part of the nucleus. Many fibres
of the systen enter the spinal V tract for ultimate distribu-
tion in the spinal cord.

The tuberculuTi acusticum is phylogenetically young. It
may have been derived from the structures of the dorsal cornu.
Pibres of tne acustico-lateral system terminate in the tubercu-
luTi acusticum. These are components of the VII, VII [, and X
nerves. Neurones are present in the tuberculum acusticum of
the molecular, granular, and Purkinje types, equivalent morpho-
logically to those of tne cerebellum.

The Oerebellim is relatively large in Mustelus. This fact
is to be interpreted from the strong development of the sense of
equilibrium in the animal. The strictural plan of the cerebel-
lum is the same as that characteristic of higher vertebrates,
the differences being due to greater simplicity of detail.

143

The aviisncs is for the orii^ia Df the cerebsllu-n, in the
phylogeny of the vertebr-ites, as a fusei outgrowth of the pair
of tuberoala acustica. The organ represents a specializatioT
of that part of the oblongata forming the original terminal
station for the acustico-lateral systeii.

The Viibrairi appears to have its organization arrange!
c:)ntribatory to the roof-nacleis. This is a group of giant
neurones, the axones of which enter into and very largely compose
the fibre of Reissner. The roof-nucleus and Reissner's fibre,
together with the motor neurones of the spinal cord, provide a
direct path for aotor reflexes between certain senses and the
body lusculature. The senses thus nediated are, primarily, the
olfactory and the visual, but the acustico-lateral and the gen-
oral cutaneous systems may be represented also.

The stratum medullars profundum is an important highway
into which there are traceable optic fibres, axones from the
tectum mesencephali, fibres of the olfactory apparatus, and
fibres fr^m posterior regions. Fibres emerge from the stratum
to terminate near the cells of the roof-nucleus.

The tectum mesencephali receives practically all of the
optic fibres in Mustelus. Three zones of neurones are recog-
nizable in the tectum, with optic terminations in all of them.
The deepest of the three layers has a more generalized struct-
ure than the corresponding zone of the teleost.

144

The Interbrali is the s3Ht of several important, relays
betflesn the forebr-iia and the posterior briin-se^ments.

The thHlanas is but slightly dii ferentiatel. Only two
thalamic nuclei are recognizable. The nucleus strati grisei
receives fibres, chiefly from the tractus strio-thalamicus;
its axones give origin to the tractus thalamo-tectalis. The
nucleas genicalatam receives collateral optic terxinations.
It is wholly inferior as an optic centre to either the tectum
mesencephali of Mustelus, or to the speciali:2e1 geniculate
nuclei of higher foriis.

The t'fo nuclei habenulae are not greatly unequal in size.
The lobi inferiores are the seat of a cro/fded group of peculiar
neurones.

The Forebrain is regarded by the author as anticipating
the forebrains of higher vertebrates in many respects. A fair-
ly well-defined epistriatum is present, the axones from which
enter both the striatum and the pallium. The striatum is an
olfacto-motor centre; its axones enter the tractus strio-tnal-
amicus for tertnination in the thalamus. An accessory bundle
of the tractus strio-thalamicus is derived from the nucleus
neuroporicus. An additional olfactory centre is provided by
the nucleus postolfactorius.

The pallium has its neurones grouped, without arrangement
into layers, chiefly in the pallial eminences. Three varieties
of neurones are to be distinguished. The pallium receives

145

imprassioris r-idiat.e:! from the adjacent olfactory nuclei; it
therefore anticipates the olfactory connections of the reptil-
ian ani higher brains. Tne tractas pallii, arising from the
pallial neurones, is regariei as giving the olfactory sense a
quite direct connection with posterior regions.

2. Conclusion.
To one fho has read this far, it must be evident that
thera is a most remarkable structural similarity bet^esn the
brain of Mustelus and the brains of higher vertebrates. The
results obtained by me do not bear out the conclusions of
Szczawinska ('98) relative to the very low plane occupied by
the selachian neurones; see Section II, 2. The neurones of
Mustelus are, of course, simpler in their external morphology,
and their architect jral relations are of a far less complicated
order, yet it is none the less true that they anticipate the
conditions found in higher vertebrates in all important partic-
ulars. Such a fact is certainly the more remarkable when the
great differences in the scale of general organology are taken
into account. The fact can only be interpreted to mean that
the nervous system of the primitive vertebrate had its essen-
tials of organization veil defined before the divergence of the
several phyla occurred. A brain of the type presented by the
selachian of to-day has not become sufficiently specialized du-
ring the lapse of time entirely to mask the ancestral charac-

146
teristios. The br-iin of one of the hii^,her vertebc'ites embol-
ies many nodi fications of the original plan, viroaght in the
course of its graiual evolution. These alterations may even
become more conspicuous than the primary structures upon flhich
they have been superposed, anl it is only through comparisons
with a less differentiated condition that we can hope to dis-
tinguish the new from the old.

A comparative study of the several brain-segments of Wus-
telus is productive of some results that might not have been
anticipated. The cerebellum is marked by an organization out
of all proportion to that of adjacent regions, an organization,
moreover, far more highly differentiated than is presented by
the cerebellum of either the amphibian or tne reptile. The
oblongata, on the contrary, has retained the plan of structure
of the primitive neural tube without the intervention of pro-
found changes. It is due to this fact that homologies between
the oblongata and the spinal cord are so readily traced in this
animal. A more extreme degree of simplicity is found in the
interbrain, the thalamus having such a slight differentiation
as to make comparisons between it and higher thalami somewhat
difficult. Finally, the forebrain is far in advance of the
forebrains of other fishes. Contrast the membranous pallium
of the teleost or the ganoid with the nervous pallium of Muste-
lus, which, as has been pointed out in the preceding pages,
anticipates the olfactory associations of higher brains to a
noteworthy degree.

