Journal of Comparative Neurology 20 (1910)

From Embryology
Embryology - 6 Dec 2021    Facebook link Pinterest link Twitter link  Expand to Translate  
Google Translate - select your language from the list shown below (this will open a new external page)

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

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


Historic Journals: Amer. J Anat. | Am J Pathol. | Anat. Rec. | J Morphol. | J Anat. | J Comp. Neurol. | Johns Hopkins Med. J | Proc. Natl. Acad. Sci. U.S.A | J Physiol. | Ref. Handb. Med. Sci. | J Exp. Zool. | Yale J Biol. Med. | Anat. Anz. | Memoirs of the Wistar Institute of Anatomy and Biology | Quart. Rev. Biol.
Internet Archive  

THE JOURNAL OF COMPARATIVE NEUROLOGY AND PSYCHOLOGY

EDITORIAL BOARD

Henry H. Donaldson Adolf Meter

The Wistar Institute Johns Hopkins University

Herbert S. Jennings Oliver S. Strong

Johns Hopkins University Columbia University

J. B. Johnston John B. Watson

University of Minnesota Johns Hopkins University

ROBERT M. YeRKES, Harvard University

C. JUDSON HeRRICK, University of Chicago Managing Editor


VOLUME 20 1910

PHILADELPHIA, PA. THE WISTAR INSTITUTE OF ANATOMY AND BIOLOGY

CONTENTS

No. 1. FEBRUARY

Donaldson, Henry H. Further observations on the nervous system of the American leopard frog, Rana pipiens, compared with that of the European frogs, Rana esculenta and Rana temporaria. With two text figures. 1

Hatai, Shinkishi. On the length of the internodes in the sciatic nerve of Rana temporaria (fusca) and Rana pipiens: — being a reexamination by biometric methods of the data studied by Boycott, 1904, and Takahashi, 1908. With three text figures 19

No. 2. APRIL

Charles Brookover. The olfactory nerve, the nervus terminalis and the preoptic sympathetic system in Amia calva, L. With thirty-five figures 49

Henry H. Donaldson. On the percentage of water in the brain and in the spinal cord of the albino rat. five figures 119

S. J. Holmes. Pleasure, pain and the beginnings of intelligence 145

No. 3. JUNE

Otto C. Glaser. The formation of habits at high speed, With two figures 165

I. W. Blackburn. On the median anterior cerebral artery as found among the insane. With six figures 185

Hansford MacCurdy. Degeneration in the ganglion cells of the crayfish, ( Y'mbarus bartonii gir. With nine figures 195

Kuntz A. The development of the sympathetic nervous system in mammals. (1910) J Comp. Neurol. 20: 211-258. With eighteen figures 211

H. H. T. Jackson. The control of phototactic reactions in Hyalella by chemicals 259

No. 4. AUGUST

Prentiss CW. The development of the hypoglossal ganglia of pig embryos. (1910) J Comp. Neurol. 20(4): 266-282. eight figures 265

Kuntz A. The development of the sympathetic system in birds. (1910) J Comp. Neurol. 20: 211-258. Ten figures 283

Landacre FL. The origin of the sensory components of the cranial ganglia. (1910) Anat. Rec. 4(2): 71-.

F. L. Landacre. The origin of the cranial ganglia in Ameiurus. Eightyeight figures 309


No. 5. OCTOBER

C. Judson Herrick. The morphology of the forebrain in Amphibia and Reptilia. Eighty-four figures 414

No. 6. DECEMBER

Karl T. Waugh. The ro1e of vision in the mental life of the mouse. From Bel6it College. Ten figures 549

Lawrence W. Cole. Reactions of frogs to chlorides of ammonium, potassium, sodium and lithium 601

Sergius Morgulis. The movements of the earthworm. Two figures 615

Editorial Announcement 625

Further Observations On The Nervous System Of The American Leopard Frog (Rana Pipiens) Compared With That Of The European Frogs (Rana Esculenta And Rana Temporaria)

Henry H. Donaldson

Professor of Neurology at The Wistar Institute

WITH TWO FIGURES

In a paper under the general title given above (Donaldson '08) I discussed some observations made in 1904 on R. esculenta at Zurich and R. temporaria at Liverpool.

On comparison with the American frog, R. pipiens, it was found that although the European species were very similar to the latter in form and proportions, nevertheless the weight of the central nervous system was significantly smaller in the European species, and in the case of R. esculenta, the number of medullated fibers in the spinal nerves was much less than in R. pipiens.

These observations made it possible to correct the records of Fubini ('81) on the weight of the brain and spinal cord, which had alone been available for the European forms, and to call attention to the possible bearing of the anatomical differences on physiological results obtained from the two European species on the one hand, and the American species on the other.

In view of the fact that on the basis of rather few observations I had ventured to designate Fubini' s records as untrustworthy, and also to suggest possible physiological differences in the responses of the central nervous system, it seemed desirable to repeat the observations on the European forms.

This I did during the past summer. For a second time I am indebted to Professor Gaule for the hospitality of the Physiological Institute at Zurich, where I had examined R. esculenta in 1904, and to Professor Sherrington for similar privileges at the Physiological Laboratory of University College at Liverpool, where I had examined R. temporaria in the same year.

To both these gentlemen I desire to express my obligations for their courtesy and aid.

The results of these latest observations support completely the conclusions based on the records of 1904.

In the present communication therefore it is not necessary to repeat the entire argument of the earlier paper, but merely to present the evidence for the similarity of the earlier and later records.

For this purpose it will be desirable to print in full only the original measurements for both years, while the important ratios can be given in condensed tables accompanied by a few charts.

The following are the tables of the principal measurements as made on the three species in 1904 and 1909


TABLE 1 Data on R. pipiens, Chicago 1904- 12 specimens


BODY WEIGHT


SEX


TOTAL

LENGTH IN

MM.


BODY LENGTH IN


WEIGHT IN GRAMS OF


PERCENTAGE OF WATER


IN QRMS.


MM.


c


N. S.


Brain


Sp. C.


Brain


Sp. C.


11.6


M.


130




0918


.0666


.0252


84.4


79.4


16.0


M.


150




1148


.0796


. 0352


85.2


80.7


17.0


F.


159




1054


.0714


.0340


84.0


80.6


20.8


M.


170




1232


.0844


.0388


85.2


81.6


22.5


M.


162




1165


.0807


.0358


84 . 5


80.4


26.4


M.


180




1372


.0946


.0426


S4.4


78.4


27.6


F.


179




1416


.1014


.0402


84.8


80.1


30.6


M.


180




1454


.0998


.0456


84.6


79.8


34.2


M.


190




1518


. 1056


.0462


85.6


81.6


41.8


M.


197




1652


.1146


.0506


86.9


82.2


43.9


M.


200




1708


.1210


.0498


85.8


SO. 7


47.0


M.


198




1664


.1140


.0524


84.4


SO. 5


NERVOUS SYSTEM OF THE FROG


TABLE 2 Data on H. esculenta, Zurich 190J,.. 11 specimen*


BODY WEIGHT


SEX


TOTAL BODY LENGTH IN LENGTH IN


WEIGHT IN GRAMS OF


PERCENTAGE OF WATER.


IN GEMS.


MM.


MM.


C. N. S.


Brain


Sp. C.


Brain Sp. C.


12.40


F.


131



.0818


. 0577


.0241


84 . 2 78 . 4


16.75


F.


144



.0926


.0634


.0292


83.4 ' 79.1


18.43


F.


144



. 0928


.0650


.0278


83.2


78.2


20.00


F.


161



.1103


.0756


. 0347


82.5


79.2


22.00



164



.1107


.0769


.0338


84.0


79.0


24 . 10


M.


167



.1217


.0841


. 0376


83.4


7S.4


33. S5


M.


175



. 1327


.0895


. 0432


83.2


78.2


36.30


M.


177



.1478


.1004


. 0474


83.4


78.6


37.56


F.


188



. 1490


.0993


. 0497


82 . 9 ! 7S . 8


37.96


F.


194



.1427


.0953


.0474


82.8


77.8


45.03


F.


196



.1578


.1078


.0500


83.9


7S.4


TABLE 3 Data on R. esculenta, Zurich 1909. 11 specimens


BODY WEIGHT


SEX


TOTAL

LENGTH IN

MM.


BODY LENGTH IN MM.


WEIGHT IN GRAMS OF


PERCENTAGE OF WATER


IN GRMS.


C. N. S.


Brain


Sp. C.


Brain


Sp. C.


18.9


M.


143


57.3


.1047


.0707


.0340


S3. 6


78.6


24.7


F.


167


65.0


.1065


.0719


.0346


S3. 6


79.3


26.5


M.


167


63.4


.1120


.0737


.0383


83.6


77.8


30.9


M.


177


69.2


.119S


.0830


.036S


83.9


78.5


32.3


F.


183


68.0


.1301


.0S73


.0428


83.6


78.1


33.0


F.


184


70.5


. 1275


.0845


.0430


83.4


77.0


35.5


F.


18S


72.5


. 1435


.0985


.0450


83. S


78.2


47.4


F.


204


80.0


.1593


.1063


.0530


83.1


79.4


48.4


F.


193


79.3


. 1545


. 1027


.0518


83.8


77.7


52.4


F.


205


82 .


.1589


.1105


.0484


83.6


79.1


58.0


F.


216


S7.0


.1858


.1278


.0580


83.6


77.6


HENRY H. DONALDSON


TABLE 4 Data on R. temporaria, Liverpool 1904


12 specimens


BODY WEIGHT


SEX


TOTAL

LENGTH IN

MM.


BODY LENGTH IN


WEIGHT IN GRAMS OF


PERCENTAGE OP WATER


IN GRMS.


MM.


C. N.S.


Brain


Sp.C.


Brain Sp. C.


14.05


F.


144



.0881


.0596


.0285


82.3


78.2


16.10


F.


151



.0991


.0690


.0301


82.7


79.0


17.65


M.


154



.0916


.0618


.0298


83.0


78.5


21.75


M.


171



.1045


.0671


.0374


82.8


78.2


23.45


M.


162



.0947


.0628


.0319


82.1


77.0


24.17


F.


173



. 1333


.0864


.0469


81.9


76.5


27.05


M.


173



.1298


.0874


.0424


82.4


77.1


28.15


M.


168



.1018


.0687


.0331


82.5


76.7


28.95


M.


174



.1324


.0813


.0511


81.3


76.8


28.95


M.


178



.1485


.0928


.0557


81.3


76.8


32.15


M.


173



. 1321


.0890


.0431


80.9


78.6


32.81


F.


178



.1161


.0766


.0396


82.7


78.0


TABLE 5 Data on R. temporaria, Liverpool 1.909. 16 specimens


BODY WEIGHT


SEX


TOTAL

LENGTH IN

MM.


BODY

LENGTH IN

MM.


WEIGHT IN GRAMS OP


PERCENTAGE OF

WATER


IN GRMS.


C. N. S.


Brain


Sp.C.


Brain


Sp.C.


14.5


F.


148


53.3


.0772


.0522


.0250


82.2


76.8


17.2


M.


154


56.0


.0903


.0598


.0305


83.6


79.6


19.1


F.


155


58.5


.0907


.0621


.0286


82 . 4


76.3


21.0


M.


162


60.3


.1014


.0663


.0351


83.8


79.5


24.2


M.


176


64.5


.1099


.0736


. 0363


83.9


78.5


25.4


M.


164


60.0


.0994


. 0672


. 0322


84.5


78.2


26.0


M.


162


60.8


.1066


.0702


.0364


S4.0


79.4


26.1


F.


174


65.8


.1191


.0787


. 0404


84.0


70.4


26.9


F.


163


63.2


.1092


.0737


. 0355


84.1


76.3


27.9


F.


175


66.7


.1149


.0786


.0363


83.8


79.5


29.2


F.


174


66.5


.1114


. 0744


.0370


83.8


79.5


29.4


M.


170


66.0


.1356


.0887


.0469


83.8


78.9


29.8


M.


16S


62.5


.1167


.0751


.0416


83.8


78.9


32.1


F.


184


73.2


.1314


.0864


.0150


84.3


79.2


33.3


M.


167


64.0


.1373


.0900


.0473


84.3


79.2


39.1


F.


196


76.8


.1452


.0964


OISS


82.3


77.8


NERVOUS SYSTEM OF THE FROG 5

The foregoing tables (1-5), representing five series, contain the fundamental data.

The plan was to have twelve specimens in each series. In the case of R. esculent a 1904 and also 1909, there are, however, only eleven in each. The absent records were excluded because the percentage of water, which was not calculated until my return home, showed the excluded specimens to be in abnormal condition.

In the case of R. temporaria 1909 sixteen records were made. In general, the grouping of these data is by threes. There are however three exceptions: In R. esculenta 1904, with a total of 11 specimens, there is one group of two (Records 7 and 8) and in R. esculenta 1909, there is one group of two (Records 10 and 11).

In R. temporaria 1909 there is one group of four. In each case this departure from the rule is indicated in the condensed tables (6, 10, 12, 13,) by a bracketed number following the average for body weight.

It will be noted that in the 1904 series, the column under the heading " Body length" is vacant. This measurement was not made in that year, but was made in the specimens collected in 1909.

It represents the length of the frog from the tip of the nose to the tip of the urostyle, the skin over the urostyle having been split in order to expose its cartilaginous tip; the measurement being taken with vernier calipers.

In the previous paper (Donaldson '08) some measurements on preserved material were introduced without correction for the effects of the reagents used. These cases were explicitly noted. It is of interest to state therefore that, in this paper, the data apply to the fresh material only. Indeed all the measurements were made on the material when fresh except in the case of the leg bones of the two 1909 series. In these cases the legs were brought to this country from Europe in 60 per cent alcohol and then the bones were measured.

A long series of control observations on the legs of R. pipiens treated in the same way and for the same time have shown that


6


HENRY H. DONALDSON


this treatment reduces the length of the femur by 0.70 per cent making it 99.30 per cent of the fresh length, the tibia by 0.73 per cent making it 99.27 per cent of the fresh length; the foot (tarsus-pes) by 1.54 per cent making it 98.46 per cent of the fresh length.

These corrections were applied before the data were used in tables 7 and 8.

TABLE 6 Body weight per millimeter of total length. Averages from groups of three


R. pipiens.


BODY WEIGHT IN GRAMS


14.9 23.2 30.8 43.2


BODY WEIGHT PER

MILLIMETER, IN GRAMS

.102 .135 .168 .218


R. esculent a 1904.


15.9 22.0

35.0 [2]* 40.2


.114 .134 .199 .20S


R. esculenta 1909.


R. temporaria 1904.



.146 .177 .225 .262

.107 137

.162 177


R. temporaria 1909.


1S.0 [4]

25.2

26.9

29.5

34.8


116 151 159 173 191


(A) AVERAGE AMOUNT OF BODY WEIGHT FOR EACH MILLIMETER

OF TOTAL LENGTH


The general form of the specimens examined is obtained by dividing the body weight by the total length (table 6). The data in this table are given in Chart 1 and show that in the years


NERVOUS SYSTEM OF THE FROG




cr -• cr

• o



• ■


t- E < « 


Chart I. Showing the average amount of body weight corresponding to each millimeter of the total length.


8 HENRY H. DONALDSON

1904 and 1909, the European frogs were similar in their general form.

The records for R. pipiens are not entered on this chart. They would run a trifle below those for the European species, showing that R. pipiens was more slender in its general build. This character of R. pipiens taken alone would imply a slightly smaller nervous system, but as we know the contrary is the case.

TABLE 7 Percentage of the total length represented by the combined lengths of the leg bones

SPECIMENS PERCENT

R. pipiens 9 66.6

R. esculenta 11 65 . 1

R. temporaria 16 66 . 2

(B) PERCENTAGE OF TOTAL LENGTH REPRESENTED BY THE

COMBINED LENGTHS OF THE LEG BONES

The absolute values of the percentages in this table are on the average less by 3.5 than those given in the previous paper (see Donaldson '08, table 2). This is the result of a change in the technique of measurement. Previously the total length of the frogs was taken when the animals were suspended, and under this condition a certain amount of flexion persisted in the legs.

In the present case the frog was measured when stretched out on the table and lying on its ventral surface. By this treatment the amount of flexion was reduced, and the total length thereby slightly increased. This naturally reduced the percentage value of the sum of the lengths of the leg bones, the measurements of which were made in the same way in both cases. The above mentioned change in technique is the only one which has been made.

The point of importance is that the percentages are nearly the same for the three species which are here compared.

(C) THE PROPORTIONAL LENGTHS OF THE SEVERAL LEG BONES

These are shown in table 8 in which the 1904 records have been repeated and a complete series of 1909 records added. It will be seen that there is no essential difference between the obser


NERVOUS SYSTEM OF THE FROG


9


vations made at the contrasted dates, and that in both instances the proportional lengths are nearly the same for the three species compared.

TABLE 8 The proportional lengths of the several leg bones


NO. OP

SPECIMENS


FEMUR



per cent


12


25.5


19


26.0


5


26.3


12


26.8


6


26.1


16


25.7


FOOT (TARSUS AND PES)


R. pipiens 1904

R. pipiens 1909

R. esculenta 1904*. . R. esculenta 1909. . . R. temporaria 1904* R. temporaria 1909.


per cent 29.3 29.3 28 . 2 28.2 28.7 28.3


per cent 45 . 2 44.7 45 5 45 . 45.2 46.0


Leg bones from frogs of the socalled "Zurich series of 1898." These frogs had been carefully fixed in 4% formaldehyde and then preserved in 80"^, alcohol. The effect of this on the lengths of the several leg bones was not at the time determined. (See Donaldson '08, p. 127).


(D) PERCENTAGE VALUE OF THE LENGTH OF THE ENTIRE CENTRAL NERVOUS SYSTEM— THE TOTAL LENGTH OF THE FROG BEING TAKEN AS THE STANDARD.

In the case of this character we have grouped the 1904 data (see Donaldson '08, table 5) into three entries and added the measurements on the new material for the 1909 groups.

The table shows that the length of the entire central nervous system is slightly greater in the European species. As this excess in length is associated with a deficiency in absolute weight, it follows, as was previously noted (Donaldson '08, p. 128) that the nervous system in R. pipiens must exceed that of the European species in its transverse diameters.


10


HENRY H. DONALDSON


TABLE 9


Percentage mine of the length of the entire central nervous system — the total length of the frog being taken as the standard



AVERAGE

TOTAL

LENGTH IN

MM.


PERCENTAGE VALUE OF THE LENGTH OF THE ENTIRE CENTRAL NERVOUS SYSTEM



Rana pipiens


Rana esculenta


Rana temporaria


9 (1904)


152 155 159 167 171 176 181 182 195 196 210


17.5

16.7 16.3


18.4 16.9



4


17.6


3



3


17.6


6


17.3


9 (1904)



3



3


17.2


3


16.6



4 (1904)


16.2



2






(E) THE WEIGHT OF THE CENTRAL NERVOUS SYSTEM

Turning now to the main character under consideration, the weight of the central nervous system, the condensed records are presented in table 10.

When these data are put in the form of a chart, (chart 2) several interesting relations between the observations of 1904 and those of 1909 at once appear. In the first place the later records follow the same line as the earlier; second, the record for each species in 1909 is somewhat less than in 1904, and as a consequence still further below the records of 1904 for R. pipiens. This result serves to establish the main conclusion, namely that R. pipiens has a heavier nervous system than either of the European forms. The fact that the values for the weight of the central nervous system in the European species as determined in 1909 are less than those determined in 1904, calls for a word of comment.

Some unpublished studies which are being made on R. pipiens at the Wistar Institute relative to the change in the weight of the central nervous system with season, indicate that in this species the greatest weight is attained about the end of July.


NERVOUS SYSTEM OF THE FROG


11


WEIGHT OF C. N. S,



• R. ESCUL. 1904

O " " 1909

■ R. TEMP. 1904

□ " " 1909


BODY WEIGHT g« 


Chart 2. Showing the weight of the entire central nervous system.


12 HENRY H. DONALDSON

TABLE 10 Weight of the central nervous system in grams. Averages from groups of three

WEIGHT OF BODY WEIGHT CENTRAL,

NERVOUS SYSTEM

14.9 .1040


R. pipiens < oq o


23.2 .1256

.1463


R. esculenta 1904.


R. esculenta 1909.


R. temporaria 1904 < 9 o q


43.2 .1674

15.9 .0890

22.0 .1142

35.0 [2] .1402 40.2 .1498

23.4 .1077

32.1 .1258 43.8 .1524

55.2 [2] .1724

C 15.9 .0929

23.1 .110S

.1213 [31.3 .1323

f 18.0 [4] .0899

1 25.2 .1053

R. temporaria 1909 i 26.9 .1144

29.5 .1212 [34.8 .1380

If this observation applies, as it probably does, to the European species, then the differences in weight as shown in chart 2 are susceptible of the following explanation :

The esculenta of 1904 were examined August 1 — 5, when it may be assumed that the nervous system of R. esculenta had attained approximately its maximal seasonal weight. In 1909 the examination was from July 5-7, or some four weeks earlier. Under these circumstances, a somewhat smaller weight was to be expected, and the records show this.

The temporaria of 1904 were examined July 11 and 12, before the central nervous system had reached the maximum for the season.

In 1909 the examination was from August 17 to 21, or some


NERVOUS SYSTEM OF THE FROG 13

three weeks after the assumed maximum, and at a time when the seasonal weight has begun to diminish. Here the difference is less than in the case of the esculenta, but is susceptible of a similar explanation.

The relation of these two series of observations can be conveniently shown in still another way.

I have been able to point out (Donalsdon '02) that a fairly accuate determination of the weight of the central nervous system in frogs can be made by the formula

C. iV. S. = (Log. Bd. W. t}~L)C

where C. A r . S. is the weight of the central nervous system, Bd.W. the body weight in grams, L the total length in mm. and C. a constant to be determined for each species. Since publishing this formula I have found that the most convenient way of expressing seasonal variations on the weight of the central nervous system is by the variations in C.

Applying this method to the series before us, and remembering that the increase in the relative weight of the central nervous system is measured by the increase in C, and vice versa, we obtain the following:

TABLE 11 To show the values of " C" for each of the several series

AVERAGE QF

BODY WEIGHT

R. pipiens 1904 \

Average of 12 records /

R. esculenta 1904

Average of 8 records > 32 . 4 24 . 6

First " weight group " omitted J

R. esculenta 1909

Average of 9 records \ 33 . 1 23 .

Last " weight group" omitted J

Difference 1 . 6 R. temporaria 1904 \ Average of 12 records/ '

R. temporaria 1909 1 Average of 16 records) '

Difference 0.9


14


HENRY H. DONALDSON


As is to be seen by inspection of the foregoing table 11 the value of C for the 1904 records is greater in both the European species than for the corresponding 1909 records, and as noted above, the greatest difference (1.6) is in the case of R. esculenta.

In connection with this table a word of explanation is required. It has been found that there is a slight increase in the value of C as the absolute size of the frog increases. This is a relation previously overlooked, but which will be discussed elsewhere. The bearing of it on the present case is that in making a comparison of the values of C in any pair of records, it is necessary in order to get trustworthy results, to compare the determinations for frogs of approximately the same range in size. In the present instance this makes it necessary in the case of R. esculenta to omit the value of C for the first weight group of the 1904 series, because there is no corresponding weight group on the 1909 series, and similarly to omit the determinations for the last weight group of the 1909 series.

A glance at chart 2 will serve to supplement the explanation.

In the case of the records for R. temporaria, the values for C in all the weight groups of both years have been used in making up the averages. It is because of this influence of the absolute size that the average body weights for each series are entered in table 11.

All through the present paper the data on R. pipiens used in 1904 have been repeated without revision. In the former communication (Donaldson '08. pp. 132-133) it was noted that the weight of the central nervous system in the series of this species was low in comparison with other data which we had. This statement still holds good, but it was thought wiser to leave the standard as represented by 1904 records on R. pipiens unchanged at this time.

As evidence that the weights here used were low for this species, I give below two other series of determinations of C on Chicago frogs as follows :


NUMBER OF SPECIMENS


DATE ABOUT AUG. 1


AVERAGE BODY WEIGHT GMS.


AVERAGE VALUE FOR C.


IN


1902 1909


22.3

27 7


28 6 29.5


4


NERVOUS SYSTEM OF THE FROG 15

It will be seen on comparison with the value of C. for the series of R. pipiens here used (C = 26.2) that these are much higher. This implies an increase in the weight of the central nervous system proportional to the differences in the values of C after correction for the differences in body weights in the several series.

Why the particular series of pipiens used by me as a standard falls below that for the two other series is a point the discussion of which must be reserved for a future paper.

In this connection it is desirable to refer to one modifying condition affecting the value of C which has not heretofore been mentioned, and the data on which are still unpublished. I find that the value of C is not the same for specimens of R. pipiens from different parts of our own country. For example those coming from northern Minnesota give a value sensibly greater than that found for the socalled "Chicago frogs" and the specimens taken about Philadelphia give a value less than that found for the " Chicago frogs," as a rule, but almost identical with that of the series used as a standard in this paper.

R. pipiens extends much farther south in this country of course, being found both in Florida and Texas. What the relation of C may be in specimens from stations farther south than Pennsylvania, has still to be determined, but the possibility of variation in this character with latitude is a matter of much interest.

(F) THE RATIO OF THE WEIGHT OF THE BRAIN TO THAT OF THE

SPINAL CORD

Omitting the tabulation of the absolute values for the brain and cord, as these can be readily found in the full tables, I give below in table 12 a condensed statement of the ratios.

It will be seen that in both 1904 and 1909, that relative weight of brain (the value given under "ratio") is higher in R. esculenta than in R. temporaria, although the difference is not so great in the later as in the earlier records. Further, this ratio in R. pipiens is always greater than in either of the European forms.

Fiuallv it is to be noted that the ratios which I find for the


16 HENRY H. DONALDSON

European species are much higher than those determined by Fubini (see Donaldson '08, table 20) and so confirm my earlier conclusions concerning the untrustworthy character of his records.