147

These ilLastritioas clearly point to an unJerlying princi-
ple. Tne orgiinizHtion of the brain is the expresision of the
adjust-nent «hich has constantly taken place betvjeen the race
of animals and the stimuli to which thay have been subjeutei.
This relationship between nervous organization and peculiari-
ties in the anvironment is such a close one that the degree of
developTient of the several parts of the brain may be very une-
qual indeed. And hence it is that the cerebellati of Mustelus
is so highly organized, for this is the correlative of the po/r-
erful swimning capacity of the animal, requiring an adequate
mechanism of equilibration. The forebrain, with its luxurious
development of neurones, nas arisen in connection with the
large place occupied by olfactory impressions in the Selachii.

Worphological data in neurology must necessarily provide
the foundation for all physiological /rork, but it is none the
less important that morphological facts should have the check
of experimental evidence /fherever this is possible. Nowhere is
there to-day a more urgent need for careful observations of this
character than in the group of the fishes, where nervous proc-
esses are of such a simple order as to introduce relatively
few complications. Thorndike ( '99) has made at least a begin-
ning in this field for the teleosts. This observer placed a
screen with but one opening across the course in which Fundulus
desired to swim. He found that after several repetitions the
animal "learned to get out". Here, it would seem, there was a
trie memory of previous activities in the absence of any nervous

143

palLiaii at all, a very suggestive fact in connection with the
interpretation placei in the present paper upon the pallium of
Mustelus. An extension of the scope of experimental work on the
fishes will certainly prove fruitful for comparative neurology
in so many ways that the writer feels impellei to bespeak a
larger place for this kini of work in oar investigations.

Since the promulgation of the neurone concept by i^aldeyer
in 1391, no work has ieraanied more critical attention than that
on the ultimate fibrillar structure of the nervous systen,
studies with which tne names of Apathy, Eethe, and Nissl will
ever be connected. An investigation of neuro-f ibrils obviously
lies beyond the bounds set for the present research, and obser-
vations concerning them in Mustelus must await another opportu-
nity for expression. In the meantime, I would join with the
protest made by A.Veyer ('99), Parker (1900), and Verworn (1900)
against the tendency to elevate the results of specific methods
into an exclusive dogma. Although spoken from a different
vantage-point, the words of Golgi (1900) may well be quoted
here: "The knowledge which we possess, either anatomical or
physiological, is not yet such as to permit us to interpret
with certainty the greater number of the facts discovered,
much less to attempt doctrinal constructions of a high order
on the functional mechanism of the nervous elements".

Section X.
Literature Cited.

Baer, Karl Ernst von,

'37. Ueber En t » ickel an^si^eschi oh te ier Thiare, Bd . 2,

S. 503-311. Koni?sberg.

Barker, L.P.

'99. The Nervous System and its Constituent Neurones.
New York.

Bethe, Albrecht,

'96. Eine neue Methode Ier Methylenblaufixation.
Anat . 4nzei?. , Bi . 18, S. 438-445.

'99. Die Locomotion des Haifisohas (Soyllium) and ihre
Beziehuni^en zu den einzelen Gehirn thei 1 en und zum
Laby ri n th .
Arch. f.d.ges. Physiol., Bd.7 6, S. 470 -4 93.

Burckhardt, R.

'92. Das Central nervensystem von Protopteras annectans:
Eine v e r i?l e i chen i - An a t omi so he Studie. Berlin.

'94a. Die Homolgieen des Zw i sohenhi rndaohes und ihre

Bedeutung fur die Morphologie des Hirns bei neider-

on Ver teb ra ten .

Anat . Anzei ? ., Bd . 9, S. 152-155.

Zur ver^l e i chen len Anatomie des Vorderhirns be:

? i sohen .

Anat . Anzei^., Bd .9, S. 375-382.

(149)

150

94o. Bemapkun^en zu P . K . S t ud n i rik a ' s Mittheilun^ uber das
Pisohgohl rn .
Anat . 4nzei^., Bd .9, S.468-4S9.

Dep Bauplan dep Wi pbel t he i pgeh i pn ,
Mopph. Apbei t ., Bd.4, (.131-149.

;ox, W.H.
'91 .

Imppagna t ion des centpalen Me pvensy s tens mit Queok-

silbepsalzen,

Apchiv f.mik.4nat., Bd.37, S. 16-21.

Edin?ep, L.

'88. Dntepsuohungen iiber die v epgl eichend e Anatomie des
Sehipas. I. Das Vordephipa.

Abhandl.bepaus.v.d.3enckenbep?.natipf.0es.. Bd.15,
S. 91-119.

'92. Ontspsuchungen iibep die vep^leichende Anatoiiie des

Sehipns. II. Das Z» i schenh i pn .

Abhandl.hepaus.v.d.Senckenbepg.natapf.Oes., Bd.lB,
S.1-55 .

'96. Untepsuchungen ubep die vepgleichende Anatoraie des
Gehipns. III. Neue Studien iibep das Vopdephipn
dep Rep t i 1 ien ,

Abhanil.hepaus.v.d .3enckanbepg.natupf.Ges., Bd .19,
S.. 51. 5 -3 86.

1900. Voplesungen iibep den Bau dep nepvosen Cen t palo pgane,
6te Auf 1 . Leipzig .

Ewapt, J.C.

'89. On the Cranial Nerves of El asuobpanch Pishes. Pre-
liminapy Oonmun ioa t i on .
Ppoc. Royal Society, Vol.45, pp. 534-537.

'92. Supplemen t apy Note on the Cpani»l Nepves of Elasno-

161
branohs . Ed i nbur>!h .

"leohsi ?, P.

'89. Debep eine noue Pirbungsme t hod e dea centralen Nerven-
systens unl deren Ergebnisse beziiglioh des Zuaam-
Tienhan?es von Oanglienzellen und Nervenfasern.
Arch. f . 4nat . und Physiol,, Phy s iol og . Ab t h . , S.537.

P r i 1 3 c h , J .

'78. Un tefsuchun^en iibep den feineren Bau des Pisch-
gebirns. Berlin.

Qaskell, W.H.

'86. On the Structure, Distribution and Punction of the
Nerves which Innervate the Visceral and Vascular
3y s tenns .
J our. of Physiolor^y, Vol.7.

'88. On the Relations betfoea the Structure, Punction,
and Distribution of the Cranial Nerves. Prelimi-
nary Oouniun i oa t ion .
Proc.Poyal Society, Vol.43, pp. 382-390.

'89. On the Relations between the Structare, Function,
Distribution and Origin of the Cranial Nerves,
together with a theory of the Origin of the Nervous
System of Vertebrata.
J our. of Physiology, Vol.10.

fJegenbaur, C.

'70. Grundzuge der ve r gl e ic henden Anatomie,2te Aufl. Leipzig.

'74. jrundriss der ve rgl eichand en Anatomie. Leipzig.

'98. Vergleichende Anatomie der Wirbelthiere mi t beruck-
sichtigung der Wirbellosen, Band 1. Leipzig.

Gehuchten A. van,

'94. Contribution a 1' etude du syst^me nerveux des te'-

1 e'o s 1 9 e n 3 .

La Collule, T.IO, p. 355-395.

Gol?i, Camil lo,

'94. Dn tepsuchun^en uber den feiaepen Bau les Gentpal

un.l Pepiphepischen Mepvensystams, ubepdetzt v.Teusohop,
Jena.

'98. Sup la stpuot'ipe d e s cellules nepveuses.
Apoh.Ital.de Biol., T.30, p. 60-71.

1900. Intopno alia stputtapa delle cellule nepvose della
coptecoia cepebpala.
Anat . Anzei?., Ep?anzuneshef t, B1.18, S. 164-176.

Spaf, Arnold,

'98. On the Dse and Properties of a New Pixii? Fluid,
( Ghroue-Oxal ic ) .
New York State Hospitals Bulletin.

Heidenhain, M.

'93. Ueber Kern und Pro topi as-na.
Pestsohr . f . Kbll ikep, S.113.

Held. Hans,

'95. Beitpa?e zup Stpuotur der Nervenzellen und ihrer
Poptsatze.
Arch. f . Anat .und Physiol.. Anat.Abth. S. 596-416.

Heppick. G.-JudsoQ.

'97, The Cranial Nepve Components of Taleosts.
Anat . Anzei? .. Bd.1-5, S. 435-431.

'98. The Cranial Nepves of Bony Pishes.

Joup.of Comparative Neurology. Vol.8, pp. 163-170,

The Cpanial and Pipst Spinal Nerves of Monidia; A
Contpibution upon the Nerve Components of the Bony
Pishes.

158

Joar.of Corapap . Neurology , Vol.9, pp. 153-455.

Herriok, C . L .

'92. Contribution to the Morpholo?/ of the Brain of
Bony Pi shes .
Jour. of Compar .Neurology, Vol.2, pp. 21-72.

Holnogren, Emi 1 ,

'99a. Zur Kentniss der Sp i nal 'anijl i en zel 1 en des Kanin-
chens und ies I^rosches.
Anat . Anzei J., B1.16, S. 161-171.

'99b. Weitere Mittheilungen uber ien Sau der Nervenzellen.
Anat . Anzeig., Bd .16, S. 588-397; Bd . 1 7, S. 113-129.

Houser, Gilbert L.

'97a. The Uses of Pormald ehyie in Animal Morphology.

Report of Iowa Academy of Soienoes, Vol.4, pp. 147-
151.

'97b. The Nerve-Cells of the Shark's Braii; Preliminary
Communication.

Proceedings of the Iowa Academy of Sciences, Vol.4,
pp. 151-153.

Johnston, J.B.

'98a. The Olfactory Lobes, Pore-Brain, and Habenular
Tracts of Aoipenser.
Zoolog. Bullet in, Vol.1, t)p. 221-241.

'98b. Hind Brain and Cranial Nerves of Acipenser.
Anat . Anzeig.. Bd . 14, S. 580-602.

Kingsbury, B.P.

•97. The Structure and Morohology of the Oblongata in
Pishes.
Jour. of Compar. Neurology, Vol.7, pp. 1-36.

154

Koellikep, A.

'96. Hanlbuch ier Gewebelehre las Menschen, 6te Aufl.,
Band 2 . Leipzig.

Kul tschi tzky, N.

'87. 7,ur Kentniss der modePQen ?ixipun?- und Conserv-
ipun^smittel.
Zeitsch.f .wiss.Mikr., Bd.4, S. 54 5-349.