TABLE 12 Ratios of the weight of the brain to the weight of the spinal cord. Averages from



groups of three








BODY WEIGHT


RATIO




f 14.9



2.33


R. pipiens




23.2 30.8



2.22





2.32




[43.2



2.28




[15.9



2.29


R. esculenta 1904



\


22.0 35.0


[2]


2.22





2.09




[40.2



2 . 05




f 23 . 4



2.03


R. esculenta 1909



J


32.1

43.8



2.09





2.05




[55.2


[2]


2.24




[15.9



2.15


R. temporaria 1904



J 23.1



1.S6




| 28.0

[31.3



1.87

1.87





'18.0


[4]


2.02





25.2



2.02


R. temporaria 1909




26.9



2.06





29.5



1.90





34.8



1.93


(G) THE PERCENTAGE OF WATER IN THE BRAIN AND SPINAL CORD

Table 13 shows the condensed results on the percentage of water. In my former communication I called attention to the differences in this character in the several species (Donaldson '08, p. 139.)

While the percentages of water in both the brain and spinal cord as determined for both European species in 1909 are less than that found in R. pipiens, they are nearly alike, and also similar to the 1904 determination for R. esculenta, so that it is


NERVOUS SYSTEM OF THE FROG


17


not desirable to give any weight to the differences as observed in 1904.

The value of this table as it stands is to show that we were dealing in all cases with healthy frogs, as the frog readily shows by changes in the amount of water in the nervous system, the effect of infections or disturbing conditions.

TABLE 13

Showing the percentage of water in the brain and in tJie spinal cord. Averages from

groups of three


R. pipieng.


BODY WEIGHT

[14.9 23.2 30.8

[43.2


PERCENTAGE OF WATER IN

Spinal Cord


Brain

84.5 84.7 85.0

85.7


80.2 80.1 80.5 81.2


R. esculenta 1904.


15.9 22.0 35.0 [2] 40.2


83.6 83.3 83.3

83.2


78.6 78.9 78.4 78.3


R. esculenta 1909.


R. temporaria 1904.



83.6 83.6 83.6 83.6

82.7 82.3 82.1 81.6


78.6 77.9 78.4 78.3

78.6

77.2 76.9

77.8


R. temporaria 1909,


f 18.0 [4]

I 25.2 i 26.9 I 29.5 [34.8


83.0 84.1 84.0 83.8 S3 . 6


78.1

78.7 78.4 79.1

78.7


The foregoing tables and the comments on them are intended to demonstrate that a second series of observations on R. esculenta and R. temporaria made in 1909 at an interval of five years yield results substantially similar to those first obtained in 1904


18 . HENRY H. DONALDSON

and that therefore so far as the conclusions of the earlier paper depend on the observations which have been repeated, they may be considered as well founded.

BIBLIOGRAPHY

Donaldson, Henhy H. Weight of the Central Nervous System of the Frog. 1902 Decennial Publications, University of Chicago, vol. 10, p. 3-15.

1908 The Nervous System of the American Leopard Frog, Rana pipiens,

compared with that of the European frogs, Rana esculenta and Rana temporaria (Fusca). J. of Comp. Neurol, and Psychol., vol. 18, no. 2, pp. 121-149. Fubini, S. Gewicht des centralen Nervensystems im Vergleich zu dem Korperge1S81 wicht des Thiere, bei Rana esculenta und Rana temporaria. Unter such. z. Naturl. d. Mensch, u. d. Thiere, Bd. 12.


ON THE LENGTH OF THE INTERNODES IN THE SCIATIC NERVE OF RANA TEMPORARIA (FUSCA) AND RANA PIPIENS: BEING A RE-EXAMINATION BY BIOMETRIC METHODS OF THE DATA STUDIED BY BOYCOTT ('04) AND TAKAHASHI ('08)

SHINKISHI HATAI Associate in Neurology at The Wistar Institute

WITH THREE FIGURES

In 1904 Boycott published an important paper on the length of the internodes under the title "On the number of nodes of Ranvier in different stages of the growth of nerve fibers in the frog." In this work he measured 5338 internodes from twenty-six frogs which ranged in body length from 14.5 mm. to 53.9 mm. Although his observations were limited to the fibers from the thigh taken at the division of the sciatic nerve into the n. tibialis and n. peroneus, they nevertheless give us definite information as to the lengthening of the internodes first in relation to the lengthening of the leg, and second in relation to the increase of the fibers in diameter.

Four years later (1908) Takahashi published a paper on the same subject under the title "Some conditions which determine the length of the internodes found on the nerve fibers of the leopard frog, Rana pipiens, " and in this investigation he examined several new points. Takahashi's observations were made not only at three different levels in the thigh, but also at levels in the shank and in the foot, as well as on some of the spinal nerve roots.

He measured altogether 3068 internodes from eight frogs mostly of a larger size than those used by Boycott. The body lengths of Takahashi's specimens ranged from 39.0 mm. to 89.4 mm.


20 SHINKISHI HATAI

The examination of the two interesting papers just mentioned, suggested several points for further study. It seemed desirable to determine first, how far the results obtained by the simple method of averages would agree with those obtained by the more elaborate statistical treatment of the data; and second, whether or not there is some definite law expressing the relation between the length of the internode and diameter of the fiber on which it occurs, both in different segments of the leg of the same frog, as well as on the same segment of the leg in frogs of different sizes.

In order to investigate the problems just mentioned, it was necessary to reexamine the original data used in the foregoing researches, and through the efforts of Prof. H. H. Donaldson, I was so fortunate as to get the generous permission of both Boycott and Takahashi to use their valuable original data for this investigation, and I wish to express here my thanks to all of these gentlemen.

METHOD OF TREATING THE DATA

As has been shown by Boycott and Takahashi, the length of the internode is highly variable, even on the same fiber, as well as on fibers of the same diameter, and therefore in order to get proper mean values in the case of Boycott's records, I have combined the data for several frogs, and out of the total of twentyfive frogs, made five groups according to the length of the sciatic nerve as determined by Boycott. 1

Group 1. Those with a sciatic nerve measuring from 15.5 mm. to 17.0 mm. : represented by 2 frogs.

Group 2. Those with a sciatic nerve measuring from 20.0 mm. to 22.5 mm. : represented by 7 frogs.

Group 3. Those with a sciatic nerve measuring from 24.0 mm. to 26.5 mm. : represented by 6 frogs.

1 The length of the sciatic nerve as defined by Boycott is the distance from the point of emergence from the vertebrae of the upper of the two larger branches of the plexus to the level of the nerve obtained by cutting across the leg through the knee joint when it is in full extension; loc. cit., p. 371.


LENGTH OF THE INTERNODES 21

Group 4. Those with a sciatic nerve measuring from 30.5 to 37.0 mm. : represented by 4 frogs.

Group 5. Those with a sciatic nerve measuring from 46.0 mm. to 53.5 mm. : represented by 6 frogs.

In the case of Takahashi's data, I found that the internodes had been measured at the lower end of the thigh in only four frogs, giving a total of 683 measurements. These frogs moreover differ so widely in body weight that the data are insufficient for biometric treatment. For this reason I have decided to analyze Boycott's data as completely as possible and then merely to compare the results obtained with Takahashi's conclusions.

I shall present my results in the following order :

I. A confirmation of Boycott's and Takahashi's conclusions together with my own.

II. Analysis of Boycott's data.

a. Analytical constants.

b. Frequency distributions.

c. Mean and standard deviation.

d. On the correlation in growth between internodal length and the diameter of the fiber.

III. Takahashi's observation on the length of the internode in different segments of the same frog.

IV. Correlation tables.

I. As my principal object was to find whether there is any definite law relating to the length of the internode and its diameter in different segments of the leg from the same frog, as well as in the same segment of the leg in different sized frogs, I shall not touch many other points discussed by Boycott and especially by Takahashi. Within the range of my examination, I confirm all their findings. Since the evidence of such confirmation will be found in the following pages, I shall present here only the main facts brought out by myself.

1. In a given specimen of Rana temporaria (fusca) whatever its size, the length of an internode and its diameter are positively correlated, though the correlation is not high; therefore it can be stated that the internodal length varies as the diameter. The degree of correlation increases as the frog becomes larger.


22 SHINKISHI HATAI

2. When however the lengths of the internodes for given diameters in different sized frogs are compared, the larger frog has for a given diameter longer internodes than the smaller frog. Thus in this case the internode varies according to the relation given by the following general exponential equation


V = Ae hx

where the constants A and h are to be determined from the observations, y is the internodal length, x the diameter and e the base of the natural system of logarithms.

3. The equation just mentioned expresses also the relation in these two characters in the different segments of the leg from the same frog.

4. Therefore the rate of the increment of the length of the internode following the increase in the diameter is proportional to the length of th? mternode itself, or in mathematical terms.

d y *,

dx = hV

This may be considered as the general formula which expresses the relation of the two characters in different segments of the leg of the same frog as well as in corresponding segments from frogs of different sizes.

II. Analysis of Boycott's Data

a. Analytical Constants

For future reference I shall present here the various values of the analytical constants as determined from Boycott's data.


LENGTH OF THE INTERNODES


23


e a


03

















^t^g


CC


©

CO


x co

















CO © ^,


OI



Ol i-H

















r* co o


X


©


© -HH


>

fa
















I> © O


-CI


©


© LO


co















i0> ^ jj


-H


-H


-H -H


p



lO


Nh^hNNMOS^'OONCC



Ol


©


co •*


o


©


©


ooooii— iMooconociK ©


-r © co


©


>o


^H -f


« 



oi


-f io oi oi co o co co -v oi x © -r


-h © Ol


X


•*


© CO


o



co


MNM>OIMOlTt<C100NCOnrO


H w oc


CI


CO


IC iO




CO


01 Ol O Ol "O O O CM CO © i-i o o


.-i o — i


^H


©


© i-H





1—1 1


co co







i



LO








1




LO







i



CO



CO

















1-1




r*

















££^3


©


Ol


r— t CC^ - —

















Tt< © CO


r


01


LO ©

















^H CO O


•*


©


© CO


>
















© CO ©


■<CH


o


© ©



lO















« ° -H


-H


-H


-H -H


EM

P


K


O


I>


o


© 1

X


© © © -H


tH i- CO iC


■* Ol ©


"O


C":


^3 I-H


CO


(N


CO i« C"


c


-t<©©X©©X©i~

Tt O lO


i>



Ol l>



CO


co lo i

o


© r



I—


© i— 1 © LO CO'


t^ © 01




X "Cj



co


o


CO -1

ir


© i—


cc r


© i— 1 CO i— 1 CO


i0 01 X


©


CO


Ol ^H




co


10) o


© CO © o c


n m o ■* o o


O O i-H


©


©


© ^H






1



■<*


H







i


^H CO


ci



i-H CO






1




T ~








I


©


CM



©

i-H

















i^^g



©


co n

© o

















© © l-H


/



Ol ©

















^n r- ©


-■


©


© X


"
















N (M O


^H



O CM


fa


iO>















Ol CO _|_|


-H


-H


-H -H


p


LO


i— i


~~


c


c


~


h- -^


pr


© Ol


©


CC


tH CO


NOW


CT:


©


CO 1^


o



©



o


CO



t^ pc


>


t^



l- r

CO t^


© iO ©



X


— t

K



©



t^


cc



cc


c


CO


i—



CO


CO ©


01 © ^H



■H

-f ©


a



Ol


X


10



co


T—


CM


•^


t^


CC


I—


CN


CO 01


© -r lo


©


1-1


-r S.




<N


OJ


I

10


-*


C


c


I—


CM


cc


© c


© o


© © ^H


,_|


©


X ©






1




©







1


©


©



©






1









1



1


t^


1-1



1^

















i^^g


'0

©


©


CM Ol

CO LO

















01 X ^


©



CO ©

















I-H CO O


CI


©


© 01


M
















CO l o ©


Ol


o


o co


fa
















-' -H


-H


HH


-H -H


P


o


lO


co


c


I—


CO


©


©


Ol


i—


>o


-t


r~


© CO


X CO ©


-*


N


co o


o



qo


1>


X


CO


r


a:


r^



CC


t^


©


CO


Ol 1^


N © CO


-*


-r


-t 1 t-

is



i

t^


©


re


i

o



o


-t


CN


r



© -tl


© r. co


©


©


x co


o



i

Tt


T—


LO


c


©


c


cc


10


cc


c


CC


© CO


© © CO


CO


i-H


iO> CM




i—i


CN


c

1


CO


oc


©

0-1


c


c


Ol


Tf


c 1


c


© o

1


(^ © 1-*

©

©


T-H


©


N ©

©

















o


. — 1


lO


i - i
















©


o


CO


o ©

















>o>


o


i-O


N X

















©


CO


©


O Tt<


H
















o


lO


©


© r>

fa P

o

K


t^















-H


-H


+1


-H -H


i-O


•*


r


CO


V


co


r


©


r


■<*


t^


10


cc


© iO


O iO X


©


N


iO ©



CO


L0


CN


©


Tt

•o


c


X


Tf


©


10


CM


CO ©


Ol © Ol


©


©


N 101



IN


co


co


co


cc


CO


LO


CM


©


10


CN


10


© Ol


© CO i-H


co


t^


TjH CO



t>


©



CN


t>


co


CN


CM


CM


T—


CM


c


O) — i


© © co


»H


©


r- i-h










©


C


CO


i0


I—


c


© o


t^ i— t i— i


©



© CM






1




"






1



1


CO


c


CO
















■~ a


u' aj u


cu


u v


u «'



02

a

a;

H















cu ^ cu



CD T3


co 'd















+3 o co c S c

g cu

C3 -M


iamet ternc iamet


CO


P S-i

§ -2


+-> o S Pi

rJ >h

TO V



03


co


a


z


1


i

s


L.



Q.


1


f c


■g


S rS


QhQ



Q^


ShS



3



c



C



C



C


-t-:


c


+:


c








CO


co



"a



a


P


~

c.


a


E a


a


e








C3 CO

£

o 6


c5


I


p

r


t

a


=

r


|


r


f —


r


e=


1


CO

s

Pi

w


F o


^5


S G

O CO •-H CO




5 £


p



^C


P —





1


^P


£


Mode Standar


'■+3

99

">

CO

T3


c9 j5

11


CO



£



7


24


SHINKISHI HATAI


TABLE 2


Length of internodes. Under each group the first column gives the frequencies ivhen the observed values are reduced to 1000. The second column (F) the observed number of frequencies.


INTER











NODAI. LENGTH


GROUP I


GROUP II


GROUP III


GROUP IV


GROUP V


P


F. '



F.



F.



F.



F.


150


25.5


4










250





1.8


2


0.7


1






350


114.6


18


14.6


16


11.2


17




0.5


1


450


242.0


38


86.6


95


54.8


S3


1.7


1


2.6


5


550


382.2


60


260.7


286


192.1


291


9.9


6


3.1


6


650


152.9


24


278.9


306


264.0


400


29.8


18


6.3


12


750


38.2


6


174.1


191


233.7


354


100.8


61


28.7


55


850


19.1


3


93.9


103


134.0


203


145.5


88


56.9


109


950


19.1


3


54.7


60


66.7


101


216.5


131


90.3


173


1050


6.4


1


19.1


21


29.7


45


175.2


106


104.4


200


1150




11.9


13


6.6


10


122.3


74


131.0


251


1250




2.7


3


2.6


4


67.8


41


101.2


194


1350







2.0


3


59.5


36


124.7


239


1450










31.4


19


81.9


157


1550







0.7


1


16.5


10


73.1


140


1650




0.9


1





13.2


8


59 .


113


1750









6.6


4


47.0


90


1850









1.7


1


30.3


58


1950






0.7


1





22.4


43


2050






0.7


1


1.7


1


12.0


23


2150










5.7


11


2250










7.8


15


2350










4.7


9


2450










0.5


1


2550










1.6


3


2650










3.1


G


2750













2850










1.0


2




157



1097



1515



605



1916


LENGTH OF THE TNTERNODES


25


TABLE 3


Diameter of internodes. Under each group the first column gives frequencies ivhen the observed values are reduced to 1000. The second column {F) the observed number of frequencies.


DIAMETER


GROUP I

F.


GROUP II

F.


GROUP III

F.


GROUP IV

F.


GROUP V

F.


3


6.4


i




0.7


1


5.0


3


0.5


1


4


12.7


2


6.4


7


4.6


7


0.0



0.5


1


5


146.5


23


62.0


68


37.6


57


1.7


1


4.7


9


6


286 . 6


45


140.4


154


70.0


106


24.8


15


11.0


21


7


305.7


48


263.5


289


163.0


247


59.5


36


50.1


96


8


146.5


23


289.9


318


206.6


313


39.7


24


7S.3


150


9


44.6


7


138.6


152


194.7


295


124.0


75


82.5


1 58


10


51.0


8


96 . 6


106


298.4


452


343.8


208


264.6


507


11




2.7


3


13.2


20


143.8


87


137.8


264


12


t





9.2


14


148.8


90


226.0


433


13






2.0


3


76.0


46


114. S


220


14








28.1


17


25.6


49


15








.5.0


3


3.7


7




157



1097



1515



605



1916


b. Frequency Distributions. Chart I

An examination of the polygons which represent the observed data shows clearly that they deviate from the normal symmetrical figures to a considerable extent. Even those curves which come nearer to the symmetrical figure have the maximum ordinates too high to be fitted with the equation of the normal curve


V =


OV2tf


2 a


An actual determination of the various constants from the data according to Pearson's method of the moment proves the correctness of our supposition, for in every case the values of p t and skewness deviate widely from zero, while in the normal symmetrical curve, these constants just mentioned should equal zero or be very near to it, and in addition the value of /? 2 is not equal to 3.


26 SHINKISHI HATAI

When we classify our curves according to the corresponding values of the constants, the frequency distributions of the diameters should be represented by Pearson's curve of type 1, except in group 4, in which it will be represented by the curve of type 4.

On the other hand, the frequency distributions of the length of the internode in groups 1, 2 and 3 will be represented by the curve of type 4, while the remaining two groups are represented by the curve of type 6.

I have not calculated the theoretical values of the curves, since the observed curves as they are answer our present purpose. In plotting these curves, we reduced the total frequency on the basis of one thousand in each case. This not only renders a comparison easier, but at the same time we can get a clear notion as to the mode of a gradual transformation from small to larger values where the total number of the internodes is approximately constant throughout life (see p. 28 for the validity of this assumption).

Thus we assume that there are one thousand internodes in the nerve of the thigh, and these numbers are distributed in different stages of the growing period in the manner shown in figs. 1 (diameter) and 2 (internode).

Let us first examine the curves for the internode. In group 1 (see p. 24) the curve is much nearer to the symmetrical figure and the range of the length of the internodes extends only from 150 to 1050 micra. We notice here that the theoretical maximum ordinate corresponds with the internodal length of 532 micra; that is it stands nearly at the middle of the abscissa.

In group 2 we notice several changes when compared with the curve for group 1 .

1. The total range of the variates has been increased and extended more towards the higher values (250 to 1650 micra).

2. The position of the node has moved from 532 in group 1, to 670 micra in group 2, and thus it is now situated at about onethird of the distance of the total range from the lower end.

3. Finally the shape of the curve shows a still greater deviation from the symmetry.


LENGTH OF THE INTERNODES 27

All these changes when compared with group 1 indicate that as a consequence of the growth of the frog, the internodes have grown in length, and therefore the shorter internodes which are so frequent in group 1, have become less frequent in this group.

With some insignificant modifications, the same statements apply to all the other curves. Finally in group 5 we see a very wide extension of variates from 350 to 2650 micra, and at the same time the ranges found in group 1 can only be found in this case in the lower third of the total range. The position of the mode has now moved from 532 to 1310 micra, indicating that as a whole, the modal value of the internode has increased 146 per cent when compared with group 1. From being nearly symmetrical, the curve itself has become highly skew. The increase in range of the variates may also be shown from the successive values of the standard deviation which increase gradually in round numbers from 139 micra in group 1 to 351 micra in group 5.

Let us examine now the curves for the diameters. For this character the distributions of the frequencies are very irregular as compared with the curves which illustrate the distributions of frequencies of the internodes.

Nevertheless in all fundamental features, the two records agree very nicely ; that is the range of variates is smaller in grouD 1 than in group 5. The position of the mode moves gradually from a less diameter to a greater diameter; from 7 to 11 micra. The position of the entire curve moves frrom left to right. The only distinct difference which can be found in the curves of the internode and diameter lies in the fact that the total amount of increment during these five successive stages is considerably less in the latter than in the former, and in fact the internodes show almost three times as much increase as the diameters (see p. 28) and consequently the changes in the form of the curves for diameter from group 1 to group 5 is less conspicuous than in the case of the internode.


28


SHINKISHI HATAI


c. Mean and Standard Deviation

As has already been pointed out by both Boycott and Takahashi, a determination of the proper average is very difficult on account of possible modifications due to the technique. The larger the animal, the longer is the internode, and vice versa. In the former case, the longer internode is easily broken and as a consequence the shorter internodes are more often measured than the longer ones, while in the smaller frogs such accidents are less frequent, and the resultiug mean will therefore be much nearer to the true value in this case than in the case of the larger frogs. We cannot however eliminate this error but it must be kept in mind when we come to draw the final conclusions. Our observations yield the following mean values for each group:


GROUP I

%


GROUP II

% inc.


GROUP III

% inc .


GROUP IV

% inc


GROUP V

% Inc.


t,. Diameter.

Mean

Internode


6.7

532 2


0.0

1.0


7.612.43 670.2 25.04


8.4 706.9


25.15 30.95


10.3

1031.2


52.43 10.6 57.01 33.761311.5146.45


As we should expect, the mean values for both diameter and internode increase from group 1 to group 5. In order to see how much increase both diameter and internode have made during these five stages, the two measurements in group 1 were compared with those of the succeeding stages. As is shown in the same table, the percentage increase for the diameter is far below that for the internode. In fact the total increment in the diameter of group 5 over that of group 1 is only slightly more than half of the value of group 1, while in the case of the internode, it amounts to more than one and a half times the value of group 1. This means of course that the internode has grown three times more rapidly than the diameter during the same period. Whether or not this relation of growth is also true for the growth of the thigh itself, when its width and length are compared, is an interesting point to be determined. At least one fact is true as will be shown later, that the growth rates of the internode and of the entire nerve are approximately equal.


LENGTH OF THE INTERNODES 29

As to the constancy of the number of internodes on a nerve fiber, I may quote first the view of Boycott from his original paper. Boj^cott says (p. 377) :

It appears from the figures that there is a small increase in the number of internodes during the growth in the length of the nerve. The increase is small, yet it is regularly progressive. It may be due to ordinary errors of experiment, while on the other hand it may represent an increase which actually takes place. There is one special circumstance which renders it suspicious. This is the fact that the longer an internode the less likely is it to remain unbroken and capable of being traced throughout its length : the longer internodes will in this way not be measured as often as they should, and hence the average internodal length will be smaller than it should be proportionately to the increase in the length of the internodes, This would account for the increase which is seen in the table, but whether it is the whole explanation it is impossible to say. Assuming that the figures are in the main correct, it must be concluded that there is a small (and somewhat doubtful) increase in the number of internodes, though the main part of the total increase in length is due to an increase in the length of individual internodes.

In the summary he says again (p. 380) :

The number of internodes thus remains approximately constant at all ages. There is a small increase in the total observed number; there are however reasons for thinking that this is due to errors of method.

This question of constancy^ in the number of internodes in the nerve cannot easily be decided until we have a strong reason to believe that the average value obtained for the internode from the given nerve is nearly correct. There is still another theoretical objection to considering the observed values as the average of the whole population, when a small fraction of the lower end of the nerve only is examined (see p. 40). I have also tested this main point by using mean values obtained by the present biometric method, and the following are the results.


30


SHINKISHI HATAI TABLE 4



GROUP 1


GROUP 2


GROUP 3


GROUP 4


GROUP 5


Sciatic length in mm


16.25 532 . 2 30.53


21.07 670 . 2 31.43

2.94


25.60

706 . 9 36.21 18.60


34.10 1031.2 33.07 8.32


4S.17


Mean internodal length in fi. . .

Number of internodes

Percentage increase


1311.5 20.31




When the figures given above are considered, they show a gradual increase in the number of internodes as the nerves grow longer, the significance of which depends of course on the correctness of the mean values. This method of treatment is crude however, owing to the fact that the observed length of the internode for any given diameter differs more or less from the theoretical value. There are some reasons to believe that the theoretical are preferable to raw mean values for this determination. When we employ the theoretical values of the length of the internodes, the results are as follows :

TABLE 5


GROUP 5


Theoretical length of the inter node 531.8

Number of nodes j 30 . 56

Percentage increase



1360.6 35.40 15. S3


From the foregoing we see that whether we use the raw mean length or the corrected mean length, the results are approximately the same; that is the data show an increase in the number of the internodes as the nerve becomes longer. In fact the total gain in the number of the internodes, when group 1 is compared with group 5, amounts to as much as 16 per cent. Even if we assume that the total number of measurements in group 1 was too small (157 observations) to trust its validity, the difference between group 2 (1097 observations) and group 5 amounts to as high as 14 per cent.


LENGTH OF THE INTERNODES 31

We can test this conclusion still another way; since the percentage increase in the internodal length when the mean length in group 1 is compared with that of group 5, is approximately 150 per cent, while the gain in the length of nerve runs as high as 200 per cent, thus the length of the nerve increases faster than that of the internode.

In this connection it is a matter of interest to note that the average number of the internodes is 33.8 in the present case, while Boycott has obtained 34.5. Although the present method of obtaining the mean values was entirely different from that used by Boycott, nevertheless the results show a difference of less than 1.5 per cent, indicating a close agreement between the simpler and more complicated treatment. Therefore we are forced to conclude as Boycott did, that so far as our present data are concerned, the number of the internodes increases with the advancing ages of the frogs.