Kupffer, C.von,

'90. Mitteilunfen zur Ent»i3klungs?eschichte des Kopfes
bei Aoipenssr sturio.

Si tzb .d .Ges . f . Mopp hoi .u , Physiol . i n Munchan, 107-
123.

Lee, Art hap Bol las,

1900. The Miopo totni s t ' s Vade-Meoum, 5th Edition. London,

Lee. Fred epic S,

Debep den Gle ichgew i ch t ss in n .
Centralbl.f. Physiol., Bd.6, S.5 08.

13.

A Study of the Sense of Equilibriam in Pishes.
Joup.of Physiol ., Vol .15, p. 311.

'94. A Stidy of the Sense of Equilibpium in Pishes,
Part II.
Joup.of Physiol ., Vol .1 7, p. 192.

'9 8. The Functions of the Eap and the Latepal Line in
Pis hes .
The Amep. Joup . of Phy s i ol . , Vol . 1 , pp. 128-144.

Lenhossek, M . von,

'94. Beitra?e zur Histologie des Nepven sy s t eras und dep

Si nneso p?ane . Wiesbaden.

Der Peinere Bau des Nepvensy s te-ns, 2te Aufl. Beplin,

155

Leylig, P.

'52. Beitra^e zup Miikroskopische Anatotnie uril flntxick-
1 un^s^esch i ch t 9 der Rochen unl ier Haie. Leipzig.

Meyer, Adolf,

'99. Critioal Review of Recant Publications of Bethe and
Nissl .

Jour. of Compar. Neurology, Vol.9, Dp.58-4 5.

Meyer, Semi,

'96. Ueber eine Ve r b i nd an gs we i s e der Neuronen. Nebst

Mittheilungen liber die Technik und die Erfolge der
Methode der suboutanen Me t hy 1 enbl aui n j ec t i on .
Archiv f . mi k . Ana t . , Bd . 4 7, S. 734-748.

Miclucho-Maol ay,

'70. Beitrage zur vergleichenden Neurologie der Wirbel-
thiere. Leipzig.

Nissl, ?.

'94. Ueber eine neue Untersuchungsraethode des Central-
organs speoiell zur Peststellun? der Localization
der Nervenzel 1 en .
Cen tralb. f . Vervenh .und °sy ohi a t . , Bd . 5, 5.537-344.

Onuf,9., and Collins, Joseph,

'9 8. Experimental Researches on the Localization of the
Sympathetic Nerve in the Spinal Corl and Braii,
and Contributions to its Physiology.
Jour. Nerv. and Mental Diseases, Vol.85.

Osborn, H . <? .

'88. A Contribution to the Internal Structure of the
Amphibian BraiT.
Jour. of Morphol . , Vol .2, pp. 51-96.

Parker, CJ.H.

1300. The Neurone Theory in the Light of Recent Discov-
eries .

Amer. Natural i s t. Vol. ^4, pp.4S7-470.

Rabl-Ruckhapl, H.

'83. Das CJrosshirn ier Knochen f i sche uni seine Anhan^s-
?ebi Me.

Arshiv f. Anat.u.Entwick, Anat.Abth.,Jahr^.l'38 3,
S. 279-322.

'93. '93. Der lobus olfactorius impar ler Selachier.
Anat . Anzei?., Bd . 8, S. 728-731,

'94. Oas Voplerhirn dep Cranioten.
Anat . Anzei?. , Bd .9, pp. 536-547.

Rara6n y 3a j al, S .

'89, Estpuotupa del l6bulo 6ptii50 de las aves y origen

ie Ids nervios feptious,

Revistpa triuestpal de Histol., No. 3 y 4.

'91a. Sup la stpuctupe de I'feoorce cferfebpale de quelques
mamfni f epes .
La Cellule, T,7, p. 125-176.

'91b. Sup la fine structape du lobe optique des oiseaux
et SLIP l'opi?irie pfeele des nepfs optiquss,
Intep,Monatschp.f.Anat. U.Physiol., Bd.8, Heft.9 u.lO.

'94, Les nouvelles id^es sup la structure du syste-ne

nepvaux chez 1 ' hornme et cbaz les vertfebpf's. Papis-

'9 6. BeitpaiJ zura Studium dep Medulla Oblongata. Deutsoh
von Bpeslep. Leipzig,

Rei ssne p, F .

'60. Beitpage zup Kentniss vora Bau des Fuc kenmapkes von
Petpotnyzon fluviatilis.
Apchiv f.Anat.und Physiol, 3,545-588,

157

Re t ai us, G .

'93. Die Cajal'sche 7.ellen der G ros 3 h i m p i ad e beitn

Mensohen uni den Sauffethioren.
Biol .'Jn t-jrauch. , N ,P ., Bd . 5 .

Rohon, J.V.

'77. Das Centralop^an des Nervensysteiis dep Selachiep.
Denksohr.d.Kaisepl . 4kad.l .Wissensoh., Wien, Math.-
nat. Classe, Bd.38, S. 43-108.

Sanders, Alfred,

'86. Contributions to the Anatouy of the Central Nepvous
System in Vertebrate Anitnals; PI a? i 0 s t oma t a .
Phi l.Transao. Roy. Society of London, Vol.177,
pp. 733-766.

Sargent, Porter E.

1900, Reissner's Fibre in the Canalis Centralis of Ver-
tebrates .

Anat , Anzei?., Bd .1 7, S. 33-44.

1901. An Apparatus in the Central Nervous System of Ver-
tebrates for the Transmission of Motor Reflexes
arising from Optic Stimuli. ,
Baltimore Meeting of the Amer . Mo rp hoi og , Soc i e ty .