It is however extremely important before drawing any final conclusions from these data to consider at least the two following points:

1. Technical difficulty in measuring the longer internodes as often as the shorter. This has been already discussed.

2. Number of newly added fibers.

This is certainly an important point to be considered. Takahashi commented on this point thus:

It follows from the foregoing result that so long as the nerve receives new (young) fibers, there will always be internodes which are relatively short, since they belong to fibers which have been subjected to the lengtheniilg process for only a short time. The presence of these fibers reduces the average length of the internodes, and hence accounts in part at least for Boycott's observation that on the average the lengthening of the internodes in the sciatic nerve is slightly less than that of the nerve itself.

Both these points require further study.


32 SHINKISHI HATAI

d. On the Correlation in Growth Between the Internodal Length

and Diameter

It has been stated by the previous investigators that the length of the internode varies with the diameter of the fiber in the sense that the fibers of greater diameter have the longer internodes. According to Birge ('82) during growth the average diameter of the fibers in the frog increases, and the average internodal length also becomes greater. In general, the results just mentioned are in harmony with the findings of Boycott and Takahashi. In this connection however, Boycott has drawn attention to an important point. He stated that

an examination of the table will show however that the internodal length increases -proportionately more than the diameter, so that in a large frog, the length of the internodes in the fiber of given diameter, is greater than in a small frog (p. 372).

Later Takahashi advanced this conclusion of Boycott a step further stating that

in the same frog, the length of the internodes at different levels on fiber of like diameter in the nerves to the leg, increases towards the periphery This increase appears to be associated with the more rapid growth of the distal segments of the leg, but the influence of the segment on the portion of the nerve within it, is less marked as the frogs become larger.

Thus Boycott's and Takahashi's investigations reveal that the relation between diameter and internodes is far more complicated than it was considered to be by the earlier investigators, and varies not only when the different sized frogs are compared, but also at the different levels in the leg of the same frog.

I shall discuss this problem under two heads :

1. The correlation of the length and diameter of the internode from a single locality in a given frog.

2. The relation between the length of the internode and its diameter.

a. From a single locality in frogs of different sizes.

6. In the different segments of the leg of the same frog.


LENGTH OF THE INTERNODES


33


1. Correlation of the Length and Diameter of the Internode from a slngle locality in a glven frog

In order to see whether or not the length of the internode is correlated with the diameter, I have determined the coefficient of correlation between the two characters just mentioned in each group from the formula given below


1


71 / a, a,


(see Davenport's Statistical Methods, 1904) and obtained the following results:


GROUP 1


Coefficient

of correlation or r =: .0797 ±.0535


GROUP 2


0.1947 ±.0195


GROUP 3


0.1480±.0169


GROUP 4


.3399 ±.0242


GROUP 5


.3456 ±.01 36


In every case the coefficient of correlation is positive and is greater than the corresponding probable error. We therefore consider that the diameter and the length of the internode are certainly correlated. Though the degree of correlation is never high, nevertheless we can say positively that so far as the present data are concerned, the longer internode is associated with larger diameter in the values given in each group, and vice versa.

We also noticed in the above table that in general the degree of correlation becomes greater as the animal becomes larger. In another trial we have obtained from Takahasi's data (measurements from sciatic in thigh only) a coefficient of correlation as high as 0.6681, his observations having been made on large frogs (body weight 26 to 63 grams). Since during the period of rapid growth the relation between the two characters must be more irregular, it follows that the degree of correlation would be less in the young than in the adult. Therefore the view maintained by the earlier investigators that "the length of the internodes


34 SHINKISHI HATAI

varies with the diameter of the fiber in the sense that the fibers of greater diameter have the longer internodes" is correct when the fibers are from the same segment of the same animal.

This conclusion is still true even when all the measurements in the five - groups are added together, since in this case again the greater diameter is still associated with the longer internode, though the degree of correlation is lessened.

The average of the coefficients of correlation in the five groups is found to be 0.2216.

2. The Relation Between the Internodal Length and

Diameter

a. From a Single Locality in Frogs of Different Sizes

We have shown already that the diameter varies with the internodal length in the same animal (this is true either for an individual or for averages of several individuals) though the degree of correlation is not high.

When however, several frogs of different sizes are compared with each other, the internodal length varies, as has been found by Boycott, proportionately more than the diameter, so that in a large frog the length of the internodes on a fiber of given diameter is greater than in a small frog. This finding of Boycott can plainly be seen when one examines either the tables or charts of his original paper. I have also prepared by a different method the tables and charts to demonstrate this point.

The following table shows the observed values of the internodal lengths corresponding to the different values of diameters in all of the five groups of Boycott. These values have been determined from the correlation tables (see tables 10-14, pp. 45-47) the internodal length being the means of arrays corresponding to the various values of the diameter.


LENGTH OF THE INTERNODES


35


TABLE 6


DIAMETER


GROUP 1


GROUP 2


GROUP 3


GROUP 4


GROUP 5


,"


P


V

P


P

"


3


450



350


517


550


4


150


579


536



950


5


515


571


561


750


72S


6


521


608


606


S70


1055


7


558


6S3


655


900


1039


8


541


698


724


S92


1123


9


464


705


71S


941


1200


10


625


660


759


1025


1306


11



817


750


1090


1373


12




757


1129


1384


13




617


1113


1420


14





1068


1523


15





1050


1993


INTERNODE


1900


Chart 2. Showing the observed values of the internodal lengths corresponding to the different values of the diameters in all of the five groups of Boycott.


36 SHINKISHI HATAI

Let us first examine the curves (see chart 2). Although the curves are irregular at the ends where the number of observations is not large, yet each can be seen to follow a characteristic course when the middle portion is examined. We notice, as Boycott found, that the length of the internode varies proportionately more than the diameter, since the length of the internode becomes longer for a given diameter as the frogs become larger. In addition these curves are not parallel to one another but diverge more from the curve of group 1 as the frog becomes larger. This divergence in the course of the curves would of course mean that the relative amount of the increment to the internode following to the increase in diameter is not equal, but is greater for the larger frog than for the smaller frog. Since the smaller frog has a shorter internode for a given diameter than in the larger frog, it follows that the longer internode gains in increment proportionately more than the shorter internode. This relation suggests that the rate of increment in the length of the internode may be proportional to the length of the internode itself.

In mathematical terms this corresponds to the expression

— = hy dx

where x and y represent the diameter and the length of the internode respectively, and h is a constant.

In order to test this hypothesis the equation has been solved in the following manner: —

— = hy, — =h dx dx y

The integration of both terms gives at once the required formula.

f^ = /if dx + k orlogy =hx + k or finally y= A e hx (1) where A = e k .

Therefore if our hypothesis is correct, the foregoing exponential equation should adequately express the relation between the


LENGTH OF THE INTERNODES


37


length of the internode and its diameter in frogs of any given size.

Assuming that our hypothesis is correct, the next step was the actual determination of the two constants h and A from the observations for each group. I have determined these constants by ordinary algebraic methods, and the five equations stand as follows :

TABLE 7

For group 1 y = 477 . 6 e- 0159x

For group 2 y = 501 . 1 e- om *

For group 3 y = 470.2 e- 0426 *

For group 4 y = 559 . 9 e 0553x

For group 5 y = 613 . 5 t - 0752x

The following table (8) shows the values of the observed and calculated length of the internodes corresponding to the various values of the diameter. The calculations have been made for each group by the formulas just given.

TABLE 8



GROUP I


GROUP II


GROUP III


GROUP IV


GROUP V


ft


Observed


Calculated


Observed


Calculated


Observed


Calculated


Observed


Calculated


Observed


Calculated


3


450


501




350


534


517


660


550


769


4


150


509


579


565


536


558



699


950


829


5


515


518


571


588


561


582


750


738


728


894


6


521


525


608


612


606


607


870


780


1055


963


7


558


533


683


637


655


639


900


825


1039


1039


8


541


542


698


663


724


661


892


S71


1123


1118


9


464


551


705


690


718


690


941


921


1200


1207


10


625


560


660


718


759


720


1025


973


1306


1301


11




817


747


750


751


1090


1029


1373


1403


12






757


784


1129


1087


1384


1513


13






617


S18


1113


1149


1420


1631


14








1068


1215


1523


1758


15








1050


1283


1993


1896




In the first place if we examine the formula for each group, one point of interest is shown, that is the value of the constant "h" which determines the steepness of the curve (see formula 1) increases regularly from the smaller to the larger frog. Since,


38 SHINKISHI HATAI

when the value of this constant is equal to zero, the resulting curve will be parallel to the abscissa with the distance corresponding to the value of the constant A, therefore the increase in the constant "h" means that the curve becomes steeper as the frog becomes larger, at the same time with an increasing distance of the entire curve from the base line. For the latter statement we find one exception in group 3, in which the value of the constant "A" is smaller than that of the smallest frog. If however we examine the observed mean values in that group, a peculiarity can be found at once; that is the mean values in both the lower and upper ends of the curve are too small compared with the values found in the middle. Since our determination of the constants is based on the observations, such an aberrant result for this particular group is inevitable. Nevertheless, the higher value of the other constant "h" indicates a higher probability that all the mean values for group 3 should have been greater than they actually were.

Therefore with a single exception in the one constant in group 3, the equations are in harmony with the general feature in the growth of the internodes. When we come to an actual test, the observed values are too irregular to make valuable a detailed comparison with the corresponding values obtained by the equation.

On account of the difficulty just mentioned, the question of fit between the theoretical values and the observed can best be judged by an actual comparison of the tw T o graphical representations of the values (see chart 3).

As we notice from the curves, the fit of the theoretical curve to the observed is very satisfactory in the first three groups, and the continuous lines run about the middle of the observed points. In groups 4 and 5 it is not as good as in the former three groups, nevertheless when the irregularity of the observed values in these two groups is considered, the continuous lines should be regarded as the best approximation to the observed values. At least the continuous lines run very close to the observation in the best part of the curve, that is the middle portion where the number of observations is large.


LENGTH OF THE INTERNODES


39


2050 1950


INTERNODE


1850



1750



1650



1550



1450



1350



1250



1150



1050



950



850


/^


750


.


650



550

1350


— •


1250



1150



1050



950



850



750


^^S


650


~ ^"^


550 850


~ •


750



650



550


- ^-T-"" 7


450



350

850


— •


750



650



550 650



550


m


450



MICRA 3


Y=6i3. 5 e 00752X


Y=


559.96


0.0553X


Y=47o. 2 e 00426x

Yr501.,e° 0400X

Yr477.6e 00159X DIAMETER


12 13 14 15 16


Chart 3. Showing the theoretical and observed values for the length of the internodes in all the five groups. The dots show the observed/ mean values, and the continuous lines the theoretical values as calculated by the formulas in table.


40 SHINKISHI HATAI

Obviously no simple curve will represent the system of points shown in group 4 and 5 better than the continuous lines obtained by the exponential equation. There is no doubt that the diminishing value of the ordinates following the increase in diameter, to be seen at the right end of the curves in groups 3 and 4, must be due to the small number of observations, as well as high irregularity in the length of the node, since it is not only opposed to the general results, but the three remaining curves show contrary relations. We are justified in assuming then that the relation between the length of the internode and diameter in different sized frogs is exponential, as it was supposed to be, and the law can now be stated as follows :

The rate of increment in the length of the internode following the increase in diameter is proportional to the length of the internode itself or

dx

In this connection we have tested three more hypotheses touching the rate of increment in the length of the internode.

1. Namely that the rate of increment is proportional to the diameter :

— = hx dx

2. That the rate of increment is proportional to the product of the length of the internode and diameter:

dy 7 — - = nxy dx

3. That the rate of increment is proportional to the quotient of the length of the internode when divided by the diameter:

dy = h y dx x

When however these three equations were tested, it was found that so far as the constants are to be determined from the observations, all these equations give a very poor fit to the observed


LENGTH OF THE INTERNODES


41


data. Thus there is no doubt as to the justification of our first hypothesis that

dx


hy


It follows from the foregoing analysis that Boycott's conclusion that "the internodal length increases proportionately more than the diameter" can now be put in the following way:

The length of the intern ode is proportional to the diameter in any growing stage of the frog in the sense that the former varies with the latter according to the relation given by the equation

y = Ae hx (1)

III. Takahashi's observations on the length of the internode in different segments of the leg of the same frog.

Takahashi found that in the same frog, the length of the internodes at different levels on fibers of like diameter in the nerves to the leg, increased towards the periphery. The following is the table quoted from Takahashi' s paper (Takahashi '08, table 8) :

TABLE 9 Showing the relative length of the internodes at T z compared with those at S it asa standard, in the case of the several diameter classes in all three frogs




LENGTH OF INTERNODES



DIAMETER


IN (JL AT



IN jX






s,


T 3


Frog 3


4.0 5.3


425 530


578



714



6.3


578


805



5.3


623


834



6.3


645


917



7.3


805


1039



5.3


711


837



6.3


828


9fi3



7.3


SS.5


1001



5.9


659


855




42 SHINKISHI HATAI

The above table shows clearly that for any given diameter the length of the internodes which are nearer to the proximal end of the leg are shorter than those away from it. This observation suggests the possible existence of an exponential relation between the internodes from different segments of the leg in the same frog. In order to investigate this point, the total averages were taken as shown in the above table. I took this average for the reason that first, all frogs are mature, and second, on account of higher variability of the length of the internode, the sum of the larger number is important. From the average we see that the diameter which is approximately 6 micra gives 659 micra for the internode at the upper end of the thigh, and 855 micra for the internode found at the upper portion of the foot (cruro-tarsal joint), that is the relation between the two is 1:1.29.

To determine whether or not a similar relation can also be found between the length and diameter of the internode in different sized frogs, I selected two values of the internode for the diameter of 6 micra; one from group 4 and the other from group 5, as these two mean lengths are the nearest values giving a diameter comparable with the averaged figures in Takahashi's table. We have here 780 micra and 963 micra for the internodes in groups 4 and 5 respectively. 2 In this case the proportion between the two internodes is 1 : 1.23 as contrasted with 1 : 1.29 in the other.

These two ratios agree very well for the diameter of 6 micra. This agreement is necessary for the argument, but does not permit us to conclude that the relation between the diameter and the length of the internode in the different segments of the leg of the same frog is also exponential, since we cannot determine the form of the curve. For this reason we must seek for further evidence.

An examination of the figs. 3 to 6 in Takahashi's paper gives us an additional reason for the above conclusion, since there we find that not only the length of the internode for a given diameter increases towards the periphery, but the three lines repre 2 Takahashi has pointed out thai the internodes on the fibers from the American frog, R. pipiens, are shorter than those found by Boycott on the fibers of R. temporaria (fusca).


LENGTH OF THE INTERNODES 43

senting the diameters of 5.3, 6.3 and 7.3 micra respectively are not parallel, but show a slight divergence (see also Takahashi '08, tables 4 to 8). The amount of divergence is very slight, but is regularly greater for the line corresponding to the greater diameter. This* relation agrees with that found to exist between the two characters in the frogs of different sizes. We feel justified therefore in concluding that our hypothesis that the relation between the diameter and the length of the internode is exponential, even when applied to the length of the internodes from the different segments of the leg of the same frog, is correct.

On the basis of the preceding argument, we present the following final conclusion: — The exponential equation


h x


y = Ae

expresses the relation between the length of the internode and its diameter either in different segments of the leg of the same frog, or in frogs of different sizes. This law seems applicable to both Rana pipiens and Rana temporaria (fusca).

Thus far we have merely demonstrated that when {he data are examined, the relation existing between the two characters under consideration is adequately expressed by an exponential formula, but we do not make any inference as to the immediate factors which bring the two characters during the growth period into such exponential relation. As one of the factors, Takahashi counts the segmental influence, by which I understand the elongation of the internodes of Ranvier depend on the elongation of the segments of the limb in which they are found.

If this view of Takahashi is correct, we shall be justified in concluding that the growth of the segments in Rana pipiens follows also an exponential formula. Although unfortunately I have notsufficient data to test this point just mentioned in either Rana pipiens or Rana temporaria, I find it to be true at least in the case of the leg of the toad as shown by the recent investigation of Kellicott on Bufo lentiginosus (1907) in which the rate of increment in the length of the segments is proportional to the length of the segment itself.


44 SHINKISHI HATAI

This relation has been determined from the coefficients of correlation and standard deviation given in his paper. We assumed of course that the regression between the body length and length of thigh, shank and foot is linear. This is admittedly a crude method, nevertheless so far as we wish at the present moment to determine the general features of the curves, such simple treatment answers the purpose.

We have however one important difference between the segments of Rana and those of Bufo, in the fact that in Rana the length of the segment regularly increases towards the periphery, while in Bufo, the length of the shank is least and that of the thigh and of the foot stand in the order named. Thus in the toad if the growth of the internodes is not at all influenced by the elongation of the segments, we shall find a progressive increase in the length of the internodes towards the periphery, and the exponential law in this case will apply to the entire extent of the leg, while on the other hand if the segmental influence is the main determining factor, we shall find the shortest mean length of the internode in the shank, and the exponential law in this case will be applicable to each segment independently.

Therefore the question of general segmental influence as the main determining factor of the length of the internode in the different segments, can best be tested from the internodes in the nerve of the different segments in the leg of the toad. I hope to be able to take up this question in the near future.

BIBLIOGRAPHY

Birge, E. A. Die Zahl der Nervenfasern und der motorischen Ganglienzellem in 1882 Ruckenmark des Frosches. Archiv.f. Anat. u. Physiol., Part 5 and

6, pp. 435-570. Boycott, A. E. On the number of nodes of Ranvier in different stages of the growth

1904 of nerve fibers in the frog. J. of Physiol., vol. 30, pp. 370-3S0. Davenport, C. B. Statistical methods. Second edition, New York.

1904 Kellicott, Wm. E. Correlation and variation in internal and external characters

1907 in the common toad (Bufo lentiginosus americanus, Le C.) J. of

Exper. Zool., vol. 4, no. 4, pp. 575-614. Takahashi, K. Some conditions which determine the length of the internodes

1905 found on the nerve fibers of the Leopard frog, Rana pipiens. J. of Comp. Neurol, and Psychol., vol. IS, no. 2, pp. 167-197.


LENGTH OF THE INTERNODES


45


CORRELATION TABLES.

TABLE 10 Showing correlation of length and diameter of internodes in micra



a


3.









©






o


o


o


o


o


©


o


©


©


©






o


©


©


o


o


o


?


©





t

t


2


1


I


1


i


00

1


i


i


TO




o


o


o


o


O


o


o


8

00


©



03


Diameter


c

1— 1


o


o


o co


§


o


g


o


©

05


©















H


3u





1








1


4


2



3


7


9


3


1





2


5


23


6


1

1



7 4 3

1


12

8 3 6

1


15

23

10

1

2


8 8 3

2


3

2


2 1


1 1 1


1


45


7


48


8


23


9


7


10


8




Totals


4



18


38


60


24


6


3


3


1


157





TABLE 11

Showing correlation of length and diameter of internodes in micra


"3



















=5.








©


©


o


©


©


©


o


o




a


o


©


©


o


o


©


©


©


©


©


©


©


o


©




©


©


©


©


©


©


©



CM


co


■*


US


CO


t^



■g


7


i


u i J


y


[


00

1


J


i


i


1


1


i


1


1


1


m



©


©


©


©


O


o


o


©


©


©


©


©


©


©


©


oj



©


o


©


©


©


©


©


o


©


©


©





©



Diameter



CO


9


>o


to


t~


00


05


©



CN


co


■>*


lO


«o


H


4/*



1


2


2




2










7


5



4


20


18


15


7


3


1









68


6



2


30


59


32


15


9


6








1


154


7



4


18


71


85


53


33


11


9


5







289


8



5


13


68


106


57


31


22


7


6


3






318


9


1



7


39


30


37


18


14


5


1







152


10


1



5


29


37


22


6


5



1







106


11






1



1


1









3


Totals


2


16


95


286


306


191


103


60


21


13


3






1


1097


46


SHINKISHI HATAI


TABLE 12

Shouting correlation of length and diameter of internodes in miera

GROUP III


V


















a.



1




©


©


g



c


= ©


_ —


_ —


c




o


o


©


o


o


©


o e


o


o


o


©


© ©


p <=


© ©





©


©



©


©


©


© ©



cn


cc


T


i- CO


1- X


OJ ©




V


T


7


>-.


©


1

/










CN



c


©


©


o


©


^



© ©


©


©


o


c


© ©


c o


- =


O


7i



©


c


©


c


©



o ©


©


©


o


c


o ©


c ©


© ©


o





CO


-r



©



00 05


o



M


re


-r 1Q


co i

X 3!




Diametei
















GN


r

Zp



1














1


4



3




3


1













7


5



4


14


23


8


5


2 1












57


6


1


6


P.i


29


22


17


11 1













106


7




20


si


77


29


23 12


3


2











247


8



2


16


57


80


73


37


24


16


1


4










315


9



1


8


44


88


87


40


18


6


2





1







2'. 15


10




4


55


111


134


80


41


20


2



3







1


1


452


11




1


1


5


4


8


1











20


12





1


5


3


2


3











14


13




1



1 400


1 ".54


203












3


Totals


1


17


83


29J


101


4"


10


4



1



1


]


1515


TABLE 13

Showing correlation of length and diameter of internodes in micro

GROVP IV


u"


u

Diameter "

©~

©

1

o o


o © ©

©

©


© © i © ©

©


s.

o o I

© ©


g


c

©

c


a

M

©

o


© c

o ©

CM


© © —

o

o


o ©

Li

o ©


© © ©

© ©


© ©

©

o ©


© ©

X

o

©


© © ©

T © ©

S

© © ©

CN

1

©

s

05


© ©

©

C

o

' CM



>■

4

5


7

8

9 10 11 12 13 14 15


1


2

1

2 1


6

4 5

2


1

4 5

7 6 22 10 4 2


3

7

5

17

33

12

6

4

1


4

7

2

24

50

10

17

11

5

1


3

6

4

13

33

15

19

6

6

1


4

3

7

21

14

14

8

2

1


1 1 1

14

10

10

2


17 3

11 3


6

5 3

1 1


3 4 1


2 3

1


2

1 1


1



1


3

1 15 36 24 75

Jl.s

^7 90 40 17 3


Totals 1 ti


is


.,1


^


131


106


71


11 36


19 10


8 1


1



1


01.5


LENGTH OF THE INTERNODES


47


TABLE 14

Showing correlation of length and diameter of internodes in micra

GROUP v


eter

node -400//

500


o o o


3


o

=

00


o o oa


o o o


o o


c o


o

o

C<3


o o


©

©


o o

CO


o © © © o o © o

N » O O


© o

o o

— CO


o ■© © o

o o o ©

re ** io CO N ?i c;i c^i


3 ©


3

3


O

o

OS


a a


o o o o

M -f


o 3


o o CO 1


o 3


o o oo


© o

05


s

o


o

o


O

c


©

©


© o


©

©


o o

© ©

CD r

© o

f


o

a


c

CI


o


©

=

IM


o

3 CO


© s

Ol


o

~

IN


3 3

(N


3 O

(N


© O / <N


"5 o


3n




1

























1


4








1





















1


5



1


1


3


1


1


1


1




















9


6



1




1


5


3


4


2


1


1


2





1












21


7



1


3


11


16


20


12


11


7


4


3


6


2














96


8


1


3



3


9


16


20


25


24


14


12


7


7


2


3


1


1



1



1







150


9




1


1


10


12


17


25


25


13


lit


11


4


8


5


4


2




1








15S


10





2


15


32


46


40


72


56


69


46


41


37


is


11


9


3


2


3


1



1


3




507


11




2



5


12


15


21


26


34


42


19


24


L914


10


9


6


4


2








264


12






2


6


28


58


70


44


44


40


33


28301614


6



•;


4


1


1


1



1


433


13






1


7


16


12


20


23


45


23


18


1213


8


5


8


3


3


2




1




220


14







2


6


2


1


2


3


6


7


4


4


6


3



1





1


1




49


15















1


3


1






1






1


7


Totals 1


5


6


12 55


109


173


200


251


194


239


157


140


11390


58 13


23


11


15


9


1


3


6



2


1916


350


POSITION OF MODE

T TT T T


350


POSITION OF MODE

TT T ( TT

It IV V

M i I


DIAMETERS



Fig. 2


15 MICRA


t n i r

oO 250 450 650 850 1050 1250 1450 1650 1850 2050 2250 2450 2650 2850 MICRA

< hart 1. Curves showing the distribution of frequencies of the several lengths of the internodes and diameters for all of the five groups. Fig. 1, Internodes. Fig. 2. Diameters.


THE OLFACTORY NERVE, THE NERVUS TERMINALIS AND THE PRE-OPTIC SYMPATHETIC SYSTEM IN AMI A CALVA, L.

CHARLES BROOKOVER

From the Anatomical Laboratory of the University of Chicago WITH THIRTY-FIVE FIGURES

Material and Methods 50

Historical Sketch 52

Embryological Development of the Olfactory Organ in Amia 57

Adult Structures of the Olfactory Organ in Amia 77

Central Relations of the Nervus Terminalis in Amia 83

Intra-cranial Sympathetic System Posterior to the Nervus Terminalis in

Amia 92

The Nervus Terminalis of Lepidosteus and Teleosts 106

Discussion and Theoretical Conclusions 108

Summary of Results 113

There are two main reasons why Amia was selected for a study of its nervous system. It is an archaic form likely to show interesting relationships with the sharks and other lower vertebrates. In the second place, it is relatively abundant in the region where the work was first undertaken, so that it could be procured in numbers to satisfy the demands of the neurological methods. It was our original intention to describe the anterior part of the central nervous system and thus supplement the work done by Kingsbury ('97), on the medulla oblongata and by Allis ('97) on the peripheral nerves of Amia, but when it was found that there are large numbers of ganglionic nerve cells among the olfactory fibers in the adult nasal capsules and that these cells are derived from the ganglion which Allis described in the young Amia, it was thought best to study first the peripheral olfactory apparatus. The work on the central nervous system has been deferred for a later paper. The present article includes a study of the em


50 CHARLES BROOKOVER

bryology and the adult morphology of the olfactory apparatus in Amia with some comparative work on Lepidosteus and a few teleosts. It has seemed best to include a description of certain sympathetic nerves and ganglia found in the meninges of the forebrain of Amia.

MATERIAL AND METHODS

The demands of the various kinds of neurological technique have required the use of a large number of young and adult Amia. At one time time 450 young Amia from 60 mm. to 75 mm. in length were employed for Golgi preparations. In addition to the neurological preparations of young and adult Amia, a great many embryos and young from two days before hatching to 100 mm. in length were sectioned, mounted serially, and stained for cytological results. Most of these sections were stained with Heidenhain's iron hematoxylin, but others were stained with Delafield's hematoxylin and by Weigert's method. Much of the first material which was available to me for cytological preparations of embryos had been fixed in bichromate-acetic fluid or in formalin. Later, when embryos and young were collected they were fixed in Zenker's fluid.

Some of the neurological methods which gave the best results in my hands will be indicated briefly. The rapid process of Golgi's method (Lee's Vade Mecum, sixth edition, p. 437) gave the best results. Double impregnation was often employed to good advantage. Good results were sometimes procured on adult brains by a preliminary fixation for from five to twelve hours in four per cent formaldehyde neutralized with lithium carbonate or with ammonia. The treatment with neutral formalin seemed to favor the impregnation of axones.

In order to cut serial sections of the Golgi preparations the pieces were imbedded in paraffin melting at about 50° C. Xylol, which dissolves the silver precipitate, was avoided and the clearing done in cedar oil. The cedar oil was used repeatedly and is probably all the better for being saturated with silver after use the first time. The cedar oil was warmed over the imbedding oven


NERVUS TERMINALIS IN AMIA 51

after each use for clearing, in order to evaporate the alcohol introduced. Dehydration was accomplished by passing the pieces quickly through the lower grades of alcohol, which have a tendency to dissolve the silver precipitate, and was made complete by two changes of 95 per cent, and of alsolute alcohol at intervals of about two hours. Pieces were finally left in cedar oil for twelve to twenty-four hours. From the cedar oil the pieces were transferred to melted paraffin which was changed three times at intervals of three or four hours in order to have the blocks free of cedar oil. By lowering an electric light bulb over the block while cutting in cool weather, sections more than 50 micra thick can be cut in ribbons with ease.

Series of adult brains by the Weigert method were cut in three different planes and a transverse series of the head of a young Amia 100 mm. in total length was made. The iron hematoxylin method recommended by Houser ('01) gave better results than the Weigert method in the forebrain where there is little medullation. Bielschowsky's formalin-ammonia-silver method was tried with some success, but by far the best results were given by the Cajal ('05) treatment for the anterior part of the brain and for certain fibers in the olfactory capsules. For showing axones by the Cajal method, the best preparations were made after fixation in 95 per cent alcohol. In some cases a small amount of ammonia was added until there was a slight reaction to litmus paper. Material thus fixed may be kept in 80 per cent alcohol until it is convenient to employ the silver bath. Small pieces were kept warm in silver nitrate of about 2 per cent, for from three to five days. The same precautions were taken as for Golgi preparations to avoid xylol and the prolonged use of low grades of alcohol in embedding. Paton's ('07) modification of Bielschowsky's method was tried on embryos and young of Amia. It showed neurofibrils in none but the larger elements of the brain in the later stages after hatching.

Nissl preparations were made of the brains of adults to show regional differentiation and to demonstrate the characters of the cells of the nervus terminalis. Material fixed in Graf's chromoxalic mixture cited by Houser ('01) gave better results than


52 CHARLES BROOKOVER

either formalin or alcohol fixation. Toluidin blue gives a more brilliant stain than methylen blue and is easier to control, as it does not diftentiate so rapidly in the alcohol. Neither does it seem to fade so rapidly afrer the sections are mounted in balsam. Methylen blue was used intra- vitam by injecting it through the ventral aorta. One half of one per cent solution in distilled water was kept in stock and mixed with ten times its bulk of normal saline solution at the time of using. Chemically pure sodium chlorid was used with Griibler's Bx brand of methylen blue. The method employed was, in the main, the one used by Wilson ('04). Pieces were cut out and examined under the compound microscope to determine when impregnation had taken place. The pieces were kept in the ice box while in the ammonium molybdate and in the alcohols. This gave better results on the peripheral than on the central nervous system.

I wish to thank the Ohio State Academy of Science for a substantial contribution from the McMillin fund toward defraying the expense for material. I am much indebted to Mr. Alex. Nielsen of Venice, O., for furnishing me adult fishes fresh from his nets and to Prof. Herbert Osborn for facilities for working and gathering material while at the Ohio State University Lake Laboratory on Lake Erie during three summers. I am also under many obligations to Prof. C. Judson Herrick for generous contributions of literature loaned or donated, as well as for his friendly advice and criticism while engaged in the work.

HISTORICAL SKETCH

In 1862 Max Schultze showed that the nucleus of origin of the olfactory neurones of the first order lies peripherally in the olfactory epithelium. Since that time his results have been confirmed by various workers repeatedly. Bedford ('04) gives a brief summary of some of the more important papers, — especially those of an embryological nature. His resume of the literature shows that the older embryologists were of the opinion that the olfactory nerve arises centrally. Then views changed after the appearance of the paper of His, Jr., in 1889, showing that in


NERVUS TERMINALIS IN AMIA 53

the human embryo the olfactory nerve arises exclusively from the periphery. The view prevails at present that the olfactory nerve differs from all other nerves in vertebrates in that its cells arise and remain in the ectoderm. Bedford's investigations on embryo swine support this view.

It would seem from the brief account just given that the embryology of the olfactory nerve is simple, but Mimes Marshall, Balfour, von Kolliker, Beard, Chiarugi, Disse, and others have described a ganglion in the developing olfactory nerve. Some thought this ganglion came from the brain or from the forward continuation of the neural crest into the region of the olfactory nerve. His ('89) described the ganglion as coming from the periphery and contributing neuroblasts to the formation of the olfactory nerve. Disse ('96) says that the cells of the ganglion give rise to the sheath cells of the olfactory nerve. Bedford did not determine their fate in swine.

In 1895 Pinkus described a new nerve in Protopterus, which is intimately associated with the olfactory nerve. Soon afterward Allis ('97) found a nerve in Amia calva near the olfactory nerve. He noted its ganglion in the young and homologized the nerve with Pinkus' nerve in Protopterus. About two years later Locy ('99) found a nerve in sharks related peripherally with the olfactory mucosa. As it had a more dorsal connection with the brain than that described by Pinkus for Protopterus, he was not inclined to homologize the nerve in selachians with the one in Protopterus. He extended his observations to other sharks and found in some species a more ventral connection of the nerve with the brain. He then thought the nerve homologous in Protopterus, Amia and selachians ('03). When the nerve was found to be present in a large number of sharks Locy ('05) named it the nervus terminalis. He also treated this nerve as a part of the olfactory nerve when he first described its development in Squalus acanthias ('99), but later ('05) he came to consider it as an independent nerve.

Bing and Burckhardt ('04) described the nervus terminalis in adult Ceratodus where Sewertzoff ('02) had previously seen it in embryos. In all probability it has not been sufficiently sought


54 CHARLES BROOKOVER

for in the remaining living lung-fish, Lepidosiren. Burckhardt, as cited by Pinkus ('05), found the nerve in Callorhynchus. Burckhardt has suggested that its ganglion might be homologous with the ganglion found by Rubaschkin ('02) in the chick, but Rubaschkin considered the ganglion in the chick to be developed as a part of the trigeminus nerve, in all probability.

Ernst DeVries ('05) described a ganglion in the course of the nerve fibers to the organan vomeronasals (Jacobson's organ) of human embryos of about three and a half to four months. In the guinea pig he found a similar embryonic ganglion. He expressed the opinion that the same relations probably exist throughout the whole series of vertebrates. Although the organon vomeronasale does not exist as such among a large number of anamniotes, he thinks the nerve described by Locy is probably homologous with the nerve to Jacobson's organ.

If the nervus terminalis of lower vertebrates is homologous with the nerve to Jacobson's organ or with some part of the nerve to this organ, we may well look for some evidence of a similar nerve or ganglion everywhere in the vertebrate series, for Jacobson's organ has been described as occuring embryologically or in the adult in forms from Amphibia to man. The literature shows that the nervus terminalis is almost universal among the living generalized fishes. We may have overlooked its presence in the remainder of the fishes. That such may be the case is indicated by the fact that the writer ('08) has found the ganglion among the fibers of the olfactory nerve of the young of two species of Lepidosteus in the identical relations described by Allis for Amia. No mention of such a ganglion has been found in the literature. Also, while this work was in progress the writer found ganglion cells indicating a nervus terminalis in the olfactory nerve of the carp, and Sheldon ('09) found its central connection with the brain. Sheldon and Brookover ('09) reported its presence in the carp at the Baltimore meeting of the American Association of Anatomists. Meanwhile, Herrick ('09) found the nervus terminalis in the tadpole and in the adult frog.

Pinkus ('05) in his third paper dealing with the nerve as found in fishes, summarizes its relations with the olfactory nerve by say


NERVUS TERMINALIS IN AMIA 55

ing there are three points of correspondence : (a) same structure,

(b) same peripheral ending in the nasal mucous membrane, and

(c) same termination in the prosencephalon. He thinks we cannot say anything definite as to its function at present. Johnston ('06, p. 106) suggests that it is probably a general cutaneous nerve. Allis ('97) noted that the ganglion in Amia develops at the same time as the ciliary ganglion and suggested that, since its cells resemble those of the sympathetic rather than cerebrospinal ganglia, the nerve is probably a sympathetic nerve.

Allis ('97) described each olfactory nerve of Amia as being made up of three bundles, of which the smaller, ventro-median, constitutes the nerve "n" of Pinkus. He states that there is an interchange of fibers between all three of the bundles in the young as well as in the adult. He says,

In the adult the ventro-median bundle of the olfactorius is distributed, so far as macroscopic observations can show, to the nasal epithelium at the extreme anterior end of the nose. In embryos such is also its apparent distribution, though I have never been able in my preparations to trace it definitely to that tissue. During the larger part of its course, in 30 mm. to 50 mm. specimens, it is easily distinguished from the lateral bundles by the presence of the large round cells which Pinkus describes in Protopterus. Near the anterior end of the olfactorius, however, these cells disappear, and the fibers into which the bundle breaks up cannot be distinguished from the other terminal branches of the main nerve.

He continues by saying that centrally,

The fibers of this bundle enter, in large part, with the fibers of the lateral bundles into the anterior end of the lobus.

A few of its fibers were traced by him in one series of sections caudally ventral of the brain membranes to enter the brain lateral to the recessus interolfactorius, which latter term he explains by saying,

in Amia, in front of the lamina terminalis, or sulcus olfactorius, and hence into the lobus olfactorius and not into the forebrain proper

(p. 512).


56 CHARLES BROOKOVER

This latter statement of Allis is based on findings in sections of a single young specimen of Amia. He found this posterior bundle in gross dissections of the adult, but says he never could find howit ended in the brain. His illustrations (plate xxxvii) of the adult brain shows the posterior root of the nerve running well back between the optic chiasm and the brain. As I understand the terms "lobus" and "lobus olfactorius" employed by Allis in the above quotations from him, he uses them to indicate what neurologists now call the olfactory bulbs in w T hich are located the mitral cells and their glomeruli. Olfactory lobes are now commonly used to indicate olfactory centers in the forebrain proper. Consequently, it would appear that Allis traced some of the fibers of the nervus terminalis in young Amia into the anterior end of the olfactory bulbs and others into the posterior ventral portions of the olfactory bulbs. He found the main bundle in the adult in the same position as in the young, but traced some of its fibers farther caudad without determining their ending. Locy ('03) examined the nerve in adult Amia and says it does not have the conspicuous separateness which characterizes it in selachians.

After examining a large number of adult Amia brains macroscopically I could never be absolutely certain that I saw anything different from olfactory fibers, which often break up into various small bundles before joining the olfactory bulbs. Anything that seemed to resemble the alleged posterior connection near the optic nerve, could never be distinguished with certainty, when cleared in xylol, from connective tissue or blood vessels. Kappers ('07) had no better results from the study of Weigert and Bielschowsky preparations of the brains of adult Amia which, however, had been removed from the cranial cavities before they came into his hands.

By examining the young stages of Amia in which Allis described the large cells that distinguished the nervus terminalis from the remaining olfactory bundles, it was easy forme to confirm his findings in the main. This led me to undertake a detailed examination of the early embryology of the nervus terminalis. As it had been described as a separate nerve by Locy ('05) in selachians, it occurred to me that it might be connected in Amia and Lepidos


NERVUS TERMINALIS IN AMIA 5/

teus with the sucking disk or adhesive gland on the snout of the young which has been described by Dean ('97) and others; but on examination it was found that the sucking disk had already atrophied to a large extent before the ganglion can be recognized in the young.

In the description that follows, I shall employ the term nervus terminalis which has been largely used in the literature recently, although it will appear that I do not have any evidence that the nerve is separate from the olfactory nerve in the fishes which I have examined, and consequently I am led to the view that it is a part or component of the olfactory nerve, as Locy ('99) in his first work on its development in the sharks considered it.

EMBRYOLOGICAL DEVELOPMENT OF THE OLFACTORY ORGAN

IX AMIA

The sections on which my study of the early embryology of the nasal pit is based were cut transversely or horizontally six micra thick and stained with iron hematoxylin and acidfuchsin. The series of embryos first described were all taken from a single nest from May 17 to May 20, 1908, and fixed in Zenker's fluid at intervals of three hours. Only such stages will be described as show differences worthy of note as compared with earlier stages.

At a period forty hours before hatching and consequently about eighty hours after fertilization, since it takes Amia eggs about five days to hatch, the nasal placodes are quite readily recognizable as solid masses of cells occupying much of the space between the optic cups posteriorly and the sucking disk anteriorly and ventrally. They are somewhat pearshaped with the smaller end pointing forward and downward between the two developing halves of the sucking disk with which they come into contact. At this stage the olfactory cups meet each other at their anterior ends in a common mass of cells (fig. 1) situated ventrally between the two developing halves of the sucking disk. The plane of the section figured slants upward and backward. There is between this unpaired portion of the olfactory placode, which is not far


58


CHARLES BROOKOVER


from what will be the anterior portion of the roof of the mouth, and the neural tube, a small button of cells (fig. 1). No limiting membrane was made out between this group of cells and the brain wall. The cells were like those of what may be termed the unpaired olfactory placode into which they grade.

In a stage six hours older which has not been figured, the button appears as a cord of cells tapering downward and forward



aired olfdc


ru placode


Fig. 1. Transverse section (slanting slightly upward and backward) of the anterior end of an embryo of Amia forty hours before hatching. Shows the paired and unpaired nasal placodes, the anterior end of the neural tube, and the sucking disk. Yolk granules are still evident in sucking disk. Iron hsematoxylin and acid fuchsin, 6 micra thick. X 215.


from beneath the anterior end of the neural tube to the unpaired portion of the olfactory placode. A similar unpaired portion of the olfactory placode has been described by von Kupffer ('93) in the sturgeon. From this fact and the unpaired condition in Amphioxus and the Cyclostomes he argues for primitive monorhinism in the vertebrates and secondary amphirhinism. The


NERVUS TERMINALIS IN AMIA 59

cord of cells just described in Amia would seem to correspond in position to the place where von Kupffer's lobus olfactorius impar ('93, fig. 15) comes into contact with his mediane Riechplatte. It was at first suggested to my mind that in Amia the cord of cells might be the ectodermal portion of the hypophysis, but a mass of cells believed to be Rathke's pouch is already present near the optic chiasm. Johnston ('05, p. 202) thinks the lobus olfactorius impar marks the dorsal border of the neuropore. As the first evidences of the nervus terminalis appear much later, no stages of Amia were cut at a date early enough to show the neuropore. Just as this goes to press Johnston ('09) publishes an article describing nerual crest near the olfactory placodes in lower vertebrates. The early development of the unpaired olfactory placode and of the neuropore should be investigated fully in Amia.

At about the middle of their antero-posterior extent the olfactory pits come into contact with the brain wall forty hours before hatching, when there is still an unpaired placode. This point of contact is some eight sections posterior to the anterior end of the neural tube, that is to say, about fifty micra. The connection is very slender, as it is comprised within two sections. It cannot be definitely stated that any fibers extend across from the placode to the brain at this time, but there is solution of the limiting membrane of the neural tube at this point. There is a delicate limiting membrane, shown by the fuchsin stain producing a red line, between the mesodermal elements and the caudodorsal part of the olfactory placodes, but none could be made out anteriorly and ventrally. The lack of a limiting membrane about the anterior ventral part of the olfactory placodes in the early stages was thought to be due to the recent connection of these placodes with the unpaired placode (fig. 1), and that no membrane has as yet formed in this position. Perhaps this may be the reason why olfactory fibers were observed to appear first at the anterior end of the placode, in most cases, when the olfactory nerve develops.

The condition of the olfactory placodes at a time six hours later than the stage described above, does not differ greatly and has


60 CHARLES BROOKOVER

not been figured. At this time, which is thirty-four hours before hatching each of the paired placodes is still connected with the ectoderm at a point between the sucking disks, although by a more slender cord of cells than was the case in the earlier stage described. But the points where these two cords of cells reach the ectoderm adjacent to and between the two halves of the sucking disk, are now removed farther laterally from each other. From a position midway between the two halves of these two paired placodal connections with the roof of the mouth, there is still a strand of cells extending from the ectoderm to the anterior ventral edge of the tip of the neural tube. This cord of cells is largest near the neural tube. There is no appreciable increase in the size of the connection of the paired placodes with the brain over the previous stage described. A slight indentation of the cuticular ectoderm shows where the external opening of the nasal sac will appear later. The embryos showed slight motions at this age.

About twelve hours later than the last stage, or some twentytwo hours before hatching, there is no unpaired nasal placode nor any evidence of a cord of cells from the roof of the mouth to the brain wall at its anterior end. The paired placodes in one or two preparations of this stage have shown a slight elongation into a point forward just beneath the ectoderm. The nasal placodes, as well as the anterior end of the neural tube, are farther removed from the anterior end of the snout than was the case earlier. At this time there appears to be a fibrous connection between the olfactory capsules and the brain (fig. 2). Whether this connection is protoplasmic or is composed of true neurofibrils such as those described by Paton ('07), was not determined, as the method of Bielschowsky which was tried, did not differentiate any fibers in young or embryonic Amia. The connection is quite slender, as it is seen in only two sections, and is in consequence not more than twelve micra in thickness. A very few cells of the placode near the place of origin of the fibers are slightly larger than the others at this age. The distance from the placode to the outer brain wall is almost nothing at this age and there are no nuclei in the course of the olfactory nerve.


NERVUS TERMINALIS IN AMIA


61


There is a definite membrane about the nasal capsule except where the olfactory nerve originates and for a slight distance anterior and ventral therefrom.

At a period three hours later, or about nineteen hours before hatching, a slight increase has taken place in the number of large cells in the placode near the origin of the olfactory fibers. The latter are not numerous, being confined to three or four sections



Fig. 2 Transverse section of a embryo Amia 22 hours before hatching. Shows contact of the paired olfactory placode with the neural tube. Preparation as above. X 450.


six micra thick. Some cells which have the characters of neuroblasts are now seen within the brain wall near the point where the brain has raised slightly to meet the olfactory nerve. This is the beginning of the olfactory bulbs. There are mitoses in the layer of germinative cells next the brain ventricle. Also, mitotic figures are seen here and there throughout the nasal placodes which have no lumen as yet. A slightly greater space exists at


62 CHARLES BROOKOVER

this stage, than formerly, between the placode and the outer wall of the brain. Mesenchyme accompanied by capillaries and blood corpuscles has pushed into this space. An outer zone of nerve fibers has appeared just inside the brain wall and extends from a point anterior of the olfactory nerve to some distance posterior of it.

Three hours later there are many more fibers in the olfactory nerve but no nuclei in its course. In one preparation of this stage a very small cavity can be made out near the center of the placode. This cavity is not connected with the outside, but lies a little nearer the lateral side of the placode opposite the slight depression already mentioned as occurring in the epidermal layer of the ectoderm. The cells of the mesial side of the placode have assumed their columnar epithelial shape with nuclei at some distance from the lumen, as in the adult ; but those on the opposite side of the lumen, where the external opening will appear later, are more rounded and irregular in arrangement. From this stage onward it is readily seen that the nasal capsule is continuous at its border just under the epidermis with the inner layer of the ectoderm in which taste-buds arise a little later. It can be said then, that at this age, some sixteen hours before hatching, the cells on the median side of the placode have assumed a more or less definite character of neuro-epithelium and it is probable that the olfactory nerve is composed of definite neurofibrils. The fuchsin stain shows a plainly fibrillar structure.

Other stages are about the same until six or seven hours before hatching, when in horizontal sections two bundles of fibers can be seen converging to a union just before entering the brain wall (fig. 3). The anterior one is slightly larger and is probably the older, as the earliest fibers are generally seen near the anterior border of the placode. Some three or four cells are seen among the olfactory fibers just as they leave the nasal capsule. Mesenchyme cells having capillaries and blood corpuscles in their midst, crowd near the nerve but in many of the preparations can be distinguished by their staining reactions from the cells inside the nerve. The two branches of the nerve can be followed well in among the nuclei of the nasal capsule. There is no external


NERVUS TERMINALIS IN AMIA


63


opening to the capsule (fig. 3), although the central cavity is larger than when it appeared ten hours earlier.

At a period about six hours after hatching the olfactory nerve has lengthened to one fourth the diameter of the nasal capsule (fig. 4). A small increase is to be noted in the number of nuclei among the fibers as they arise from the capsule. The two main peripheral divisions of the olfactory nerve are more evident than previously. An outer fibrous membrane covers the brain



Fig. 3. Horizontal section of Amia 6 hours before hatching, showing anterior end of the neural tube, the olfactory placode with beginning of nasal cavity, and the origin of the olfactory nerve from the placode. Iron hematoxylin and eosin. X 450.


except where the olfactory fibers pierce it. A marginal zone of fibers lies just inside the brain at this point, almost entirely devoid of nuclei so that it does not seem possible that any nuclei are migrating from the neural tube into the nerve.

The olfactory capsules remain closed in this series of embryos until about thirty hours after hatching, although slight cavities are to be detected twenty-four hours earlier among the laterally


64


CHARLES BROOKOVER


placed cells of the capsule. At this time the olfactory nerve has a large number of fibers and its length is equal to about half the diameter of the olfactory cup (fig. 5). The fibers spread peripherally, ramifying as small bundles among the cells of the capsule. The thickness of the capsular epithelium is not so great in the middle of the mesal side of the capsule. This is opposite


-k ?






X»th ar inr M,M ol


Fig. 4. Horizontal section of Amia 6 hours after hatching. Shows two rami of the olfactory nerve as it arises from the placode, the migrating placodal nuclei and the fiber zone devoid of neuroblasts just inside the neural tube. Stain as above. X 400.


the middle of the spread of the fibers into the nasal capsule, and causes the capsule to appear flattened laterally. The thin nasal epithelium at this point has more nearly its adult character than the epithelium at the margins of the capsule reached by the laterally spreading olfactory fibers (fig. 5). Laterally where the walls of the nasal capsule are thicker there are nuclei arranged


NERVUS TERMINALIS IN AMIA


65


irregularly at the outer margin of the capsule. These irregularly arranged nuclei form a continuous series with the nuclei within the olfactory nerve, which have migrated out of the placode. At this time these nuclei appear to be nuclei of indifferent cells that are migrating along the olfactory nerve. Probably they are of the same general nature as the nuclei producing the sustentacular cells of the nasal epithelium. Later the larger part of these nuclei along the olfactory nerve produce sheath cells for



Fig. 5. Horizontal section of Amia about 30 hours after hatching. Shows growth in length of the olfactory nerve. Nuclei are migrating from the placode along the olfactory nerve. Note the position of the blood vessel at the anterior side of the olfactory nerve. Stained in thionin and acid fuchsin. X 400.


the olfactory fibers. They are more numerous peripherally than in previous stages, but disappear entirely just before the brain wall is reached. Inside the brain there is a fiber zone, as was the case earlier, that is devoid of nuclei (fig. 5) . The mesodermal elements are closely crowded about the outside of the connective tissue sheath which is seen at the outer border of the nerve peripherally where it is continuous with the sheath around the nasal capsule. Where this sheath is present the mesodermal


66 CHARLES BROOKOVER

elements are outside it and the nuclei within it among the nerve fibers belong to the sheath cells. This membrane, however, cannot always be made out. When this is the case, it is difficult to differentiate the mesenchyme cells from the cells among the fibers of the nerve. Their nuclei are about the same in size, but in favorable preparations stained with iron hematoxylin the nuclei of the mesodermal cells appear darker with a greater number of chrolmatin granules in them. The sheath cells within the nerve have a rather clear nucleus with one or two nucleoli.

Unfortunately an accident seems to have befallen the male fish guarding the nest from which the stages above described were taken and a turtle was found in possession of the nest about thirty hours after hatching. This necessitated taking eggs from a different nest for the later stages. This second series began during the second day after hatching and was continued for more than a week. They were preserved in Zenker's fluid at intervals of about four hours for the first two days. They were stained as in the previous series.

In the first stage of this series it is evident that there has been a large increase in the number of cells in the course of the olfactory nerve (fig. 6) , but the distribution is the same as in the last stage of the previous series. The plane of section does not show the whole course of the olfactory nerve. The cells are more numerous peripherally but are entirely absent near the brain wall. The brain wall has raised up into a cone pointing into the olfactory nerve. This cone is an early stage of the olfactory bulb and its apex is devoid of cells, although there are numerous neuroblasts at its base where the mitral cells are forming. The mitral cells can be distinguished by their larger size and vesicular nuclei, aside from the fact that a wide fiber zone free of nuclei intervenes between them and the nuclei in the olfactory nerve. ( onsequently, there does not appear to be any evidence at this stage that cells are migrating from the brain into the olfactory nerve.