Sauerbeck, Ernst,

'96. Beitrage zur Kentniss vom feineren Bau des Sela-
ohierhi rns .
Anat . 4nzeig ., Bd .1 2, S. 41-52.

Schaper, Al f red ,

'94. Die morphologische unl histologische Entwickelung
des Kleinhirns der Teleostier.
Morph. Jahrb ., Bd .21, S. 625-708.

'98. The Finer Structure of the Selachian Cerebellum,

(Mustelus vulgaris) as shown by Chrome-Silver Prep-
arations .

158

Jour. of Compar , Neu pol ogy, Vol.8, pp. 1-20.

'99. Zur Moppholo^ie des Kleiohirns.

Ana t . Anze i ? . , Erijanzun^shef t, Bd.l6, S. 102-115.

Stnrr, Strong, and Lea-nin?,

'96. Atlas of Nerve Cells. New York.

St led a, Lud* i?,

'73a. Deber den Bau des Ruckenmarkes der Rochen und der
Haie.
Zeitsch.f.i»iss.Zool.,Bd.2 5, S. 435-442.

'73b. Ueber die Deutan? der einzelnen Theile des Pisch-
?ehi rns .
Zeitsch.f.»iss.Zool.,Bd.23, S. 443 -4 5 0.

St ron?, 0. S .

'95. The Oranial Nerves of Amphibia. A Contribution to
the Morphology of the Vertebrate Nervous Systen.
Jour. of Morphol ., Vol .10, pp. 101-230.

'96. Review of the Golgi Method.

Jour. of Compar . Neurology, Vol.6, jjp. 101-127.

Studni^ka, P.K.

'94a. Zur Losung eiiiger Pragen aus der Morphologie des
Vorderhirnes der Cranioten. ( Vo rl auf . Mi t t h . ) .
Anat. Anzeig., Bd9, S. 307-320.

'94b. Eine Antnort auf die Benoerkangen R . Buro k hard t ' s

zu meiaer vorlaufigen Mittsilung uber das Vorder-

hirn der Cranioten.

Anat . Anzeig . , Bd . 9, 5.691-693,

Beitrage zur Anatoroie und En t « i ck 1 un g s gesch i o h te
des Vorierhirn der Cranioten. Abth.l.
Sitzungsb.d.konigl.boehmisch. Resell. d.Wissensch.;
Math. -Nat. Classe, No. 33, S.1-42.

'98. Noch eini?9 Xopte zu meiaen Ab h an1 1 j n i^an uher die
Anatoraie das Vopdaphirns.
Anat . Anzeig . , Bd .14, S. 561-568.

Szozawinska, H .

'98. Rechepches sup Is systeme nepveux des Selaci^ns.
I. Stpuctupe des cellules nepveuses.
Archives de Biologie, T.15, p. 463-509.

Thopndike, E.L.

'99. A Note on the Psychology of Pishes.
Amep.f^at uralist, Vol.33, pp. 923-925.

Vepatti, E.

'97. Uebep eini?e Stpuctureigenthumlichkeiten dep Hipn-
rinde den Saugethiepen.
Anat. Anzeig., Bd.l3, S. 379-389.

Vepwopn, Max,

1900. Das Neupon in Anatomie und Physiologie. Jena.

Viault, Pp.

'76. Rechepches histologiques sup la struct u re des Gentpes
nepveux des PI agios tomes .
Apohiv.de Zool . expep . e t gen.,T.5, p. 441-525.

WappiTgton, W.B.

'98. On the Stpuctjpal Altepations Observed in Nerve
Cells.
Joup.of Physiol .. Vol .23, pp. 112-129.

Wolteps, M.

'90. Dpei neue Methoden zup Mapk- und Ac hsen cy 1 i n i e p-
f arbung .
Zeitsoh.f.wiss.Mikp., Bd.7, S. 466-473.

Section XI.
Description op the FnuREs.

All of the figures have been dra«n with the ail of the
camera lucila. A uniform scale of magnification coull not be
aioptel because of the wiie ran^e in the sizes of tha neurones.
In order that co-nparisons -nay be facilitated, the scale of
diamettTs of the drawing as reproduced is ?iven in the descrip-
tion of each figure. The orientation is accurately nnainlBined
in all of the drawings, the dorsal parts being uppermost as the
figure appears en the plate.

REP&PKNCE LETTERS.
a.f.-- Ascending fibre
a.], .f.-- Acustico-lateral fibre
aq.-- Aqueduct of Sylvius
ar.-- Terminal arborization
as. p.-- Associative neurone
as t . -- As t rocy te
ax.-- Axone
ax.h.-- Axone-hillock
b.f.-- Basal nerve-fibre
bv . -- Blood vessel
cb . -- Oerebel 1 um
cb.cr.-- Cerebellar crest

cb.inf.-- Inferior lobe of the cerebelluir
c.f.-- Communis filire
c.g.m.-- Central gray matter
ch . -- Chi asma

(160)

es t r .
ex t . -
f b.--
f . c . -
f .1 .d
f . s . t
? . c . f
e . c . n
?.l .-
hy.—
Ill .-
i .1 .-

inf .-
i ri f . T
i rj t . -
IX,--
i . 7 . -
1 . i .-
\ .V .-
1 . v? .
mb . —
ir . f . -
m.t'.-
rr.l .-
IT . n . -

31. S . t

n.?.-
n . gen

1«1

- Commissural neurone

.m.p.-- Decussation of the stratum medullars profundutr

- Deeper neuron es

- External division of the lateral ramus of the stratunc

medullare profundum

- External division of the median ramus of the stratum

medullare profundun
Ependyma
-- Epistriatum

- Limitans externa
Pore brain

- Pasci cuius communis

.-- Fasciculus lon^itudinalis dors&lis
.-- Pihre of the tr actus strio-thalarr, icus
.-- General cutaneous fibre
.-- iSeneral cutaneous nucleus

- Granular 1 ayer
Rypophy s i s

- The oculomotor ius

- Internal division of the lateral ramus of the stratum

anedullare profundum

- Internal division of the median ramus of the stratum

medullare profundum

- Infundibulum

. -- Ventricle of the infundibulum

- Limit ans in terna
fhe glossopharyngeus

- Inner zone

- Lobus inferior

- Lateral ventricle
-- Lobus vagi

Midbrain

- Median nerve-fibres,

- Vedian nerve-fibres,

- Molecular layer

- Middle neurones
. -- Median bundle of the tractus s t r io-t hal am i cus

- Neuroglia
.-- Nucleus ?eniculaturt

longitudinal bundle
transverse bundle

n.h.--

n. in .

n, IV.-

np.--

n . po . -

n . s . g .

obi .--

ol .f .-

ol.v.-

op. f . -

p.c.--

p.e.--

P.I.—

p . t. . r .

p. v.--

r .1 at .

r . n . --

r .n . t .

s.an.p.

s . n . —

sp.V.-

str.--

t . a. --

t . c . p .

t . c . s .

til.--

t . n . —

t r . c- b .

tr.p.-

v.m.f .

V , m . D .

162

Nucleus habenulae
-- Nucleus of the ocul omo tor i us

- Nucleus of tbe trocMearis
Recessus neuroporicus

- Nucleus pos tol f 8C tor i u s
-- Nucleus strati grisei

Oblongata

- Olfactory fibre

- Olfactory ventricle

- Optic fibre
Pall i al com-ni ssure
Pall i al eani nence
Layer of tbe neurones of Purkinje

- NeuroDe of the tractus pallii
Ventricle of the pallial eminence

X.-- Ramus lateralis vagi

Roof-nucleus of the midbrain
-- Tract of the midbrain roof-nucleus
-- StratuiE medullars profunduor.
Superficial neurones

- Spinal V tract
General striatum
Tuberculum acusticum

-- Tela choroidea posterior
-- Tela choroidea superior
Thai amus

Tract-neurone
s.-- Tractus cerebello-spinalis

- Tractus pallii
-- Vi scero-mo to r fibre
-- Viscero-mo tor nucleus

PLATK I,

Pig.l. Median sagittal section through the brain of an
adult Mustelus. The right lateral ventricle of the fore brain
is indicated in broken outline. Natural size.

Pi?.?. Composite transversa Eection of tbe oblongata at
the level of the IX nerve. The Golgi Tethod. Outline * 14.

Pi?.?. Structural elements froiri the formatio reticularis
on the ri?ht side of the oblongata. The commissural neurone
(c.n.) sends its axone across the iredian raphe. Tbe upper
neurogliar cell (ng.) has its longer axis extended in the radius
of the oblongata. The Golgi nretbod, * 230.

Pig. 4. Two neurones frcm the lobus vagi of the oblongata,
and a communis fibre (c.f.) having its terminal arborization
near a cell not drawn. The axones are directed into tbe deeper
nervous matter. The Golgi method, x230.

Pig. 5. Oblongata; a small area from the substantia gelat-
inosa of the general cutaneous nucleus. The minute neurone (a)
has a profusely branching axcne, w i t ti which the dendrites of (b)
interlace to form a socplex tangle; in this, the genei-al cuta-
neous fi>ire (g.c.f.) has its termination. The axones of (b) and
(o) penetrate the deeper levels of the nucleus. The Golgi
method, x 230.

Pig. 6. Oblongata; a neurone from the deeper part of the
left general cutaneous nucleus. A general cutaneous fibre
(g.c.f.) is seen breaking up into a terminal arborization.
Tbe Golgi method, x 230.

Pig. 7. Oblongata. Two Purkinje neurones from the cere-
bellar crest of the tuberculum acusticum. An acustico-lateral
fibre (a.l.f.) is seen terminating near the one on the right.
The Golgi method, " 230.

Pig. 8. Oblongata; ependymal fibres from the outer level
of tbe general cutaneous nucleus. The Golgi method, " 250.

Pig. 9. Bpendymal fibres from the ventral oblongata,
forniing a bundle ic the formatio reticularis. The Golgi ffethcd,
X 230.

Pi?. 10. Ependyma from the lobus va?i of the oblongata
The Golfji method, x 230.

Pig. 11. Heuroglia cell from the general cutaneous nucle-
us of the oblongata. The Golgi irethod, » 230.

Pig. 12. Sagittal section through the entire cerehellutn
showing its folds, the form of its ventricle, its neurone-layers,
and its principal masses of nerve-fibres. The Wolters method.
Outline x 8.

Pig. 15. Neurone of Purkinje from the cerebelluir. The
Golgi method, " 230.

Pig. 14. Neurone from the molecular layer of the cerebel-
lum. The greatest extension of the dendrites is parallel with
the surface of the cerebellar fold. The Golgi method, x 230.