In sections of one or two of the fishes of the earliest stage of this series, there can be seen a slight aggregation of cells among the peripheral olfactory fibers at their anterior side (fig. 6, a).


NERVUS TERMINALIS IX AMIA


67


This is the first indication of the ganglion of the nervus terminalis which, however, cannot be made out with certainty in some cases with specimens nine to twelve hours older. This indicates a rather wide range of variability with regard to its time of appearance. The cells of the aggregation at this early date are not greatly different from those elsewhere among the olfactory fibers. In other words, all of the cells along the olfactory nerve are in


5*


' ] fa -M&^^z


Fig. 6. Horizontal section of Amia during the second day after hatching. Shows the anterior end of the neural tube and part of the olfactory placode and nerve. Note the nuclei at a in the olfactory nerve near its origin from the placode. This is the point of origin of the cells of the nervus terminalis, not to be distinguished at this time from sheath cell nuclei. Iron hematoxylin and acid fuchsin. X 340.


different in character and still embryonic. Occasional mitoses have been found along the nerve in this and earlier stages. When the ganglion cells can be recognized they are slightly larger than the indifferent cells and their nuclei are more vesicular while their cytoplasm stains more deeply. Wherever the indifferent cells are taking on the characters of sheath cells, their nuclei have become elongated and stain more deeply than their


68


CHARLES BROOKOVER


cytoplasm. On account of its deeper staining in hematoxylin preparations, the ganglion of the nervus terminalis is easily made out at a glance in stages two or three days older than this, i. e., in this series about five days after hatching (fig. 7, a). I was not able to recognize the ganglion until about ten days after hatching in another series preserved at longer intervals in



Fig. 7. Horizontal section of Amia about five days after hatching when the ganglion of the nervus terminalis is first recognizable. Note their position at a near the artery b. The olfactory nerve has two main rami to the nasal capsule c Olfactory bulb at d. Iron hematoxylin and eosin. X 340.


the previous year. This may be due to more rapid development during the warm weather of this present season. But the recognition of the ganglion at an earlier date may be due in part to better fixation and staining, as well as to the fact that the sections were cut horizontally so as to show the whole length of the


NERVUS TERMINALIS IN AMIA 69

olfactory nerve and a greater surface of the ganglion, whereas the sections of the previously collected material were cut transversely. At all ages the ganglion appears more striking in sagittal and horizontal sections than in transverse, probably because the ganglion is slightly elongated in the direction of the axis of the body of the fish.

I have searched carefully for any placode or other source of origin of the nervus terminalis outside the olfactory capsule and nerve, but to no purpose, although Locy ('05) describes it as arising separately in Acanthias. If there is a separate placode in Amia, I have not been able to recognize it unless, indeed, the unpaired nasal placode already described is the beginning of the nervus terminalis, for this placode is the most anterior part of the nervous system I have found in Amia. In that event it would have to be said that the placode of the nervus terminalis is absorbed into the paired olfactory placodes from which the ganglia arise some days later.

Some extirpation experiments were made on very young Amia to determine, if possible, the mode of origin of the nervus te minalis. The nasal capsule was cut away from one side of the young, one or two days after hatching. The object was to remove the olfactory capsule completely without too much injury to adjacent parts. Probably complete removal was accomplished in only two cases out of over twenty operations, for there were some olfactory fibers found in all the nerves on which operations were performed except two. The fishes on which operations were made, were fixed and sectioned about a week or ten days after the extirpations. It is impossible to state to what extent regeneration of the olfactory nerve and placode took place in the short time the operated fishes were permitted to live, but in cases where one felt confident that all the placode was taken away as a more or less adherent mass of cells, there were found olfactory fibers later. It is probable that in all those cases where olfactory fibers were found on sectioning, a few cells of the olfactory placode were not removed in the operation. The number of fibers varies from a very few to half as many as were found on the uninjured side which was used for control. Along the fibers


70 CHARLES BROOKOVER

were found sheath cells and a small ganglion of the nervus terminalis somewhat proportional to the size of the olfactory nerve as compared with the ganglion and olfactory nerve on the uninjured side.

Taking the inhalent nasal tube as a landmark, in some cases it appeared that more of the posterior part of the nasal placode had been removed and in other cases the anterior part. However, it did not seem to make any appreciable difference in the development of the ganglion whether the anterior or the posterior part of the nasal capsule was removed. From this fact we may infer that the ganglion of the nervus terminalis does not originate from any particular part of the olfactory placode. In the two cases where there were no olfactory fibers, no ganglion appeared, although small blood vessels were found near the brain and it is to be noted that the ganglion always appears near the blood vessels (fig. 7). From these experiments it appears, also, that the ganglion does not regenerate from the brain. The extirpation experiments taken by themselves do not absolutely prove that the ganglion does not arise from some other source near the olfactory placode, which source was disturbed in the operations that destroyed the placodes but, taken in conjunction with the embryological history of the ganglion previously given, they contribute some evidence to show that the ganglion is derived from the olfactory placode.

We may briefly sum up the early embryological history of the peripheral olfactory apparatus in Amia by saying that at a time about forty hours before hatching, two olfactory placodes exist in connection with what may be termed an unpaired olfactory placode. The latter disappears and the paired placodes immediately establish fibrous connection with the neural tube. Later, nuclei migrate out from the nasal sacs along the olfactory fibers as Disse ('96), and others have found for other forms of vertebrates. Still later some of the nuclei which have wandered out along the nerve form a ganglion, as Carpenter ('00) found for the ciliary ganglion in the chick, while others remain scattered among the olfactory fibers as sheath cells. In case of the ciliary ganglion, however, the nuclei migrate from the neural tube


NERVUS TERMINALIS IN AMI A 71

along the developing oculomotor nerve, while in the case of the nervus terminalis the migration is from the placode toward the neural tube.

The nuclei of the ganglion on the olfactory nerve of Amia, can be distinguished readily from the nuclei of the sheath cells or of surrounding mesodermic elements when the young are about 10 mm. to 12 mm. long. These nuclei are larger than any others near them and are situated on the ventro-median side, as Allis, ('97) described them, about midway between the outer wall of the neural tube and the olfactory cup. There are perhaps two dozen nuclei on each olfactory nerve at this stage. The capsule which Allis describes as surrounding the ganglion has not been very evident to me, but there are smaller cells partly surrounding it on the side next the olfactory nerve, which I have taken for sheath cells of the olfactory nerve.

The ganglion increases in size and number of cells, but generally remains a single compact mass until about the 25 mm. to 30 mm. stage. In a few cases its cells have been found distributed in two or three aggregations along the olfactory nerve. It keeps its position about midway between the anterior end of the brain (bulbus olfactorius) and the most rostral part of the developing nasal capsule. The olfactory nerve is lengthening at this time and. the brain lies farther caudad with respect to the eyes and some other structures when the fish has grown to maturity, and as Allis ('97) states, the olfactorj^ nerve in the adult is about one and one half times as long as the brain. From the first appearance of the ganglion it is located near the point where a branch of the external carotid artery joins the olfactory nerve to be distributed to the nasal capsule.

When the young of Amia are from 25 mm. to 30 mm. in length, some of the cells, of which there are about forty at this stage, have wandered rostrad from the main ganglion and are scattered among the peripheral olfactory fibers. There are some five or six folds in the mucous membrane at this stage in each of two series lying one on either side of a median fold or " mid-rib" in each nasal capsule. Earlier the opening of each nasal capsule to the outside was a slit which has now closed in the middle to pro


72


CHAELES BROOKOVER


duce two apertures. One is larger and posterior near the eye, while the other is anterior and more median near the end of the snout and has already begun to develop a slightly protrusible nasal tube for the intaking of water. As the folds of Schneiderian membrane are smaller at the anterior end of the mucosa, it appears that new folds are added at this point as the animal grows. The cells from the ganglion scatter in a line beneath

distal end



Fig. 8. Golgi impregnation of part of the ganglion cells of the nervus terminalis in young Amia about 25 mm. long. A few olfactory fibers are shown. The latter are smaller and have slight varicosities. X 333.

the median fold or mid-rib. The cells are easily recognizable because their cytoplasm takes Delafield's hematoxylin more strongly than surrounding cells. That some of the ganglion cells are true nerve cells at this stage, is indicated by the fact that Golgi preparations (fig. 8) show branching processes, — toward the brain in this instance.

The 50 mm. stage is instructive and important, as the cells have attained practically the same distribution that they have


NERVUS TERMINALIS IN AMIA 73

in the- adult. In a section cut sagittally through the head and passing along the entire length of the olfactory nerve, the nasal capsule and its mid-rib, one can see the main ganglion and the scattered ganglion cells along the ventral side of the olfactory nerve (fig. 9). About two hundred and fifty cells on a single olfactory nerve were counted at this age. Nearly half of them are in the main ganglionic mass or near it, while the remainder are distributed almost uniformily from the ganglion rostrally to near the anterior end of the nasal capsule. They lie beneath the mid-rib rather than laterally under the secondary folds. There are some thirteen of these secondary folds on each side at this age, indicated by the number of undulations of the epithelium in the section (fig. 9). As the figure shows, the cells are practically all located between the anterior and the posterior



olf.bulb


Fig. 9. Diagram from camera lucida outline to show the distribution of the ganglion cells of the nervus terminalis in young Amia 50 mm. long. Delafield's hematoxylin. X 20.

limits of the nasal capsule. One or two cells were found proximally of the ganglion along the ventral margin of the olfactory nerve, and two or three cells of the same appearance were seen ventrally near the point where the olfactory nerve joins the bulbus olfactorius.

In a series of Weigert sections cut transversely through the head of a fish whose total length was 100 mm. the ganglion cells along the olfactory nerve show quite well. Twenty cells were counted on the two sides along the proximal part of the olfactory nerves between the nasal capsules and the olfactory bulbs. Most of these cells lie within the cranial cavity. However, three


74


CHARLES BROOKOVER


or four cells were found anterior to the cranial cavity at the level of the eye-muscle canal of Allis. They are located ventromedianly between the two main rami of the olfactory nerve (fig. 10), but are more intimately associated with the median bundle. Fig. 10 may be taken as a typical cross-section of the olfactory nerve in this stage at any point anterior to the cranial cavity. In nearly every section across the region of the nasal capsules, from one to four ganglion cells may be found in the wedge between the two rami of the olfactory nerve. This is


lateral ramus



Fig. 10. Transverse section of olfactory nerve and nervus terminalis at level of the eye-muscle canal from Weigert preparation of young Amia 100 mm. long. Two ganglion cells show in the ventro-median wedge representing the nervus terminalis which is more intimately united with the median ramus of the olfactory nerve. X 210.


the portion which Allis ('97) denominated the nerve "n" of Pinkus. The fibers of this wedge, which is only occasionally slightly separated from the median bundle of the olfactory nerve by a few connective tissue fibers, show no trace of medullary substance. It appears similar to the olfactory nerve proper and cannot be traced in the series of sections except by its position and the presence of the ganglion cells. Allis has well said that there is interchange of fibers between the three bundles of the olfactory nerve.


NERVUS TERMINALIS IN AMIA


75


When this ventro-median wedge is traced posteriorly into the cranial cavity it furnishes the fibers for a separate bundle (fig. 11) which passes to a more median position. This bundle has the ganglion cells already referred to, occasionally along its course, as figs. 11 and 12 show. However, there are variations with regards to the distinctness of the nervus terminalis in different cases as is well shown on the opposite, or left, side of tnis same fish where it is never more distinct within the cranial cavit v than is indicated in fig. 10. Consequently, I was able to trace a nervus terminalis on the left side of this specimen, only by a more or less detached wedge of the median half of the olfactory


Tierire ■"nervus te rmiTidks



Fig. 11. From same series as fig. 10, but farther caudad within the cranial cavity. The nervus terminalis is distinct and shows a ganglion cell. X 210.


nerve, in which the characteristic large ganglion cells were found occasionally. It may be mentioned in this connection that I found a separate bundle of the left olfactory nerve in this fish, located on its dorso-lateral side at the level of the eye-muscle canal, but no ganglion cells were found along its course from the time it separated from the olfactory nerve until it reunited with it. On the right side, the farther the nervus terminalis is traced caudad within the cranial cavity in this Weigert series, the farther it becomes separated from the main olfactory nerve (fig. 12). This figure is drawn at a level where it includes a large


76


CHARLES BROOKOVER


ganglion cell outside the limits of the nervus terminalis and not far from a blood vessel. This section shows two other vessels which are the main branches of the internal carotid artery at this level. As the right nervus terminalis is followed caudad it becomes somewhat smaller and finally is lost in this preparation on account of poor fixation at this depth because the cranial cavity was not opened when the fish was killed and fixed. It will be shown later that the nervus terminalis joins the olfactory bulbs ventro-mesially and that the number of ganglion cells increases at this point in the adult.


|oolfddoru Tie ire



nenrns terminalis -^


Fig. 12. From same series as figs. 10 and 11, but farther caudad within the cranial cavity near the olfactory bulbs. The blood vessels are quite near the nervus terminalis. A nerve cell is located between the nervus terminalis and the most dorsal vessel shown. X 340.


In the above mentioned Weigert preparations the blood corpuscles are stained a beautiful light blue, and since the blood vessels can be traced with great ease and they probably have the same distribution in the adult, they will be described here. Allis ('97) shows that the internal carotid artery enters the cranial cavity near the optic chiasm and sends a branch rostrad. As was said previously, this artery shows beneath each olfactory nerve intracranially (figs. 11, 12). Possibly it passes forward


NERVUS TERMINALIS TN AMIA 77

beyond the cranial cavity to reach the nasal capsule, but this last point could not be determined with certainty from the preparation.

However, the main blood supply to the nasal capsules is derived from the external carotid artery. The branch which supplies each nasal capsule joins the olfactory nerve just anterior to the main mass of the ganglion of the nervus terminalis. This position of the artery with reference to the ganglion was found to be true for all stages of development subsequent to the appearance of the ganglion. The artery approaches the nerve from the ectal side and circles ventrally beneath it to come into close proximity with the ganglion. The artery divides as it runs forward and its branches passing near the scattered cells of the nervus terminalis, turn dorsad with the fibers of the olfactory nerve until they reach the basement membrane of the nasal epithelium lying along the mid-rib. From this point the arterioles turn laterally right and left, along each secondary fold of the Schneiderian membrane. An occasional nerve cell lies slightly dorsal of the olfactory nerve, but for the most part, the cells are located, as indicated in fig. 9, between the two main rami.

ADULT STRUCTURE OF THE OLFACTORY ORGAN IN AMIA

Amia lives much in swamps and Reighard ('03) has shown that when an adult is released from a boat it quickly buries itself in the slime at the bottom. The distensible nasal tube has a very small aperture which would be of advantage in filtering the water taken into the nasal capsules (fig. 13). The size of the nasal capsules, which almost equals that of a shark of the same size, indicate that Amia is macrosmatic as compared with other fishes. There are in the adults from fifty to seventy secondary folds on either side of the mid-rib of each nasal capsule (fig. 13).

Some experiments were made to determine how the water circulates in the nasal capsule. When colored fruit juice was introduced by means of a long pipette into the water of an aquarium containing a living adult, the nasal tube (fig. 13) inhaled


78


CHARLES BROOKOVER


the juice. Later the colored fluid appeared at the exhalent aperture. In this experiment the current was intermittent and synchronous with the movement of the opercles in respiration, but there is also, a continuous egress of water between exhalations. This is probably due to the ciliary action to be described next.

In a pithed fish the nasal capsules were opened as shown in fig. 13, by removal of the nasal bone of Allis ('97, plate xx, fig. 1). When studied with a high power binocular dissecting microscope,


.-nasal tube


exhalent aperture 7



-mid-rib


Fig. 13. Dorsal view of the opened nasal capsule of adult Amia to show the folds of the mucous membrane. The arrows show the direction of the water in passing through the nasal capsule. The water enters by the nasal tube, which is anterior. X 3.


the blood can be seen circulating laterally each way from the midrib along the dorsal margin of each of the secondary folds, as described previously for the young fish. The nasal capsules are very vascular and the circulation seen here reminds one of the circulation within the gills. When powdered carmine in water is introduced into the opened nasal capsule, it shows a current due to ciliary action directed posteriorly along the mid-rib and thence


NERVUS TERMINALIS IN AMIA 79

laterally down between the secondary folds, as indicated by the arrows in the figure, to emerge at the sides of the capsules into a sort of drainage trough. From this trough the water can reach the posterior opening and pass to the outside. The mid-rib is at a lower level than the lateral ends of the secondary folds and some of the water passes posteriorly along the mid-rib and directly out between some larger folds there, or over them. A mucous substance is continually thrown out from the nasal epithelium and soon entangles the powdered carmine in ropy masses. This is probably produced by the goblet cells to be mentioned next.

Intra-vitam methylen blue and various cytological preparations show that there are in Amia three main types of cells i which have also been described by various workers on fishes) in the Schneiderian membrane, viz., olfactory cells of various shapes, ciliated supporting cells, and goblet cells secreting mucus. The ciliated cells are most numerous, while the olfactory cells come next in point of numbers. A ciliated supporting cell has a larger surface on the cavity of the nasal capsule than an olfactory cell. The mucous cells are isolated among the supporting cells and can be readily recognized in preparations stained with Delafield's hematoxylin.

In the adult fish there are only one or two important changes from the 50 mm. stage, in regard to the position of the ganglionic cells of the nervus terminalis. Most of the cells come to lie in the dorsal groove between the two main olfactory rami instead of ventrally, as in the early stages (fig. 9). This position is probably due to the increase of fibers in the lateral rami as the surface of the nasal folds to be supplied increases in area. But also, it may be that the cells migrate to a point nearer the place to be innervated. There are slight variations in the position of the cells in the adult. Sometimes a greater number of cells are found ventrally of the olfactory nerve than in others. In the adult there has ceased to be any decided ganglion at the posterior end of the nasal capsule. There are a few more cells posteriorly in the nasal capsule than farther rostrad, but the surface of the folds and the number of blood vessels are greater at the posterior end. The cells are nowhere aggregated in gan


80 CHARLES BROOKOVER

glia of any considerable size. Occasionally two or three cells may lie close enough to exert some mutual pressure on each other and flatten their adjacent sides (fig. 14), but generally they stand at some distance from each other. The figure was drawn from a slightly more crowded region than is typical in order to include as many cells as possible in one drawing.



£&


/•>■


m


ill


Fig. 14. Ganglion cells of the nervus terminalis in a sagittal section of the olfactory nerve in adult nasal capsule of Amia. Taken from a slightly more crowded locality than is typical. Toluidin blue stain. X 333.

In a medium sized adult a little over one-third of a meter long, there were found to be nearly one thousand cells that were attributed to the nervus terminalis beneath a single olfactory capsule. The counting was done in sagittal sections of Nissl preparations, cut thick, and generally only such cells were counted as showed nuclei, in order to avoid counting a cell more than once. Fig. 14 is drawn from these preparations (toluidin blue stain), and shows


NERVUS TERMINALIS IN AMI A 81

tigroid bodies and the large vesicular nuclei characteristic of functioning nerve cells.

More than twenty adult nasal capsules were treated by the Golgi method, but there were only two or three doubtful cases of impregnation of the ganglion cells of the nervus terminalis. There were many fibers shown in all of the preparations. The fibers belonging to these cells can be distinguished from those of the olfactory nerve (fila olfactoria) on account of their larger size and the fact that they often branch. Also, the olfactory fibers nearly everywhere show slight varicosities in their course (figs. 8 and 15), and are of lighter color, owing perhaps to their covering of sheath cells. The fibers of the nervus terminalis branch as they rise toward the mid-rib of the nasal epithelium (fig. 15), and were often seen following the course of the arteries


ia.sal epithelium. rar. term, fibers



Fig. 15. Golgi preparation of the adult nasal capsule of Amia cut sagittally through the mid-rib. The coarse branching fibers belong to the nervus terminalis, the slender varicose fibers to the olfactory nerve. X 40.

(fig. 16). In no case were the branches followed into the epithelium of the nasal capsule, although careful search was made in both Golgi and intra- vitam methylen blue preparations where the finest end-branches were shown. In many cases they were followed into a reticulum of nerve fibers beneath the basement membrane of the epithelium (fig. 15). This reticulum is located at the point where the arterioles turn laterally along the secondary folds of the Schneiderian membrane.

Favorable preparations of the nasal capsules of adult Amia by the Cajal methods show the cells of the nervus terminalis, as well as the fibers related to them. For the reason that they show


82 CHARLES BROOKOVER

these cells and have not in my hands inpregnated the olfactory fibers or cells, they are much preferable to Golgi preparations. In some cases neurofibrils are shown (fig. 17), thus indicating that they are functioning neurones in the adult fish. There are nerve processes ending on the surface of some of the cells or near them (fig. 17). It is difficult to say whether these are axones or dendrites. This is often the case in the symphathetic system, as for instance in the myenteric plexus (Auerbach's plexus) of the small intestine of mammals. In the case of the processes of


fcVtvav.naHal. ° %


Fig. 16. Golgi preparation of adult Amia cut as above. Shows fibers of the nervus terminalis in the adult nasal capsule following an artery which is indicated by broken lines. X 80.

the nervus terminalis, they behave in the Cajal treatment like axones rather than dendrites. Cajal preparations and hematoxylin staining show that many of the cells are bipolar and spindle-shaped, but some of them are stellate with three or more processes. The third type comprising about one-half of the total number of cells, is fusiform in shape. So far as could be determined, these latter have only one nerve process that tapers gradually from the cell body like a dendrite. These are the three types of cells usually described in sympathetic ganglia.


NERVUS TERMINALIS IN AMIA 83

CENTRAL RELATIONS OF THE NERVUS TERMINALIS IN AMIA

Near the olfactory bulbs each olfactory nerve breaks up into a large and variable number of small bundles. In Nissl preparations on one of the ventro-median bundles can be found cells of the same character as those found in the nervus terminalis in the nasal capsule. These cells (fig. 18) increase in numbers as the olfactory bulbs are approached. They are fusiform with the

clistdl



Fig. 17. Fibers and cells from nervus terminalis in adult Amia by the Cajal method. Taken from the nasal capsules to show neurofibrils and fibers branching toward the periphery to end near the cells of the nervus terminalis. X 350.

pointed end generally directed toward the olfactory bulbs (figs. 18 and 19), while the opposite end is frequently capped with a cell of the same size and general appearance as the sheath cells. There are a few smaller ganglion cells, as fig. 18 shows. Nearly one hundred ganglion cells of this type were counted intracranially along the two olfactory nerves. I use olfactory nerve rather than nervus terminalis in the previous sentence because


84 CHARLES BROOKOVER

I have never been able to satisfy myself that there are not olfactory fibers in this bundle even when separate. We have seen how it is sometimes united with the olfactory nerve in the 100 mm. stage, and the previous embryological account shows that it is developed as a part of the olfactory nerve.

Nissl preparations show that cells of the same general character as those just described, are continuous over upon the surface of the olfactory bulbs without any line of demarcation. Here they lie in groups of from two to as many as ten. They are situated outside the glomerular zone among the olfactory fibers and just inside a more or less continuous layer of neuroglia (fig. 21), from which they can be distinguished by their staining reactions in the Nissl preparations. Their size is larger than that of the supporting elements, but they are only about half as


sheaik cells



Fig. 18. Cells in proximal part of nervus terminalis as it joins olfactory bulbs. Sagittal section showing the dendrites tapering proximally. Shows sheath cells characteristic of the olfactory nerve. Toluidin blue stain. X 340.

large as the mitral cells which lie at a deeper level beneath the glomerular zone. An estimate showed that there are at least two hundred cells that may confidently be said to belong to this category, on a single olfactory bulb. They are slightly more numerous on the mesial side of the bulbs but some were found in other positions as well. The majority of these cells appear to be bipolar. Catois ('01) has described such extra-glomerular cells in the olfactory bulbs of teleosts. Disse ('96) has found cells in Golgi preparations of embryo birds in the same position as those I have described in the olfactory nerve (nervus terminalis) of Amia where it joins the bulbs. Also, Rubaschkin ('03) has found peripheral cells in the frog's olfactory bulbs, which he denominates "sub-glomerular. " Similar cells have been shown by Rubasch


NERVUS TERMINALIS IN AMIA


85


kin and Cajal to be in relation to the olfactory glomeruli. It may be that their endings in Amia have no connection with the nervus terminalus.

The course of the fibers of the nervus terminalis after joining the olfactory bulbs, was not made out with any degree of clearness until a large number of Golgi preparations were made of young Amia about 75 mm. long. One hundred of these small



C&fOT

tertj


olfactotlus


Fig. 19. Reconstruction from four sagittal Golgi sections 60 micra thick, of young Amia head, total length of fish 75 mm. All non-essential details omitted, to show the course of the nervus terminalis as seen from the median plane projected upon the outline of the most lateral of the four sections. The numbers indicate the sections from which a given part was taken. X 50.


iishes' heads were treated by the Golgi rapid process and transferred from the osmium-bichromate mixture to the silver in lots of ten or more at intervals of five or six hours from the second to the fifth day after immersion in the fixing fluid. In two of the earliest of these lots the fibers of the nervus terminalis were impregnated, while few or none of the fila olfactoria were shown. Fig. 19 is reconstructed from camera drawings of four sagittal


86 CHARLES BROOKOVER

sections cut 60 micra thick and represents the nervus terminalis with all non-essential details omitted, as seen from the median side of the olfactory bulb and projected upon the deepest or most lateral of the four adjacent sections. This is one of more than a dozen fishes that show essentially the same thing.