Pig. 15. Three representative neurones from the granular
layer of the cerebellum. The intervening cell-bodies are omit-
ted for the sake of clearness. The axones are cut across at
the juncture of the granular with the molecular layer; see
Pig. 16, and the description in the text. The Golgi method,
« 230,

Pig. 16. Axones in the molecular layer of the cerebellum
passing transversely across the organ, derived from the neu-
rones of the granular. Several ascending axones are to be
seen just previous to their T-shaped division. The Golgi
method, x 230.

Pig. 17. Neurone of the Golgi II type from the granular
layer of the cerebellum. The Golgi method, * 230.

Pi?. 19. Neuroffliar cells froir. tbe cerebellum. The cell-
bodies lie between the neurones of Purkinje. An astrocyte is
seen at (ast.); and a Bergmann's fibre at (bg.f.). The Golgi
method, " 230.

Pi?.?0. Transverse section of the midbrain. The Kolters
method , "14.

Pig. 21. Transverse section of the entire thickness of the
left optic lobe, showing the neurones of the tectum mesenceph-
ali. The Golgi method, * 46.

PLATE III.

Pig. 18. Ependymal elemsnts from the cerebellum. Por the
sake of clearness, only a few of the fibres proper to the region
have been represented. The Golgi method, " 2 5 0.

Pig. 22. Neurone from the middle layer of the tectum mes-
encepbali. The Golgi method, * 300.

Pig. 23. Neurone from the deeper layer of the tectum mes-
encephali. The Golgi method, x 230.

Pig. 24. Interbrain. Transverse section through the left
thalamus at the level of the chiasma. The dotted lines indi-
cate the extent of the thalamic nuclei. The Golgi method.
Outline " 30.

Pig. 25. Interbrain; a neurone from the nucleus habenulae.
The Golgi netbod, " 230.

Pig. 26. Interbrain. Two neurones from the right side of
the infundibulum, together with the structures adjacent to then
The Golgi method, " 46.

?i^ .27 . Inter brain. Transverse section of tLe ri?ht
lobus inferior. The Gol^i uietbod. Outline " 30.

Fi?.30. The inner half of an ependysial eleirent from the
hypothalamus. The Gol?i method, * 150.

PLATF; IV.

Fi?.?8. Neurone of the hypothalamus from the ri?ht lobus
inferior. The T-shaped branching lies parallel with the lim-
itans externa. The Solfji method, " 230.

Pig. 29. Neuro?liar element from the nucleus habenulae.
The Gol?i method, " 230.

Pic. 31. CoTiDosite transverse section of the forebrain.
The 3ol?i methcd. Outline x 9.

Pi?. 32. A small area from the left epistr latum. The
neurones send their axones into deeper levels. The 3ol?i
method, X 46.

Pi?. 33. Neurone from the right epistriatum, its axone
passic? into the general striatum. The Golgi method, x 250.

Pig. 34. An area from the left general striatum to show
the arrangement of the neurones. The Golgi method, x 45.

Pig. 35. Neurone from the general sti-iatum. The Golgi
method, x 150.

plate: V,
'ig.56. A small area of the nucleus postolfactorius to

187

sho« the character and arrftn^ement of its neuroneE, together
with the termination of an olfactory fibre. The Go\P,i method,
« 250.

Pi?.?7. Two neurones fro* the nucleus n eu ropo r i cus . The
section was taken in the sagittal plane. The Golgi method,
" 230.

Pi?. 38. A group of neurones from the right pallial em-
inence. The Golgi nethod, " 67.

Pig. 39. Neurone from the right pallial eminence. Its
axone finally enters the tractus pallii. The Golgi method,
X 230.

Pig. 40. Neurone from the right pallial eminence sending
its axone into the pallial commissure. The Golgi metbod, x 230,

Pig. 41. A neurone of Cajal from the left pallial eminence,
The Golgi method, x 230.

Pig. 42. Ependyma from the left striatum. The fibres
really extend entirely to the limitans externa, but only the
inner portion has been represented. The Golgi method, " 230.

Pig. 43. Neurogliar element from the pallial eminence.
The Golgi method, " 230.

PLATE VI.

Pig. 44. A tract- neurone lying just to the left of the
median raphe of the oblongata, sending its axone into a tract
of the opposite side. The Nissl method, * 900.

Pig. 45. Commissural neurone from the right side of the
oblongata; its axone passes to the left across the median raphe.
The Nissl method, » 1120.

Pi?. 46. Neurone from the right lobus vapi
method, « 1120.

The Nissl

Fig. 47. Neurone from the right viscero-motor nucleus of
the oblongata. The axone passes dorsal to the fasciculus com-
munis on its way to the fasciculus longi tud in al i s dorsalis.

The Nissl nethod, " 900.

Fig. 48. Oblongata, Two of the deeper neurones frow the
right general cutaneous nucleus just dorsal to the lobus vagi;
they lie between the bundles of the spinal V tract, indicated
in cutlir. e. The Nissl method, " 1120..

PLATR VII.

Pig. 49. Two Purkinje neurones from the cerebellum. The
Nissl aiethod, " 1120.

Pig. 50. Two neurones from the molecular layer of the
cerebelluai. These lie unusually close together. The Nissl
aethod, « 1120.

Pig. 51. Pive neurones from the granular layer of the
cerebelluir. A nerve-fibre is seen ascending between them to
the molecular layer. The Nissl method, " 1120.

Pig. 52. Neurones of the superficial and middle layers of
the tectum mesen cep hal i . The Nissl method, * 900.

Pig. 53. A group of the spindle-shaped neurones character-
istic of the deeper layer of the tectum mesencephal i . The
Nissl method. " 900.