The above preparations show the nervus terminalis with varying degrees of distinctness from the region of the main ganglion of cells rostrad of the eye-muscle canal along the ventro-median edge of the olfactory nerve until its fibers join the olfactory bulb. A very few impregnated cells have been seen in my preparations (figs. 19, 4), but that this bundle is the nervus terminalis rather than fila olfactoria, is shown by the following facts: it is composed of slightly coarser fibers than the fila olfactoria, arises from the ventro-median part of the olfactory nerve, turns ventro-caudad over the median surface of the anterior one-third of the olfactory bulbs, then caudad, and finally ventro-caudad into the prosencephalon proper. Some of its fibers may end in the olfactory bulbs, but a number of them continue into the forebrain.

The course of the nervus terminalis as seen in horizontal sections is shown in fig. 20. This figure was reconstructed from Golgi preparations in the same way as fig. 19. Fig. 20 shows the nervus terminalis as seen from the ventral side of the brain, projected upon the outline of the most dorsal of the five sections showing the nervus terminalis. In consequence of the more ventral sections of the olfactory bulbs being smaller in area and the fact that the mass of the bulbs was slightly shrunken and contracted away from the median surface of the olfactory bulbs, the nervus terminalis appears more deeply embedded in the bulbs than is really the case. The real depth of the nervus terminalis is more accurately shown by fig. 21 which is taken from a Cajal preparation of the median side of the adult olfactory bulb at about the middle of its antero-posterior extent. Also, fig. 21 shows the cells previously mentioned as of a different nature from the mitral cells, and as being more superficial in position. In one or two Golgi preparations I have found some evidence that fibers believed to belong to the nervus terminalis end in relation to cells on the surface of the olfactory bulbs medianly.


NERVUS TERMINALIS IN AMIA


87


The maximum number of fibers of the nervus terminalis impregnated in any one of the Golgi preparations of fishes 75 mm. long has not exceeded twenty. One cannot say whether these are all of the fibers that belong to the nervus terminalis, but often




ievDu,stettain&Yi\


oUa.c\oru W\\>


olfactory


Fig. 20. Golgi preparation from a young Amia about 75 mm. long. Reconstructed from camera lucida drawings as in fig. 19, but taken from horizontal sections and shows the nervus terminalis as seen from the ventral side of the olfactory bulbs projected upon the most dorsal of the sections. X 50.


the Golgi process impregnates a majority of the fibers of a given kind. My impression from all the preparations and the difficulty with which the fibers were found in any of the preparations of young or adult, is that there are not more than about twenty


88


CHARLES BROOKOVER


five fibers in the nervus terminalis at this age at the point where it joins the olf actor}' bulb, although there are not less than two hundred and fifty ganglion cells peripherally at this time. The maximum number of fibers at this point in the adult as shown by the Cajal process, which is supposed to show all the fibers of a given kind, did not exceed forty. Herrick ('09) and Sheldon ('09) traced the nervus terminalis posteriorly into the anterior



A^Thirme H.U


Fig. 21. Cajal preparation cut perpendicularly to the median surface of the olfactory bulbs of adult Amia, 15 micra thick. Shows what were thought to be fibers of the nervus terminalis at a, and ganglion cells at 6. Supporting elements on the median surface of the bulbs at c. X 444.


commissure, but the bundle is diffuse in Amia and has not been traced into the commissure as yet.

As noted in an early part of this paper, Allis traced aroot of the nervus terminalis posteriorly ventral of the prosencephalon to the region of the optic chiasm, but I failed to find it in gross dissections. However, I have frequently found a bundle of non-medullated fibers accompanying the internal carotid artery of each


NERVUS TERMINALIS IN AMIA


89


side in this position. Fig. 22 taken from a single sagittal section of a Golgi preparation of a young Amia 75 mm. long shows the relation of the nerve fibers to the artery at the anterior end of the olfactory bulbs in the position where the internal carotid artery is shown in fig. 19. It will be noted (fig. 22) that a fairly compact bundle of five or six fibers accompanies the main artery. From near it two fibers run dorsad along a blood vessel. The



^f^olj^cW»j \su\ls


9 - —


Fig. 22. Sagittal section in the median plane from a Golgi preparation of a young Amia about 75 mm. long. Shows a bundle of about five fibers following the course of the internal carotid artery at the level of the anterior end of the olfactory bulbs. The fibers become mingled with those of the olfactory nerve in the region of the nervus terminalis, in the adjacent section. X 225.


main bundle, which can be followed from posterior to the sulcus olfactorius between olfactory bulb and hemisphere, runs forward beyond the olfactory bulbs to be lost among the fibers of the olfactory nerve. I have found fibers behaving like those that run dorsad in fig. 22 in Cajal preparations accompanying the small arteries among the bundles of the olfactory nerve in adults at the level of the anterior part of fig. 19. Also, I have


90


CHARLES BROOKOVER


more than once found fibers accompanying the blood vessels that run into the adipose tissue of the cranial cavity rostrad of the olfactory bulbs.

In Cajal preparations of adults I was able to trace a bundle of fibers, not exceeding fifteen in number, from the olfactory bulbs posteriorly to the region of the optic chiasm, in the same relation to the internal carotid artery as shown in the Golgi preparations just mentioned (see also fig. 32). In one Cajal preparation the fibers seemed to diminish in numbers posteriorly, but I was not able to connect this bundle with the nervus terminalis on account of a defect in the preparation, as I judged, although I succeeded in tracing the bundle within less than a millimeter of


ol^cto™ Wlb



Fig. 23. Golgi preparation from young Amia as for previous figure. Shows what was thought to be a nerve cell beneath the olfactory bulb, the most posterior of any cells in the meninges attributable to the nervus terminalis. X 2*25.


the bundle of the olfactory nerve containing nervus terminalis fibers.

What was thought to be a nerve cell with its processes branching in the neighborhood of the internal carotid artery beneath the olfactory bulbs (fig. 23) was found as far as posteriorly as the bit of artery shown in fig. 19. This is the farthest caudad that a cell separated from the main branch of the nervus terminalis has been found within the cranial cavity, except the groups of nerve cells to be described later. The position of the nerve cell in this instance (fig. 23) appears quite similar to that of the


NERVUS TERMINALIS IN AMIA 91

undoubted nerve cell shown in fig. 12 near the nervus terminalis. Golgi preparations of adult Amia show that there are branchingnerve fibers along the arteries at the anterior end of the olfactory bulbs (fig. 24), which are apparently derived from the nervus terminalis. They are more numerous ventral of the olfactory bulbs, but they have been found on every side of the bulbs and oftentimes seem to arise from the bulbs or are following the vessels to the very surface of the bulbs. There is a rich network of vessels at the surface of the olfactory bulbs and injected Nissl preparations, as well as Golgi impregnations, show that every large mitral cell has an arteriole coursing along the surface of its two or three main dendrites.


Fig. 24. Golgi preparation cut sagittally from adult Amia. Shows nerve fibers proceeding posteriorly from the position of the nervus terminalis, near where it joins the olfactory bulb ventro-mesially. X 225.

From the above description, from the findings of Allis, and from the fact that Pinkus found the nervus terminalis running far caudad ventral of the brain in Protopterus, it would not be unnatural to infer that the nervus terminalis in Amia extends caudad of the olfactory bulbs. However, we shall be in a better position to judge of this matter after a description of certain fibers found at all levels in the meninges of the prosencephalon of Amia.


92 CHARLES BROOKOVER

INTRA-CRANIAL SYMPATHETIC SYSTEM POSTERIOR TO THE NERVUS TERMINALIS IN AMIA

In addition to the fibers already described as occurring in the cranial cavity on the blood vessels, there were found in numerous Cajal and Golgi preparations a number of fibers among the blood vessels and glandular tubes of the paraphysis. These fibers could often be graced into a more or less distinct bundle near the posterior lateral portion of the paraphysis. This is just outside the brain membranes near the anterior edge of the optic tracts. As Huber ('99) had found nerves entering the cranial cavity of mammals along with the internal carotid artery near the optic chiasm, search was made for a long time to discover the entrance of fibers at this point in Amia, but to no purpose. Nerve fibers were found entering the cranial cavity dorsally opposite the anterior end of the epiphysis (fig. 25) through a foramen by which a vein apparently leaves the cranial cavity. This foramen is near the alisphenoid ossification (Allis, '97) and is probably the one marked "foramen for the anterior cerebral vein" (plate xxi). In Weigert preparations of the head of a young Amia 100 mm. long there are seen to be five or six medullated nerve fibers entering at this point. There are probably some non-medullated ones as well, From dissections of adults there were found to be three rami of this bundle diverging from a point just inside the cranial cavity (fig. 25). The first ramus runs mesially at right angles to the long axis of the body to a point near the anterior end of the stalk of the epiphysis. A second branch runs forward to be lost in gross dissections in the fat surrounding the brain. The third and largest ramus turns posteriorly to a point just in front of the habenular body and is distributed forward among the tubules of the paraphysis, as already mentioned. Some fibers are sent ventrally along the arteries to the neighborhood of the hypophysis, where nerve fibers have been found on the blood vessels. Gross dissections show some variations in size and distribution of these rami in different specimens.

In one adult Amia there were found more than thirty large nerve cells along the intra-cranial part of these rami. About


NERVUS TERMINALIS IN AMIA


93


half of them were situated near the entrance of the nerve into the cranial cavity (at point a in fig. 25), while the remainder were found along the posterior branch of the nerve near the paraphysis and the pallial wall of the forebrain (at point b in fig. 25). The latter ganglion is situated in close proximity to the base of the



rue


cartilage


--ant. di.en.ceph.aUc sac


-ciliary dano"! ion -pineal stalk -■--profundus dantflion paraph, ysis profundus nerue


Fig. 25. Diagram showing dorsal view of the anterior end of adult Amia brain. Made from gross dissections with the aid of Weigert preparations of a young Amia 100 mm. long, to show the relation of the profundus nerve to the ramus which innervates the meninges of the forebrain. Also, it shows the relation of the preoptic sympathetic system to the pineal nerve and to the nervus terminalis. X 4.

dorsal sac (fig. 30) and near the origin of the diencephalic sacs of Kingsbury ('97), from the third ventricle. In total mounts of the intra-cranial part of the nerve, sheath cells can be seen surrounding the large nerve cells (fig. 26). The cells show tigroid bodies and are of the same size and character as those of the pro


94 CHARLES BROOKOVER

fundus nerve outside the cranial cavity. However, some of the cells are only half as large as others and the smallest are about the size of the cells of the nervus terminalis.

The fibers that enter the cranial cavity arise from the ramus ophthalmicus superficialis trigemini where the latter is joined by a portion of the profundus nerve as it extends forward from its main ganglion (Allis, '97, plate xxx, fig. 39, opt). From Weigert sections it seems pretty certain that most, if not all, of the fibers that enter the cranial cavity arise from the profundus nerve. As the profundus nerve furnishes a connection between the posterior part of the sympathetic system and the ciliary ganglion, it may well establish a connection from the sympathetic system



Fig. 26. A portion of the ganglion cells from point b in fig. 25. Camera drawing from the most crowded portion of the ganglion, the meninges being mounted without sectioning. Shows sheath cells around the nerve cells and Nissl bodies within. X 200.

to the cells inside the cranial cavity, and to the nervus terminalis. It may be said here, that there are nerve cells along the profundus nerve up to the point where it joins the ramus ophthalmicus superficialis trigemini and sends its fibers into the cranial cavity. Also, in Weigert preparations the profundus nerve was found to send fibers into the trochlear nerve as it passes near (fig. 25).

An interesting question arises as to the origin of the nerve cells within the cranial cavity of adult Amia. Total mounts of the nerve did not always show the cells and when found, they are quite variable in number. Moreover, it appears that the nerve fibers as well as the cells are sometimes asymmetrical as regards the two sides. In the cases noted there was greater development


NERVUS TERMINALIS IN AMIA 95

on the right side of the specimen. As the cells are very much like those of the profundus nerve outside the cranial cavity, it was at first inferred that they migrate into the cranial cavity along the fibers from the profundus nerve. This seemed all the more probable because the cells inside the cranial cavity were not found in sections of young specimens, but the number of cells is very small in the adult and might be overlooked very easily in the young among so many blood vessels. There is an interesting observation to be mentioned here as having a possible bearing on the origin of the posterior group of cells within the cranial cavity. This group lies not far from the position of the evanescent thalamic nerve described by Miss Piatt ('91) in Acanthias. In fact, in one Cajal preparation of adult Amia, it was thought that a few fibers of the thalamic nerve were found entering the brain laterally anterior to the habenular bodies. For some reason there was a break in the continuity of the fibers at the brain wall. Inside the brain, fibers were seen running from this point. As no evidence of a nerve in this location has been found in any other specimen the presence of a thalamic nerve in Amia is in doubt.

When the nerve cells at this point (b in fig. 25) are examined in surface views of total mounts, they are found to be well scattered. Fig. 26 shows five cells in their natural relation to one another in the most crowded portion of the ganglion of nearly twenty cells. Some of these cells are closely applied to the membranous pallial wall of the forebrain. In Cajal preparations I have found one or two nerve cells lying near the median line among the paraphysis tubes situated between the dorsal sac and the pallium of the forebrain. Fibers were often found here among the paraphysis tubes establishing what may be considered as a commissure between the two halves of the intra-cranial sympathetic system.

As already mentioned, Golgi impregnations show that the blood vessels everywhere within the cranial cavity have nerve fibers branching on their walls, but it is not so easy to determine whether the paraphysis and the ciliated epithelium of the dorsal and diencephalic sacs have their intrinsic nerves or not. The paraphysis is a glandular structure with tubes gathering to a duct which pours its secretion into the brain ventricle just beneath the middle


96


CHARLES RROOKOVER


of the pineal stalk (fig. 30). Kingsbury ('07. a, fig. 4) shows the opening of to the duct of the larval paraphysis. There are many instances in Golgi impregnations that seem tome to show that there is an intrinsic nerve supply to its tubes (fig. 27). Fibers often lie in the closest proximity to the blood vessels on one side and to paraphysis tubes on the other. Also, fibers are occasionally found between the tubes of the paraphysis and the ciliated epithelium of the pallium {fig. 27). In the basement membrane of the dorsal and diencephalic sacs there are often found nerve fibers in the same intimate relation with the epithelium on one side and the

forebrain ventricle X^^N^^cd Lated epithelium


^■pT; -- jjarapkysis tube



Fig. 27. Golgi preparation of the meninges of adult Amia showing the relation of the nerve fibers to the paraphysis tubes and to the blood vessels near the pallial covering of the forebrain. X 200.


blood vessels on the other. The richest supply of nerves is found among the paraphysis tubes to which the majority of the medullated fibers entering the cranial cavity were traced. The next richest supply of nerves is furnished to the blood vessles near the ciliated pallial epithelium, but some fibers are found on the walls of the blood vessels in all positions in the cranial cavity.

The ciliated epithelium just mentioned merits a closer examination into its structure and function. It may be said at the outset that there is the same essential structure of the pallium of the


NERVUS TERMINALIS IN AMIA


97


forebrain, of the dorsal sac, and of that side of the diencephalic sacs farthest from the brain wall. The ental side (toward the brain) of the diencephalic sacs is made up of delicate flattened epithelial cells, as Kingsbury ('97, a) has noted. Also, he has pointed out the fact that the high columnar epithelial cells of the pallium and diencephalic diverticula are glandular in appearance and that they have a more copious blood supply than the flat cells on the ental side of the diencephalic sacs.

Favorable fixations and staining in iron hematoxylin show that these columnar cells are ciliated (fig. 28) with from three to six long stout cilia to each cell. Portions of the live epithelium mounted in normal saline solution show that the cilia are active in producing motion in the encephalic fluid. Strips of the epithe


Fig. 28. Section of the membranous pallium of the forebrain of adult Amia. Shows the motile cilia, the granular contents of the columnar epithelium and its cuticular border. Two "mast" or wander cells are shown between the cells of the epithelium proper. Iron hematoxylin stain. X 444.


lium running parallel with the long axis of the brain show by the motion of the blood corpuscles which have escaped from the vessels, that the general direction is anterior in the common forebrain ventricle. Strips cut from the ventricular walls of the forebrain showed that there is ciliary action by its cells also, producing motion rostrad along the slit between the halves of the prosencephalon. Likewise, there is ciliary action on the walls of the rhinoccels. The few experiments made seem to show that the return path of the encephalic fluid runs laterally from the rhinoccels and posteriorly along the lateral everted portion ( Kappers '07) of the forebrain. The cilia have basal bodies and the free borders of the epithelial cells possess a cuticula which is striated


98 CHARLES BROOKOVER

perpendicularly to the surface of the cells so as to make them appear as if there were a second set of shorter cilia. The contents of the cells are granular with large nuclei located slightly deeper than the center of the cell. Frequently there appear to be two or three nucleoli. There are vacuoles in the cytoplasm in certain preparations, but not enough work has been done on the finer structure of the cells to determine whether fat or other substances have been dissolved in the treatment with alcohol or not. In preparations stained with intra-vitam methylen blue or by the Nissl method, there appeared certain cells occasionally, that were a deeper blue than the majority of the columnar cells. They were brought out in some of the iron hematoxylin preparations in which they seemed not quite so granular as the other epithelial cells. They are closely applied to the basement membrane from which they taper to a narrow end at the free surface. When the epithelium is viewed from its deeper surface, these cells show radiating arms that seem almost to set them in connection with one another. Thus they look very much as if they form a nerve plexus. These cells call to mind the supporting cells which Johnston ('01) described among the ciliated epithelial cells of the saccus vasculosus in Acipenser.

Our knowledge of the structure and function of the diverticula of the neural tube of vertebrates is not very extensive. Meek ('07) has studied the choroid plexus of the lateral ventricles of some mammals and has shown that there is a single layer of cubical cells in theepithellium. These are ciliated in the young but devoid of cilia in the adult. He found motile cilia in the adult on the ependymal walls of the ventricles. He shows that nerves are present in close proximity to the blood vessels and that the epithelium sometimes has intra-cellular fat globules. There is sometimes a cuticular border on the free surface of the epithelial cells of the choroid plexus that, as in Amia, gives them the appearance of being ciliated. Johnston found nerve fibers in the basement membrane of the ciliated epithelium of the saccus vasculosus in Acipenser and, as we have seen that cilia are present and cause motion in the encephalic fluid in Amia where there are nerve fibers in the basement membrane, we might infer that cilia are con


NERVUS TERMINALIS TN AMIA 99

trolled by nerves. However, Putter ('03) states that it has not been proven that ciliary action is under the control of the nervous system except perhaps in a few molluscs and annelids.

The pineal stalk of Amia has an innervation that shows some characteristics of sympathetic nerves. Cajal preparations show that there are more than fifty nerve fibers joining the brain from the pineal stalk. Some of these fibers pass to the habenula>, but a larger number turn caudad to the region of a gland-like structure beneath the superior commissure. Johnston ('01, p. 108) has called this structure in Acipenser the epiphysial sac and shows that nerve fibers end among its cells but he tells me these fibers do not come from the epiphysis. It is prominent in many preparations of Amia since it stains deeply with methylen blue

pineal stalk ciliated epithelium /•'"" )



Fig. 29. From the same preparation as fig. 27, showing the relation of nerves of the meninges to the pineal stalk. X 64.

and with hematoxylin. Most of the nerve fibers from the pineal stalk are soon lost in all of my preparations near the walls of the third ventricle, after passing the glandular epiphysial sac.

In a fortunate Golgi preparation of a young Amia about 12 mm. long I was able to show the details of a neurone of the pineal stalk (fig. 31), in a single section. This neurone has its cell body in a position near the distal end of the stalk. Its dendrites branch near the surface of the stalk, while its axone passes centrally to be lost between the habenula and the superior commissure. At this time the pineal stalk comes into contact with the ectoderm and I have often noticed in very young specimens of about this same size that there is a light spot in the skin free of


100


CHARLES BROOKOVER


pigment in this location. However, as the young fish grows, bone and cartilage intervene between the skin and the pineal stalk. The distal end of the pineal stalk in the adult fish is slightly enlarged (fig. 30) and adheres to the cartilage on the roof of the brain case. The nerve fibers and cells of the adult pineal stalk of Amia were found to be particularly susceptible to methylen blue used intra


cereb. optic lobe

habenul


dorsal sac



epiphysis

paraphysis tubes

paraphysis duct velum transversum pallial fold

rhinocoele


ant. com.


olfactory bulb


Fig. 30. View of the left half of the anterior part of the adult brain of Amia as seen from the median plane, to show the relation of the pineal stalk, the paraphysis, and the diencephalic sacs to the brain ventricles. Partially schematic from camera lucida outlines. X 6. a. d. s., anterior diencephalic sac; ant. com., anterior commissure; cereb., cerebellum; hyp., hypophysis; z. I., inferior lobe; mam., mammillary body; m. I. dorsal sac, median and lateral portions of dorsal sac; op. n., optic nerve.


vitam. A number of fine impregnations were procured by injecting methylen blue through the ventral aorta just anterior to the heart. Fig. 33 shows a drawing made from such a preparation mounted without sectioning and represents the pineal stalk at about the middle of its length. The longitudinal nerve fibers already mentioned as occurring in Golgi and Cajal prepara


NERVUS TERMINALIS IN AMIA


101


tions, are shown in numbers, with nerve cells at intervals. The nerve cells have from one to four branching processes establishing anastomoses which set the longitudinal fibers and the cells in connection with .each other. When the cells are examined with a high magnification there does not seem to be any marked differentiation into axones and dendrites (figs. 34 and 35). All the cell processes look quite similar, as is the case in an intestinal sympathetic plexus. Moreover, the fibers appear to be continuous from one cell to another. The cytoplasm of some of the cells shows a copious supply of tigroid bodies (figs. 34



Fig. 31. Approximately transverse section of a Golgi preparation of young Amia about 12 mm. long, to show the details of a neurone of the pineal stalk. X 200.


and 35). There is an epithelium lining the tube of the pineal stalk. The nerve cells and fibers lie among the bases of these epithelial cells near the basement membrane. The structure of this epithelium has not been studied to learn whether it shows evidence of being glandular or not, but I have noticed a set of capillaries here with finer meshes. The enlarged distal end of the stalk has essentially the same structure and manner of innervation as the stalk proper.


102 CHARLES BROOKOVER

In fig. 33 there is a transverse fiber that runs across the stalk of the epiphysis at a higher level than the longitudinal fibers and appears to leave the confines of the pineal stalk. This seems to provide for connection with the nerves already described among the tubes of the paraphysis. Also, in a number of Golgi impregnations the fibers of the nerves of the meninges reach the basement membrane of the pineal stalk (fig. 29), if indeed, they do not pass into it. This provides a nervous mechanism capable of correlating the blood supply or the fluid secretions of the paraphysis and the pineal stalk.

In the diagram of the intra-cranial course of the sympathetic system of Amia (fig. 25), I have connected the fibers from the profundus nerve with the pineal stalk on the above evidence. The pineal structures of vertebrates are generally considered atavistic remnants of a former eye. If such an eye had a sympathetic component, the nerve supply to its stalk in Amia might perhaps be considered as remaining after sight degenerated. Or there might have been a complete change of function resulting in the present nervous structure of the pineal stalk. However that may be, it can be seen (fig. 25) that there is provision for a longitudinal connection between the supposed ancient nerve, the hypothetical thalamic, and the post-optic sympathetic system. It should be explained in this connection that point "b" is anterior to point a (fig. 25) in young specimens of Amia so that the sympathetic chain runs forward instead of bending posteriorly. The brain is carried farther caudad in its cavity as the fish comes to maturity. The above described sympathetic chain furnishes connections with the posterior part of the sympathetic system for two of the three supposed neuromeres which Johnston ('05) and others have assigned to the forebrain. The last link in the forward extension of the sympathetic chain is probably represented in the connection of the intra-cranial sympathetic just described with the nervous terminalis, now to be considered more fully than before.

In discussing the central connections of the nervus terminalis I have pointed out the fact that non-medullated fibers were found along the arteries ventral of the olfactory bulbs and that this


NERVUS TERMINALIS IN AMIA


103


bundle of fibers passes caudad of the sulcus olfactorius. Fig. 32 shows a reconstruction from nine horizontal Golgi sections drawn with the aid of a camera lucida. The same essential facts are shown by a number of other Golgi preparations of fishes 75 mm. long which belong to the same lot of fishes previously mentioned as showing the nervus terminalis favorably. On the



Fig. 32. Reconstruction from ten camera lucida drawings of consecutive sections of Golgi preparations of young Amia 75 mm. long. Shows the course of the fibers following the internal carotid artery from the optic chiasm to the region of the nervus terminalis. Arteries indicated in broken lines. The numbers indicate the section from which parts were taken. X 25.


left side, which is more fully impregnated (fig. 32), a bundle of fibers follows the main branch of the internal carotid artery from the region of the optic chiasm to the region of the nervus terminalis. The arteries are plainly visible in the preparations and are shown by broken lines in the figure. The main bundle of fibers


104 CHARLES BROOKOVER

impregnated does not show more than five or six fibers, and there is slight diminution in numbers forward. As we have seen (fig. 22), there may be as many as six fibers impregnated at this age at the level of the anterior end of the olfactory bulbs. It will be noted (fig. 32) that a lateral branch is given off medianly at the level of the anterior edge of the anterior commissure which is drawn in position by the aid of the camera lucida. There is always a pair of arteries entering the forebrain at the level of the anterior edge of the anterior commissure and in one instance in Cajal preparations of adults I thought I found a fiber entering the brain along with the artery, but this is the only instance I have ever been able to find of fibers entering the ventral surface of the brain between the posterior part of the olfactory bulbs and the optic chiasm, although I have searched carefully in all of my preparations.

On the. opposite side (fig. 32) fibers reach the same point near the anterior commissure and continue forward along the blood vessels near the mid-line as far as the anterior median margin of the olfactory bulbs. At the level of the sulcus olfactorius the internal carotids always give off one or more branches medianly. These branches turn dorsad in the fold of pallium which Kappers ('07) shows as separating the common forebrain ventricle into lateral halves at its anterior end (fig. 30). In a fish which had its medulla oblongata and spinal cord pithed, I have watched the blood circulating dorsad in this region to the neighborhood of the anterior end of the pineal stalk (fig. 30). In Cajal preparations I had traced fibers along the course of these arteries from a compact bundle of six or eight fibers at the level of the posterior ventral margin of the olfactory bulbs until they scattered beneath the paraphysis (fig. 30). Finally in three out of four Golgi preparations of adults made at one time, I confirmed my findings in Cajal sections and came to regard the fibers in this position as a constant feature. They are non-medullated and their maximum number does not exceed twelve in adult Amia. Almost invariably when the conditions have been favorable for impregnation of these fibers in their protected position they have failed to impregnate in Golgi and Cajal preparations the fibers in the


NERVUS TERMINALIS IN AMIA 105

region of the posterior ventral margin of the olfactory bulbs . If one had found the nervus terminalis but had failed to find the intracranial sympathetic fibers posterior of the olfactory bulbs, he might have supposed that the fibers in the fold of the membranous pallium between the forebrain and in the meninges ventrally were roots of the nervus terminalis as Allis apparently did, but I have satisfied myself from my preparations that the bundle of fibers in the median fold of the membranous pallium is a part of the system of fibers innervating the meninges of the fore-brain (figs. 22 and 32.)

In concluding the description of the intra-cranial sympathetic fibers of the meninges of the forebrain of Amia, it can be said that there is ample opportunity for connection between the nervus terminalis and the post-optic sympathetic system. There is a constant bundle of about six non-medullated fibers that can be traced in fishes 75 mm. long from the optic chiasm to the -region of the nervus terminalis (fig. 32). Cajal preparations of the adult show that there may be three times as many fibers in this bundle at maturity. This bundle has not been traced into the brain near the optic chiasm as Allis ('97) seemed to think might be the case, but in many preparations I have connected it by a bundle of fibers following the blood vessels dorsad, with the fibers entering the cranial cavity from the profundus nerve (fig. 25). The nature of the Golgi impregnations on which I have had to depend to a large extent for tracing these intra-cranial fibers does not permit of demonstrating the connection between the nervus terminalis and the posterior portion of the sympathetic as clearly as would be the case with medullated fibers by the Weigert method, but the slightly diminished bundle of fibers of fig. 22 certainly continues rostrad along the carotid artery beneath the olfactory nerve, while the fibers of the nervus terminalis just as certainly become more or less distinctly separate from the olfactory nerve, after it enters the cranial cavity, and run near this same artery (figs, 10, 11, and 12).

It seemed quite probable to me at first that a connection might exist between the peripheral ganglion cells of the nervus terminalis in the nasal capsules of Amia and the post-optic sympathetic


106 CHARLES BROOKOVER

system by way of the fifth nerve, since a sphenopalatine ganglion is mentioned in anatomical works in connection with the fifth nerve which sends fibers into the nasal capsule of higher vertebrates and Sheldon ('08) reported a ramus of the trigeminus nerve running into the nasal capsule of the carp. But in none of my Weigert preparations could I find fibers in Amia entering the nasal capsule in close proximity to the ganglion cells of the nervus terminalis. It may be that a few non-medullated fibers reach the ganglion cells here but no clear evidence of such a condition was found in searching my Golgi and Cajal preparations. The combined ophthalmic branches of the 5th and 7th nerves pass above the cavity of the nasal capsule in the membrane lining the inner side of the nasal bone (Allis/ 97, fig. 20), while the superior maxillary branch of the 5th and the buccalis branch of the 7th nerve send rami beneath the nasal capsule. It may be mentioned in this connection that no ganglion cells were found in Amia on the ophthalmic branch of the 5th nerve anterior to the point where it sends its ramus into the cranial cavity (fig. 25). Yet this may be different in other fishes and probably any branch found entering the cranial cavity of other fishes will be smaller than in Amia on account of the high degree of development of the meninges of the latter. It may be mentioned here that I have found nerve fibers along the blood vessels ventral of the forebrain of Ameiurus in Golgi preparations of young fishes, and that these preparations show evidence of essentially the same innervation of the pineal stalk as was found in Amia.

THE NERVUS TERMINALIS OF LEPIDOSTEUS AND TELEOSTS

As has already been mentioned, a ganglion was found on the olfactory nerve of Lepidosteus in a similar position to the one discovered by Allis in Amia. It can be readily recognized in specimens longer than 10 mm., and is located on the ventro-median side of the olfactory nerve. Stages of known age at close intervals have not been available for working out its early embryonic history. To give a detailed account of its later embryology would be to repeat much of what has been said of Amia. Con


NERVUS TERMINALIS IN AMIA 107

sequently, only a few points will be mentioned here. In a specimen about 12 mm. long cut sagitally, there was found a slender fibrous connection from the ganglion of about a dozen cells, running posteriorly beneath the olfactory nerve to the brain wall. As there was a small blood vessel on the brain wall just at this point, it cannot be said whether the fibrous connection is a root of the nervus terminalis or a fiber to the blood vessel. The cells are embryonic in appearance and many of them certainly do not possess nerve fibers until a much later date. As in Amia, the ganglion develops during very late embryonic stages. In a Lepidosteus over 85 mm. long, the ganglion of fifty cells or more, is located about half way from the olfactory bulbs to the nasal capsule; but in adults the main mass of cells lies peripherally in the nasal capsules. Nissl preparations of adult nasal capsules show ganglionic masses of these cells lying among the olfactory fibers at the base of the main folds of the Schneiderian membrane, of which there are about twelve. Also, the ganglion has been found in the young of what was thought to be the short-nosed gar (Lepidosteus platostomus).

While this manuscript was being written, preparations have been made which show conclusively that we have in teleosts a nerve very similar to the nervus terminalis of Amia and Lepidosteus. In the carp (Cyprinus carpio) there were found about three hundred ganglion cells scattered along a more or less distinct and separate strand of the olfactory nerve of a specimen about one-fourth meter long. In the historical sketch we have already cited Sheldon ('09) as having found in the carp the central connection of the nervus terminalis with the brain. The cells are somewhat larger than the sheath cells of the olfactory nerve, as in Amia, and are situated on the ventro-median side of the olfactory nerve in the nasal capsules. Their number diminishes as the olfactory bulbs are approached. The single Cajal preparation so far made to show the fibers, makes it evident that they are slightly coarser than the olfactory fibers, as in Amia, and that they are distributed everywhere in the nasal capsules, and that the main bundle turns dorsad from the ganglion cells to the region of the mid-rib. A full description of the condition in teleosts


108 CHARLES BROOKOVER

will be reserved for another paper, but it may be said in this connection that with the. help of Mr. T. S. Jackson, the development of a ganglionated nerve in two species of Ameiurus has been worked out in detail. The paper will soon be ready for publication and will show that there is the closest similarity in the development of the olfactory nerve and the nervus terminalis in Ameiurus when compared with the account given in this paper for Amia.


DISCUSSION AND THEORETICAL CONCLUSIONS

From the embryological history of the nervus terminalis given above for Amia, Lepidosteus, and the teleosts, it is clear that it is to be considered a component of the olfactory nerve rather than a separate segmental cranial nerve. This is in accord with the condition which Locy ('99) first described for Acanthias, but later ('05) he came to consider it as arising from a separate placode in the sharks. In the historical sketch I have cited the three points of similarity between the olfactory nerve and the nervus terminalis as pointed out by Pinkus ('05), and I may add here that the ganglion cells of the nervus terminalis have never been found farther caudad than the sheath cells of the olfactory nerve among which they arise in the fishes that I have studied.

We will next consider the homology of the nervus terminalis in the fishes. Something remains to be done embryologically on other fishes and there is need of bringing Locy's second account of its development in the shark into agreement with the work on other forms before the homology can be strengthened on the side of embryology, but its adult morphology shows that it is always distributed peripherally to the nasal capsules, that it is in close proximity to the olfactory nerve ventro-mesially and enters the forebrain not far from the neuropore. Also, it has generally been recognized by ganglion cells distributed along its course or more or less aggregated into ganglionic masses. It would be strange if such a nerve were not homologous throughout the fishes. We have quoted Locy ('03) and Pinkus ('05) as thinking so, and Sheldon ('09) is of the same opinion.


NERVUS TERMINALIS IN AMIA 109

Herrick ('09) says the nerve which he found in the frog in a like position to the nervus terminalis in fishes is morphologically similar so far as our information extends. DeVries ('05) went farther and expressed himself as believing the nerve in fishes homologous with the nerve to Jacobson's organ (organon vomeronasale) in higher vertebrates and to be looked for everywhere in the vertebrate series. In reptiles Leydig ('97) found branching cells in the " inter-epithelial gland" at the base of the columnar epithelium of Jacobson's organ. Also in amphibians Rubaschkin ('03) reports nerve cells in the olfactory bulbs sending fibers peripherally into the nasal epithelium. We do not know what relation these fibers bear to the glandular structures sometimes found in the nasal capsule of amphibians.

Although Jacobson's organ is not well understood physiologically, it is pretty clear that it is morphologically a part of the nose. It develops as a cavity which evaginates from the nasal capsule, and in the adult of some macrosmatic mammals has been found by Miss Read ('08) to possess neurones similar to typical olfactory neurones. She found these to end centrally in glomeruli of the olfactory bulbs. Jacobson's organ has not been found as such in the fishes, but the olfactory nerve is quite generally divided into two rami peripherally, and this is true of some amphibians. It seems to me probable that the nerve to the organon vomeronasale is homologous with the median of the two rami in fishes rather than with the nervus terminalis, as DeVries has suggested. The nervus terminalis has generally been described as more intimately connected with the median of the two rami of the olfactoy nerve in fishes, and it may be that all, or a part only, of the nervus terminalis component is included in the nerve to Jacobson's organ in different species of higher vertebrates. In macrosmatic animals a large number of olfactory fibers remain, while in other cases the nerve to the organon vomeronasale may contain only the nervus terminalis component.

The condition of the nervus terminalis in sharks offers an apparent exception to the statement made above that the nervus terminalis of fishes is more intimately associated with the median ramus of the olfactory nerve. The nerve in sharks is median,


110 CHARLES BROOKOVER

or ventromedial!, of the olfactory nerve centrally, but takes a dorsal position peripherally and apparently is distributed mainly to the lateral half of the olfactory capsules. However, if the nasal capsules of the shark were rotated outward and upward into the position of the nasal capsules of most other fishes, it would have the same ventro-median position.

Details of anatomy of peripheral nerves are of most value when they are brought into relation with the function of the part of the body concerned. This has been the merit of the work of Sir Charles Bell, of Gaskell, and of the American neurologists on nerve components. We have already cited Pinkus as saying we are in ignorance of the function of the nervus terminalis, while Johnston has suggested its general cutaneous nature. We can readily see that if it is of somatic type, it is sensory rather than motor on account of the peripheral position of its ganglion cells. In my own experiments upon Ami a there was no response detected in the nasal capsules when the olfactory nerve was stimulated near the olfactory bulbs with a strong faradic current. Neither did I detect changes in the rate of water flow through the nasal capsules nor blanching of the mucous membrane of the opened nasal capsules in these experiments. However, I do not think that the experiments prove that there is not vaso-motor control exercised by the nervus terminalis, since there was always much loss of blood in the operation of pithing, the parts are very small, and I have failed to get inhibition of the heart-beat when the vagus nerve was stimulated in the same way.

The fibers of the nervus terminalis probably do not belong to the same functional type as the fila olfactoria, for they differ in the type of their nerve cells, their location, time of development and central connections. No other specialized sense organ was found in the nasal capsules of the fishes studied and those described by Blaue ('84) in fishes and amphibians have been shown by subsequent authors to be collections of ordinary olfactory epithelium. As previously noted, Johnston has suggested that this nerve is somatic sensory of general cutaneous nature. In chat event it might be thought to serve the tactile sense or some unspecialized sensibility similar to that which Parker ('08) or Sheldon


NERVUS TERMTNALTS IN AMIA 111

('09) have shown to be present in the skin of fishes. It is possible that it serves a visceral sensory function although the olfactory pit seems to deve'op as an invagination from the ectoderm.

The 'ast suggestions receive but little support from the embryological development of the nervus terminalis. Both general cutaneous and some unspecialized visceral ganglia are developed from neural crest (Landacre, '08 and later unpublished observations), while the epibranchial and dorso-lateral placodes are thought to give rise mainly to ganglion cells of special sensory functions (gustatory and lateral line). The origin of the nervus terminalis from the olfactory placode, therefore, does not support the theory of its general cutaneous function. However, it should be pointed out in this connection that Beard ('88) described neural crest in the developing olfactory nerve and Johnston ('09) has just published an article in which he shows neural crest elements present in lower vertebrates in the region of the neuropore quite near the unpaired olfactory placode. The relation of neural crest with the olfactory placodes in this region has not been worked out fully in the earliest stages of Amia. It is possible that the unpaired olfactory placode (fig. 1) receives neural crest elements that enter the paired placodes to become the ganglion of the nervus terminalis later, but I was not able with my staining methods to differentiate the cells of the nervus terminalis until fully four days after the unpaired olfactory placode had disappeared. Yolk granules (fig. 1) obscure the details in very early stages of Amia and increase the difficulty of tracing possible neural crest elements through four days of embryological history.

Another possible interpretation of the nervus terminalis has alread}^ been referred to; viz: that its fibers are of sympathetic (visceral) type, probably vaso-motor. The embryological evidence here also is obscure for we know of no other case where sympathetic neurones are derived from ectodermal placodes. But if the paired olfactory placodes can be shown to receive neural crest cells, as suggested in the last paragraph, it offers a possible solution of the difficulty; for it is commonly thought that sympathetic neurones originate from neural crest (Jones, '05). We have mentioned that Carpenter ('06) found an apparent excep


112 CHARLES BROOKOVER

tion to the general rule that sympathetic ganglia are derived from neural crest. He found the cells that migrate along a motor nerve (the oculo-motor in the chick) giving rise to the sheath cells of that nerve and contributing neurones to the ciliary ganglion. The case is parallel with the condition found in the olfactory nerve and the nervus terminalis in Amia except that the migration of cells is from the olfactory placode rather than from the neural tube. Also, if it is granted that the cells in the pineal stalk in Amia are sympathetic, we have a case where the sympathetic cells originate from the neural tube direct without any apparent connection with neural crest in development.

We need further embryological and morphological data, as well as physiological evidence, in order to determine the function of the nervus terminalis. It may contain general cutaneous components along with sympathetic and possibly other elements in some forms of vertebrates. If it is largely vaso-motor, in the forms studied, as I have been led to think from the evidence, we may tentatively consider the fibers of the neurones entering the forebrain as preganglionic. The postganglionic neurones may be considered to be those that put the cells on the olfactory bulbs, and possibly some of those along the olfactory nerve intracranially into connection with the posterior part of the sympathetic on the one hand and with the cells in the nasal capsules on the other. We may summarize the evidence given in various places in the present paper that points to the sympathetic type of the nervus terminalis in Amia, as follows:

In addition to the point Allis made that the nervus terminalis develops at a time when the ciliary ganglion is developing, we have seen that its development is pari passu with the blood vessels which are always near it in the fishes I have examined. The same thing is true of other forms where the literature mentions the blood vessels. The cells in the periphery are many times more numerous than the fibers that were found entering the prosencephalon. These cells are multipolar in some instances and always aggregated into one or more ganglia or scattered like typical sympathetic ganglion cells. The same statement can be made of other forms mentioned in the literature. The fibers of


NERVUS TERMINALIS IN AMIA 113

these nerve cells, peripherally as well as intra-cranially, were often found branching along the walls of the arteries. In other cases the nerve fibers arborize about the ganglion cells of the nervus terminalis, as sympathetic fibers are supposed to do. There is ample provision in Amia for connection with the post-optic sympathetic system, and it is difficult to account for a compact bundle along the arteries beneath the olfactory bulbs and the forebrain on any other supposition.

It is evident from the literature cited that the nervus terminalis cannot be considered a nerve peculiar to primitive vertebrates, as seemed probable so long as it was found in the generalized fishes exclusively. It appears more and more probable that there is a ganglionated nerve associated with the olfactory nerve throughout the vertebrate series. Aichel ('95) cites a number of authors who have found fibers in the nasal capsules differing from olfactory fibers. Some of these fibers are described as coarser than the olfactory fibers, while others are said to be smaller. In most instances they were attributed to the trigeminus nerve, but in light of our present knowledge the whole matter needs to be gone over again to determine whether they belong to the nervus terminalis, although we know that in some instances a ramus of the trigeminus nerve enters the nasal capsule.

SUMMARY OF RESULTS

1. This paper confirms the work of Allis in finding a ganglionated nerve in Amia, which is probably homologous with the nerve first discovered by Pinkus in Protopterus, and with the nervus terminalis found by Locy in a large number of sharks.

2. The ganglion of the nervus te v minalis in Amia originates in common with the olfactory nerve, from an ectodermal placode.

3. In early stages the cells of the ganglion cannot be distinguished from the undifferentiated mass of cells which produce sheath cells of the fila olfactoria.

4. Incidentally, we have confirmed the results of recent investigators who find the olfactory neurones of the first order arising in the ectoderm and remaining there in the adult.


114 CHARLES BROOKOVER

5. About one thousand nerve cells develop from the embryonic ganglion during late stages. Most of these lie in the nasal capsules in the adult but a few are found intra-cranially on the ventral side of the olfactory nerve in a more or less distinct ramus of this nerve.

6. The cells of the nervus terminalis lie in the dorsal sulcus between the two rami of the olfactory nerve in the adult nasal capsules. A few of them are scattered slightly dorsad of this position.

7. The nerve processes of the cells of the nervus terminalis in the adult possess neurofibrils and tigroid bodies are found in their cytoplasm, thus showing that they are functional nerve cells.

8. These cells show the three main forms characteristic of sympathetic ganglion cells.

9. Some of their nerve processes follow the arteries, while others arborize about other cell bodies of the ganglion of the nervus terminalis.

10. Not more than about forty axones of the thousand ganglion cells of the .adult were found joining the olfactory bulbs. Many of these fibers pass through the bulbs to end in the prosencephalon proper.

11. Ganglion cells of the same general character were found increasing slightly in numbers as the nervus terminalis approaches the olfactory bulbs.

12. We have confirmed Allis in finding fibers that seem to belong to the nervus terminalis continuing caudad ventrally of the olfactory bulbs to the region of the optic chiasm. These fibers were not found to enter the brain, as Allis seems to have suspected they would do, probably on account of the condition described for Protopterus; but connect with the post-optic sympathetic system by way of an intra-cranial sympathetic system.

13. Golgi and Cajal preparations show that the blood vessels everywhere within the cranial cavity are innervated. Also, the paraphysis is innervated.

14. A ramus of the trigeminus nerve, probably derived from the profundis nerve, enters +he cranial cavity dorsally opposite the anterior commissure. Some of these fibers are medullated and supply the paraphysis tubes.


NERVUS TERMINALIS IN AMIA 115

15. Ganglion cells are found along this ramus of the trigeminus nerve intra-cranially. Their maximum number is about thirty on a side in the adult.

16. There is reason to think the paraphysis has its intrinsic nerve supply.

17. The epiphysis, or pineal stalk, is innervated richly with a type of cells and fibers much like the sympathetic plexus found in the intestinal walls of vertebrates. It sends about forty fibers centrally into the brain past a glandular structure at its base. Some of these fibers seem to pass to the habenulae, but the great majority were lost in proximity to the walls of the third ventricle.

18. There is some evidence that the innervation of the pineal stalk, also, is connected with the post-optic sympathetic system through the trigeminus nerve.

19. In a number of specimens there was found to be a bundle of about a dozen fibers running in the fold of pallium between the halves of the forebrain in adults. This is capable of connecting the nervus terminalis with the post-optic sympathetic system, but the main connection is probably by a bundle ventrad from the entrance of the trigeminus nerve into the cranial cavity and thence along the internal carotid artery to the olfactory bulbs.

20. The nervus terminalis has been found in Lepidosteus and in teleosts.

Buchtel College, Akron, Ohio. November 20, 1909.


BIBLIOGRAPHY

Aichel, Otto. Kurtz Mittheilung tiber den histolog. Bau der Riechschleimhaut

1895 der embryonalen Teleostier. Sitz. der Gesell. f. Morph. u. Physiol,

zu Miinchen. p. 73, Bd. 11. Allis, E. P. The cranial muscles and cranial and first spinal nerves in Amia

1897 calva. Jour. Morph., vol. 12, no. 3.

Beard, J. The ciliary and motoroculi ganglion and the ganglion of the ophthal 1887 micus profundus in sharks. Anat. Anz., vol. 2, no. 18, 19, pp. 565 575.Bedford, E. A. The early history of the olfactory nerve in swine. Jour. Comp.

1904 New. and Psych., vol. 14, p. 390.


116 CHARLES BROOKOVER

Bing and Burckhardt. Das Zentralnervensystem von Ceratodus forsteri. Anat.

1904 Anz., Bd. 25, no. 23, p. 588.

Blatje, Julius. Untersuchungen iiber den Bau der Nasenschleimhaut bei Fischen

1884 und Amphibien, namentlich iiber Endknospen als Endapparate des

Nervus olfactorius. Archivf. Anat. [u. Physiol.] Hft. 3 u. 4, p. 231.

Brookover, Chas. Pinkus' nerve in Amia and Lepidosteus. Science, n. s.,

1908 vol. 27, no. 702, June 12, pp. 913-914.

Catois, E. H. L'Histologie et L'Anatomie Microscopique de L'Encepahle chez

1901 Les Poissons. Bull. Scient. de la France et de la Belgique, Tome 36.

Cajal, S. R. y Une Methode simple pour la Coloration Elective du Reticulum

1905 Protoplasmatique. Bibliogr. Anat., T. 14, no. 1.

Carpenter, F. W. The development of the oculomoter nerves, the ciliary gang 1906 lion and the abducent nerve in the chick. Bull. Mus. Comp. Zool., Harvard, vol. 48, no. 2.

Dean, Bashford. On the larval development of Amia calva. Zool. Jahrb. Abt.

1897 /. SysL, Bd. 9, p. 639.

DeVries, Ernst. Over het Ganglion Vomeronasale. Koninklij. Akad. Weten 1905 schappen te Amsterdam, 11 May.

Disse, J. Ueber die erste Entwickelung des Riechnerven. Gesell. z. Beford.

1896 d. ges. Naturw. Marburger, Silzungb., Oct.

Herrick, C. Judson. The nervus terminalis (nerve of Pinkus) in the frog.

1909 Jour. Comp. Neur. and Psych., vol. 19, no. 2, p. 175.

His, W., Jr. tlber die Entwickelung der Riechlappens und des Riechganglions

1889 und iiber diejenige der verlangerten Marks. Verhandl. d. Anat.

Gesell. Houser, Gilbert L. The neurones and supporting elements of the brain of a

1901 selachian. Jour. Comp. Neur., vol. 10, p. 65.

Huber, G. Carl. Observations on the innervation of the intra-cranial vessels.

1899 Jour. Comp. Neur., vol. 9. p. 1.

Johnston, J. B. The brain of Acipenser. Zool. Jahbr., Abt. f. Anat. u. Ontog.

1901 Bd. 15, p. 59.

1905 The morphology of the vertebrate head from the viewpoint of the

functional divisions of the nervous system. Jour. Comp. Neur. and Psych., vol. 15, p. 175.

1906 The nervous system of vertebrates. P. Blakiston's Son and Co., Phila. 1909 The morphology of the forebrain vesicle in vertebrates. Jour. Comp.

Neur. and Psych., vol. 19, p. 457. Jones, Walter C. Notes on the development of the sympathetic nervous sys1905 tern in the common toad. Jour. Comp. Neur. and Psych., vol. 15.

p. 113. Kappers, C. U. A. Untersuchungen iiber das Gehirn der Ganoiden Amia calva

1907 u. Lepidosteus osseus. Abh. Senck. Naturf. Ges., Bd. 30, Heft 3.

1908 Die Phylogenese des Rhinencephalons, des Corpus striatum und Vorder hirnkommissuren. Folia Neuro-Biol., Bd. 1, p. 173. Kingsbury, B. F. The structure and morphology of the medulla oblongata in

1897 Fishes. Jour. Comp. 2V< ur., vol. 7, p. 1.

1897a The encephalic evaginations in ganoids. [Jour. Comp. Nt ur., vol. 7, p. 37.


NERVUS TERMINALIS IN AMIA 117

Landacre, F. L. The epibranchial placodes of Ameiurus. Ohio Naturalist,

1908 vol. 8, p. 251.

Leydig, F. Zirbel und Jacobsonsche Organ einiger Reptilien. Arch. f. Mikr.

1897 Anal., Bd. 50, p. 402.

Locy, W. A. New facts regarding the development of the olfactory nerve. Anat.

1899 Ariz., Bd. 16, p. 273.

1903 A new cranial nerve in fishes. Mark Anniversary Volume.

1905 On a newly recognized nerve connected with the forebrain of selachians Anat.Anz., Bd. 26. Meek, Walter J. A study of the choroid plexus. Jour. Comp. Neur. and Psych.,

1907 vol. 17, p. 286.

Parker, G. H. The sense of taste in fishes. Science, n. s., vol. 27, no. 690,

1908 p. 453.

Paton, Stewart. The reactions of the vertebrate embryo to stimulation aDd the

1907 associated changes in the nervous system. Mitt. Zool. Stat, zu Neapel, Bd, 18, p, 535.

Pinkus, Felix. Die Hirnnerven des Protopterus annectens. Morph. Arb. (G.

1895 Schwalbe), Bd. 4, p. 275.

1905 Ueber den zwischen Olfactorius und Opticusursprung das Vorderhirn (Zwischenhirn)verlassenden Hirnnerven der Dipnoier und Selachier Arch. Physiol. Jahrb., Sup. Halfte 2, p. 447. Platt, Julia B. A contribution to the morphology of the vertebrate head based

1891 on a study of Acanthias vulgaris. Jour. Morph., vol. 5, p. 79.