Pig. 54. A group of the stellate neurones found at inter-
vals in the deeper layer of the tectum rresencepball. The Nissl
method. « 900.

Pig. 55. The region immediately dorsal to the aqueduct of
Sylvius, showing the disposition of the neurones of the roof-
nucleus. Methylen-blue, * 89.

Fig. 56. Neurone frorr the left half of the midbrain roof-
nucleus. The Nissl TetLod, " 900.

Pig. 57. Neurone from the roof-nucleus of the^midlrain,
right s id e, toge t her with a part of the stratum medullare pro-
fundus. Two fibres are seen to eirerge from the stratum to end
in arborizations near the neurone. Iron haematoxylin staining.
Outline x 496; details from the homogeneous immersion.

PLATE VIII.

?ig.58. Midbrair. Sagittal section through the nuclei
of the oculomotor and trochlear nerves. Methylen-blue, * 89.

Pig. 59. Neurone from the nucleus of the III nerve. The
tigroid masses ir. the region of the nuclear membrane are re-
markable for their large size. The Nissl method, * 900.

Fig. 60. Neurone from the nucleus of the IV nerve. The
Nissl method, x 1120.

Pig. 61. Ependyma from the optic lobes. The ependymal
fibres are represented to the boundary of the central gray
matter, only; see Pig. 21 for the entire fibre. Combination
of iron haematoxylin and methylen-blue staining, * 1120.

Pig. 62. Interbrain. Two neurones from the nucleus strati
grisei of the thalamus. The Nir. si method, " 1120.

Pig. 6 3. A characteriftic group of neurones from the nucleus
geniculatum of the thalairus. The Nissl method, x 1120.

Fig. 64. Two neurones froir. the left lob us inferior. Their
dendrites are directed toward the limitans externa. The Nissl
method, *■ 1120.

Pi?. 65. Porebrain. A cfroup of three neurones from the
epistrifttum. The Nissl nethod, x 11<;0.

FiS.66. A neurone of the general striatum. The Nissl
method, " 112 0.

Pi?. 67, A ?roup of five neurones from the nucleus
postolfactorius. The Nissl method, * 1120.

Pi?. 68. Neurones from the left pallial eminence. The
neurone of the traclus pallii (p.t.n.) shows an axone-hillock.
The smaller neurone (as.r.) is of the associative type. The
Nissl method, *■ 1120.

Pi?. 69. Cajal neurone from the lateral part of a pallial
eminence. The Nissl method, x 1120.

LIKE.

I, Gilbert Logaa Houser, was born on an Tos-d fam in 136*?.
My preparatory training /ras given by several academies ani col-
leges not far from home, ^-ntering the University of Iowa, 1
pursued an undergraduate course in that institution for four
years, receiving the degree Bachelor of Science in 1891. A
year was subsequently spent in graduate study at the University
of Iowa, leading to the degree Master of Science, 1B92. My
graduate studies have since been continued in connection with
the University of Chicago and the Johns Hopkins University.

My earlier investigations were chiefly in the fields of
geology and paleontology. The dissertation for the Master's
degree was a critical study of the "Genera of Paleozoic Corals
of the Order Madreporaria". Later, the Iowa Geological Survey
comrnissioned me to examine certain building-stones and lime-
burning rocks of eastern Towa; my report on this subject is
published in the Report of the Iowa Geological Survey, Vol.1.
Receiving an appointment as Instructor of Biology in the Uni-
versity of Iowa, my studies have since flowed entirely in biolog-
ical channels.

A biological expedition to the 'fest Indies of tnree months,
and four season's work at the Marine Biological Laboratory of

(171)

172

Woods Holl have provided both the opportunity ani the inspira-
tion for the pursuit of numerous researches. Two papers have
be^n published as the result of these studies: -- "The 'Jses of
Por-naliehyle in Animal Morphology", and "The Nerve-Cells of the
Shark's Brain".

Since my appointment as an Instructor in 1^92, the Univer-
sity of Iowa has ^^radually advanced me to positions of larger
and larger responsibility, culminating in my election to the
Professorship of Animal Morphology and Physiology in June 1897,
a position which I still hold. At my urgent request, I was
granteJ leave of absence in September 1900 for a year of grad-
uate study. I have chosen to spend tnat year at the Johns
Hopkins University.

I hold membership in The American Morphological Society,
The American Society of Matjralists, and The Iowa Academy of
Sciences.

Baltinore, May, 1901.

^v*' •w*vwvv»i,^,,

>v^^*^wi^^'

■^tm^Mf^^u*.

'^^^v:.>'uwv^^ww^

'^^^^W«^^^'-'^&fc

'*^^u^v;,^v

.^#s»«»«i

/u^^^^Vw^^g^^wwlJg^

iE^^^^^is^i::^^^^**^^^^^^

'^-wvA;''V>rt*www'"*yu«

;>w-w-ww/vw'w,v^--,

iujW^ *^«W;uuvui

%ww*v»

yvwy^Vj^v ^,*v^vv^

^2a*I^P»

H^^iJW^*^

SS^^;::u^^e^^

.<^1^^.

I'^^'U'J^

;-M;vviJi^i^iy.iXSuw>

^v'y^^^«i''''..V

-'W^UOui^^

viivv«si«'

¥^^

'sJU^wW^^W"'

,w-^^!^-

'.nP^

'/^'-.i-'

■«rf»»»«

-vi^V^^^Wwwwwi^

.W**V**rv,

iijCip^/