Putter, A. Die Fliminerbewegung. Ergeb. d. Physiol., vol. 2, p. 66.

1903 Read, Effie A. A contribution to the knowledge of the olfactory apparatus in

1908 dog, cat and man. Amer. Jour. Anat., vol. 7, May. Reighard, Jacob. The natural history of Amia calva Linnaeus. Mark Anniver 1903 sary Volume, p. 59.

Rubaschkin, W. Ueber die Beziehung des Nervus Trigemimus zur Riech 1902 schleimhaut. Anat. Am., Bd. 22, p. 407.

1903 Zur Morphologie des Vorderhirnes der Amphibien. Arch. f. Mikros.

Anat., Bd. 62, p. 207.

Schultze, Max. Untersuchungen uber den Bau der Nasenschleimhaut, nam 1862 entlich die Structur und Endigungsweise der Geruchsnerven bei

dem Menschen und den Wirbelthieren. Abhandl. d. Naturj. Gesell.

zu Halle, Bd, 7.

Sewertzoff, A. N. Zur Entwickelungeschichte des Ceratodus forsteri. Anat.

1902 Anz., Bd. 21, p. 606.

Sheldon, R. E. The participation of medullated fibers in the innervation of the

1908 olfactory mucous membrane of fishes. Science, n. s., vol. 27, no. 702, June 12.

1909 The nervus terminalis in the carp. Jour. Comp. Neur. and Psych., vol.

19, p. 191. 1909a The reactions of the dogfish to chemical stimuli. Jour. Comp. Neur. and Psych., vol. 19, p. 273. Sheldon, R. E. and Brookover, Chas. The nervus terminalis in teleosts. 1909 Anat. Record, vol. 2, no. 4.


118 CHARLES BROOKOVER

Von Kupffer, C. Studien zur vergleichenden Entwickelungengeschichte des 1893 Kopfes der Kranioten. Heft 1, Munich and Leipzig.

Wilson, J. G. The relation of the motor endings on the muscle of the frog to 1904 neighboring structures. Jour. Cornp. New. and Psych., vol. 14,

p. 1.


Description of Figures

Fig. 33. Intra-vitam methylen blue impregnation of the nerves of the pineal stalk (epiphysis) of adult Amia, as seen from the surface of a total mount showing the relation of the cells and their processes to the longitudinal fibers. Lateral limits of pineal stalk at b. X 408.

Fig. 34. From the same preparation as previous figure to show details of a cell at point a. Note the Nissl bodies and the similar structure of the different cell processes. X 1500.

Fig. 35. From the same preparation as the two previous figures, near the edge of the mount where there were not so many overlying layers of the meninges to obscure the details. X 1500.


NERVUS TERMINATES IN AMIA Charles Brookover


PLATE I




The Journal of Comparative Neurology and Psychology — Vol. 20, No. 2.


CORRECTI ONS

Page 122, line 23

For "15 to 16," read "17 to 18" Page 138, line 5

For "P. 19" read "P. 134" Page 143, next to bottom line

For "constituents the" read "constituents of the"


ON THE PERCENTAGE OF WATER IN THE BRAIN AND IN THE SPINAL CORD OF THE ALBINO RAT

HENRY H. DONALDSON

The Wistar Institute of Anatomy and Biology

WITH FIVE FIGURES

The object of this study has been to obtain a continuous record of the change in the percentage of water in the central nervous system of the albino rat during its life cycle, and to correlate this with the other important changes in the nervous system which are commonly recognized. These results in turn should put us in a position to determine to what extent and in what way this character may be modified.

Although it has long been known that at birth the percentage of water in the central nervous system was much greater than at maturity, yet the change in this character through the life cycle has never been systematically followed, and it thus happens that there are no other extensive records with which to make comparison. The relations of existing data to this investigation will be discussed later on.

The data used for the following study were largely obtained from the same animals which furnished the records employed for the two previous researches on the weight of the brain and of the spinal cord of the albino rat under different conditions of age, body-weight and body-length (Donaldson '08 and '09) although many cases have been necessarily excluded because the percentage of water had not been determined. On the other hand, a few new records have been added to the original series.

In carrying on this work, which has extended through a number of years, I have been greatly assisted by Dr. Hatai, as well as by two of my former students, Dr. Polkey and Dr. Whitelaw, both of whom made a number of the determinations of water under my


120 HENRY H. DONALDSON

directions, and to all of these gentlemen I wish here to express my obligations for assistance.

Technique. The determination of water has been made for the entire encephalon severed from the cord at the level of the first spinal nerve, and for the entire cord, the spinal nerves having been clipped away at their origin from the cord. The rats used were chloroformed, eviscerated and rapidly dissected. No special device for preventing evaporation during dissection was used.

The percentage of water applies therefore to the nerve structures proper, surrounded by the meninges and containing such blood as usually remains after the foregoing treatment.

The details of the technique according to which the brain and spinal cord were removed have been already given (see Donaldson '08, p. 346). Each brain or cord was placed in a small glassstoppered weighing bottle, and after being weighed in the fresh state, was dried in a closed water bath which had a temperature ranging from 85°-95° C. and then was cooled in a dessicator over sulphuric acid, and reweighed.

The brain took somewhat longer to dry as a rule than the spinal cord, but usually seven days in the water bath served to bring it to a constant weight. At various times objections have been raised to the determination of the percentage of water by the use of heat. The other method which is most approved is that of drying the material at the room temperature or somewhat above, in a vacuum over sulphuric acid. A comparison of these two methods has been made for the brain and cord of the rat, but no significant differences have thus far been found. I shall, however, reserve the discussion of the data on which this statement is based for another occasion.

The percentage of water in the brains of albino rats of different body weights. The number of cases is 409 males and 212 females. The mean values for the percentage of water in the brain for given body weights differing by 10 grams, as determined by a correlation table, are entered in table 1.

The examination of table 1 shows for the brain a relative loss of water amounting to about ten units between birth (body weight 5 grams) and the end of the series. This loss is most


- PERCENTAGE OF WATER



PERCENTAGE OF WATER

BRAIN AND SPINAL CORD

ALBINO RAT

ACCORDING TO BODY WEIGHT

MALE .FEMALE


SPINAL CORD


10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 320 330

Birth BODY WEIGHT Gms.

Chart 1

To show the changes in the percentage of water in the brain and in the spinal cord of the albino rat of different body weights. The data for the two sexes are plotted separately.


PEECENTAGE of water


121


TABLE I. The mean values of the percentage of water in the brain and spinal cord of the albino rat. 1 Both sexes arranged according to body weights, increasing by 10 gram increments.




Percentage op Water



BODY






WEIGHT


BRAIN


SPINAL CORD



Male


Female


Male


Female


Grams.






5


87.6


87.9


85.7


85.5


15


84.8


83.9


81.4


80.7


25


81.2


80.4


77.0


76.2


35


79.7


79.1


74.9


74.1


45


79.6


79.0


74.4


73.3


55


79.5


78.8


74.4


73.2


65


78.9


78.8


73.2


72.8


75


78.9


78.6


73.2


73.2


85


78.8


78.5


72.5


72.5


95


78.8


78.5


72.1


72.0


105


78.4


78.2


71.o


71.5


115


78.7


78.3


71.7


71.5


125


78.6


78.3


71.4


71.7


135


78.4


78.2


71.8


70.6


145


78.2


78.0


71.3


70.4


155


78.5


78.3


71.2


71.0


165


78.4


78.2


70.6


71.0


175


77.8


77.5


70.0


71.0


185


78.3


77.8


70.7


69.8


195


78.0


77.2


70.7


68.6


205


78.0


78.0


69.6


69.3


215


77.5


77.7


69.5


68.7


225


78.0



69.8



235


78.0


77.0


69.5


68.3


245


78.0


77.3


70.0


68.0


255


78.0



70.0



265


77.8



69.2



275


78.3


78.0


69.5


68.0


285


77.0



68.3



295


78.0



70.0



305


77.3



69.0



315


77.5



68.5



325


78.0



68.0



1 For reasons similar to those previously given (see Donaldson '08, pp. 156-157) > the individual records are not printed. These however are on file and copies c^ them may be had by application to the Director of the Wistar Institute.


122 HENRY H. DONALDSON

rapid at the time when the brain is growing most actively. Table 1 further shows that the percentages for the females are in general slightly less than those for the males of the same body weights. Chart 1, which is based on table 1, exhibits this relation.

As we shall see later, the percentage of water in the central nervous system is more closely correlated with age than with the body weight or brain weight. Nevertheless, it will most often occur that it is desired to estimate the probable percentage of water in cases where the weight of the body or brain alone are known, and the foregoing table 1 furnishes the means of doing this for animals which have been grown under the ordinary normal conditions.

It has been already demonstrated (Donaldson '06) that for a given age, the body weight of the female is less than that of the male, consequently the comparison in each case is here between males that are younger than the females with which they are contrasted, and as increasing age is an important factor tending to reduce the percentage of water, it follows that the males, which are younger, should show, as they do, a slightly greater percentage of water.

Percentage of water in the spinal cord. In the spinal cord the relative loss of water with increasing body weight is greater than in the brain, being from 15 to 16 units. Although the initial percentage is somewhat less, yet the subsequent loss is regularly more rapid than that in the brain. The percentage of water in the two sexes is related in the same way as in the brain. The observations are given in table 1 and in chart 1.

Percentage of water in relation to age. To support the suggestion that the males in the foregoing tables show a greater percentage of water, because they are younger than the females, the data have been rearranged according to age. In many cases the age was not known, and this reduces the number of records to 358 males and 169 females. The results in the form of mean values, based on a correlation table are given in table 2 and plotted in chart 2, the entries being made for ten day intervals. When thus arranged, it appears that in the brains of males and females of like age, the percentage of water is similar.


PERCENTAGE OF WATER 123

For the brain, the records show in both sexes ranges in the percentage of water in the different age groups as follows : AGE PERCENTAGES

0-10 days total range 3 units 86-89

10-20 days total range 5 units 82-87

From 20 to 100 days the range diminishes, and after this latter age it does not amount to more than one unit. The ranges for the spinal cord are less than those for the brain.

It will naturally be asked whether among individuals belonging to the same litter, reared under similar conditions and killed at exactly the same age there is any difference in the percentage of water between those having relatively heavy brains and spinal cords and those in which these organs are relatively light. This question seems to be answered in the negative by the result of 25 pairs of observations recently made.

The figures are as follows : —


PERCENT PERCENT

OF WATER OF WATER


Heavy brains 78.651 72.436 Heavy cords

Light brains 78.649 72.465 Light cords

In both instances as is seen, the differences found are too small to be significant. It may be added that the weight of the light brains was about 96 per cent that of the heavy, and similarly the weight of the corresponding light spinal cords about 93 per cent. Such differences as we find therefore among specimens of the same age must depend on some other cause than the individual variations in the weight of the central nervous system.

I feel sure that the irregularities seen in the curve for the cord, chart 2, 95-115 days, are purely incidental and would not appear on repeating the observations.

At the same time it is seen that the percentage of water in the female spinal cord after the period of rapid growth, is in general a trifle higher than in the male. This is an unexpected result. The mean difference as determined from those entries in table 2 where there are data for both sexes at a given age (i. e., up to 230-240 days) is 0.36 per cent. At the moment this difference is most readily explained as one effect of the passive lengthening


124


HENRY H. DONALDSON


TABLE 2

The mean values of the percentage of water in the brain and spinal cord of the albino

rat. Both sexes arranged according to age, increasing by 10 day increments. 2




Percentage


of Water



AGE IN DAYS


BRAIN


spinal


cord



Male


Female


Male


Female


0-10


87.4


87.4


84.8


84.8


10-20


83.7


83.4


80.5


80.3


20-30


81.3


81.6


77.2


77.1


30-40


79.4


80.0


74.3


74.8


40-50


79.2


79.0


73.9


73.7


50-60


79.0


79.3


72.9


74.2


60-70


79.3


78.8


74.5


73.2


70-80


78.9


78.8


72.9


73.8


80-90


78.3


78.8


72.8


73.8


90-100


78.7


79.0


73.0


74.1


100-110


78.3


78.0


70.0


70.8


110-120


78.6


78.7


71.3


72.5


120-130


78.3


78.2


71.6


71.7


130-140


78.2


78.0


70.0


71.0


140-150



78.0



72.0


150-160


78.1


78.0


70.6


70.8


160-170


78.2


78.3


71.0


71.3


170-180



78.0



71.0


180-190


78.0


79.0


71.0


71.5


190-200






200-210


78.0


79.0


71.0


72.0


210-220


78.3


78.3


71.0


71.7


220-230


78.7


78.3


72.2


71.0


230-240


78.5


78.0


71.0


71.0


240-250






250-260






260-270






270-280






280-290






290-300


78.5



72.0



300-310



77.4



68.2


310-320



77.3



68.0


2 Note that the values here given begin with 0-10 days, i..e, a mean age of five days after birth. Hence the initial percentages are less than those in table 1 which gives the values at 5 grams, approximately the weight at birth.


PERCENTAGE OF WATER


PERCENTAGE OF WATER

BRAIN AND SPINAL CORD

ALBINO RAT

ACCORDING TO AGE



10 20 30 40 50 60 70 80 90 100 HO 120 130 140 150 160 170 ISO 190 200 210 220 230 240

PAYS AGE

Chart 2

To show the changes in the percentage of water in the brain and in the spinal cord of the albino rat at different ages. The data for the two sexes are plotted separately. It is to be noted that the first entry is at the mean age of five days.


PERCENTAGE OF WATER 125

of the spinal cord which for a given age is relatively somewhat greater in the male than in the female (Donaldson, '09 pp. 163164). The effect of this lengthening would be to diminish the percentage of water. The influence of passive lengthening is discussed more fully later on.

Theoretic curves. When we take the more extensive series of mean values which is that for the males as given in table 1, and draw the theoretic curves based on them, we obtain the relations shown in chart 3, the entries being arranged according to body weight. The data for this chart are given in table 3. For the formulas for these curves, I am indebted to Dr. Hatai.

The formulas for the percentage of water in the brain of the male albino rat are as follows:

Up to a body weight of 30 grams

y = 99.5 - - 12.6 log (x + 3.5) [1]

and from a body weight of 30 grams on

y = 82.62 — 2 log (x - 10) [2]

In the case of the percentage of water in the spinal cord of the male albino rat we have for body weights up to 35 grams

y = 94.9 — 12.8 log (x) [3]

and from a body weight of 35 grams on

y = 85.2 — 6.5 log (x) [4]

In all these formulas y = the percentage of water and x = the weight of the body in grams.

The formulas are of the same type as those used to express the growth changes described in several previous investigations (Donaldson '08, '09; Hatai '09), and have their main value as convenient expressions of the several series of observations.

Calculations (based on table 1, down to and including the entries for 275 grams body weight) show that in general for given body weight, the females which are under these conditions relatively older as compared with the males, have a percentage of water lower by 0.37 per cent in the brain and 0.60 per cent in the spinal cord. The theoretic values for the female can therefore be obtained approximately by applying these corrections to the determinations here given for the males.

Having thus presented the data on the percentage of water


126


HENRY H. DONALDSON


TABLE 3

Giving the values of the percentage of water in the brain and spinal cord of the male albino rat, calculated according to the formulas given above. Brain: formulas 1 and 2; spinal cord: formulas 3 and 4. For comparison the observed values for the male, taken from table 1, are repeated here. Arranged according to body weights




Percentage


. of Water



BODY WEIGHT


BRAIN


male


SPINAL CORD MALE



Calculated


Observed


Calculated


Observed


grams






5


87.79


87.6


85.96


S5.7


10


85.26



82.10



15


83.50


84.8


79.80


81 4


20


82.24



78.26



25


81.17


81.2


76.98


77.0


30


80.22



75.96



35


79.83


79.7


75.16


74.9


45


79.53


79.6


74.45


74.4


55


79.31


79.5


73.89


74.4


65


79.14


78.9


73.42


73.2


75


78.99


78.9


73.01


73.2


85.


78.87


78.8


72.66


72.5


95


78.76


78.8


72.34


72.1


105


78.66


78.4


72.06


71.5


115


78.58


78.7


71.81


71.7


125


78.50


78.6


71.57


71.4


135


78.43


78.4


71.35


71.8


145


78.36


7S.2


71.15


71.3


155


78.30


78.5


70.96


71.2


165


78.24


78.4


70.79


70.6


175


78.19


77.8


70.62


70.0


185


78.13


78.3


70.46


70.7


195


78.09


78.0


70.31


70.7


205


78 . 04


78.0


70.17


69.6


215


78.00


77.5


70.04


69.5


225


77.96


78.0


69.91


69.8


235


77.92


78.0


69.79.


69.5


245


77.88


78.0


69.67


70.0


255


77 s 1


78.0


69.56


70.0


265


77.81


77.8


69.45


69.2


275


77.77


78.3


69.34


69.5


285


77.71


77.0


69.24


68.3


295


77.71


78.0


69.15


70.0


305


77.68


77.3


69.05


69.0


315


77 . 65


77.5


68.96


68.5


325


77.62


78.0


68.87


68.0


470


77.30



67.80



PERCENTAGE OF WATER


J I l_


PERCENTAGE OF WATER

BRAIN AND SPINAL CORD

ALBINO RAT

ACCORDING TO BODY WEIGHT

MALES ONLY.



SPINAL CORD


10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 320 330 lb BODY WEIGHT Gms.


Theoretic curves showing the changes in the percentage of water in the brain and in the spinal cord of the male albino rat at different body weights. The dots • show

the observed mean values.


PERCENTAGE OF WATER 127

in the brain and cord of rats according to age and to the normal body weight, we pass to the question of the extent to which the percentage of water may be modified experimentally under special conditions. The amount of modification which has been experimentally induced is thus far extremely slight, nevertheless some deviation seems to occur. The evidence is as follows:

(a) Some conditions which increase the percentage of water in the brain and cord. Dr. Watson ('05) when working on the effects of the bearing of young on the weight of the central nervous system in the albino rat, noted that the mated animals had both heavier brains and heavier cords. He noted also that the mated rats, as compared with the unmated of like age, had the following percentages of water:

NO. OF PERCENTAGE OP WATER

cases Brain Cord

Female mated (8) 77 . 47 68 . 51

Female unmated (10) 77 . 37 68 . 29

This shows that the mated rats had in the brain 0.10 per cent more water than the unmated, and in the spinal cord 0.22 per cent more. Thus, even though the brain and cord in the mated series were absolutely heavier, yet if the higher percentage of water be taken as an index of a lesser maturity, the central nervous system of the mated rats must be regarded as physiologically younger. Such slight differences would, of course, not be worthy of remark if they had been obtained merely by the averaging of widely varying data, but in this case comparisons were made by Watson in five different groups for the brain, and five for the cord, and in only one (in the cord) out of the ten comparisons, did the mated rats show a smaller percentage of water. Thus though the difference is small, it was found to occur in the same sense in nine cases out of the ten. This seems to justify the conclusion which Watson drew that mated female rats had a slightly higher percentage of water in the brain and spinal cord than the unmated females belonging to the same litters.

Hatai ('07) also has made observations on the modifications of body growth as the result of which the percentage of water in the central nervous system was slightly increased.


128 HENRY H. DONALDSON

When young rats were underfed for three weeks and then returned to a normal diet, Hatai found that their subsequent increase in body weight was somewhat more rapid than that of the control group, and in the case of the males, the final weight even greater. Hatai's table IV ('07) is here repeated.


TABLE 4.


ENCEPHAI.ON SPINAL CORD PER CENT PER CEN1


Male controls 77.50 69.71

Male experimented 77 . 75 70 . 05

Female controls 77.50 69.40

Female experimented 77 . 75 70 . 10

Taking both sexes together, the experimented groups, as shown in the above table 4, had on the average a percentage of water in the brain greater by 0.25 per cent and in the cord by 0.52 per cent. As will be observed, this treatment produced a rather greater alteration in the percentage of water than was obtained by Watson in the case of the mated and unmated females.

In the foregoing instance there were fourteen pairs of brains between which comparisons were made, and in thirteen of these the experimented rats show a greater percentage of water. In the case of the spinal cord, eleven pairs out of a total of fourteen show the experimented rats to have the greater percentage of water, so that here again although the variation induced by the treatment is not great, yet a slight change seems to be really effected.

In another series of observations Hatai ('08) got still more marked differences in the percentage of water. In this case there were seven pairs of contrasted individuals. Seven individuals were used as controls and seven others, from the same litter, fed with small quantities of a varied diet and thus stunted. When these latter had attained an average age of about 140 days, they were put on a full normal diet for thirty days and then both lots were killed and examined at the same time.

During the thirty days of normal feeding, the stunted rats grew in weight and length. When killed at this time it was found that the stunted rats had in both brain and cord a distinctly


PERCENTAGE OF WATER 129

greater percentage of water than did the controls. The difference is in the same sense in all the pairs and for both the brain and spinal cord. The average figures are as follows. — ■

PERCENTAGE OF WATER

average ag"e no. of cases Rrain Cord

lTOdays 7stunted 78.618 72.613

170daj's 7controls 78.378 71.076

As the figures show, the percentage of water in the stunted group is greater by 0.24 per cent in the brain and 1.53 per cent in the cord.

The difference in the case of the brain is about that found in the preceding investigation, but that in the cord is much greater. The reason for the greater percentage of water in the case of the spinal cord requires still to be investigated.

The foregoing conditions are the only ones which at the moment have been shown to increase the percentage of water in the central nervous system of the rat, and in all cases this increase seems to be associated with more vigorous growth processes.

(b) Some conditions which decrease the percentage of water in the brain and spinal cord. On the other hand, in 1904 Hatai determined that in rats killed at the end of three weeks of underfeeding the experimented rats, though initially heavier, had on the average only about 44 per cent of the body weight of the controls.

This result was due not only to an arrest of growth, but to an actual loss, as measured by changes in the weight of the entire body and also of the brain. At the termination of the experiment, the brain weight in the underfed group was about 11 per cent less than in the controls, approximately two thirds of this deficiency being due to the arrest of growth, and one third (4.3 per cent) to actual loss (see table IV, Hatai '04). On the other hand, the percentage of water in the brain was

79.11 per cent in the controls 78.91 per cent in the experimented,

thus showing a deficiency of 0.2 per cent in the latter. If the process of the reduction of the percentage of water had been


130 HENRY H. DONALDSON

stopped by the underfeeding, which stops the growth as indicated by the body weight and the brain weight, we should have found the higher percentage in the experimented rats.

As further evidence that the disturbance of the growth process involves but slightly the changes in the percentage of water correlated with increasing age, we have the data in this same paper by Hatai given in Table IV, series II where the control group was killed and examined at the beginning of the experiment. Here the percentages are

79.01 for the control rats 78.71 for the experimented rats

giving a difference of 0.3 per cent.

The difference in this case is greater than in the preceding because not only is the percentage of water in the experimented group slightly diminished by the treatment, but also because the experimented group was three weeks older than the controls at the time of killing, thus giving a total loss of 0.3 per cent in series II against 0.2 per cent in series I, where both controls and experimented rats were killed at the same time. This again supports the view that underfeeding does not arrest the changes in the percentage of water characteristic for advancing age, but may rather hasten them.

The weight of water in the brain and spinal cord. The precedingdescriptions have been given in the terms of the percentage of water. A better view of the changes taking place can be obtained however by following the suggestion of my colleague, Dr. Hatai, and showing the changes in the absolute weight of the water in the brain and cord at different weights of these parts. This eliminates the time factor which has modified the previous forms of presentation, and gives a simple and suggestive picture of the changing relations between the water and the solids.

The determinations thus made are given in table 5 and have been also plotted in charts 4 and 5

The following table 5 shows that for the successive increments of weight, the female brain has less water than the male brain of like weight. This is undoubtedly due to the fact that under


PERCENTAGE OF WATER


131


the conditions of comparison, the female brains are the older. Owing however to the relatively large interval of brain weight used in the correlation tables, from which the means in table 5 were obtained, the absolute weights for the amounts of water increase irregularly, and this in turn makes the progressive per


TABLE 5


The weight of water present in the brain and in the spinal cord according to the absolute weight of these organs. Sexes distinguished. Based on the entire series of records for both sexes. Mean values determined from correlation tables.



Amount of Water


SP. CD. WEIGHT


Amount <


5F Water


BRAIN WEIGHT


brain


SPINAI


cord



Male


Female


Male


Female


grams


grams


grams


grams


grams


grams


0.25


0.208*


0.175*


0.03


0.028


0.027


0.35


0.325


0.290


0.07


0.067


0.062


0.45


0.350


0.400


0.11


0.085


0.083


0.55


0.510


0.450


0.15


0.116


0.110


0.65


0.600



0.19


0.147


0.146


0.75


0.650


0.650


0.23


0.176


0.183


0.85


0.736


0.700


0.27


0.199


0.191


0.95


0.817


0.800


0.31


0.230


0.226


1.05


0.860


0.850


0.35


0.250


0.248


1.15


0.950


0.938


0.39


0.280


0.275


1.25


1.025


1.012


0.43


0.308


0.308


1.35


1.088


1.067


0.47


0.340


0.318


1.45


1.150


1.143


0.51


0.354


0.358


1.55


1.234


1 . 232


0.55


0.390


0.390


1.65


1.304


1.294


0.59


0.412


0.398


1.75


1.359


1.355


0.63


0.434


0.430


1.85


1.450


1.444


0.67


0.465


0.450


1.95


1.530


1.520


0.71


0.473


0.470


2.05


1.636


1.550


0.75


0.520



2.15


1.650






Not plotte

i.






centages still more irregular. However, a second series of calculations based on the theoretic curve for the percentage of water (see table 3 and chart 3) agree so well with the observed results here given that the general correctness of the latter may be accepted.


132 HENRY H. DONALDSON

The significance of table 5 is made more evident by plotting the data on a base line giving either the weight of the brain or of the spinal cord. It is then seen that the records for the weight of water lie in an approximately straight line.

Weight of water in the brain. Beginning with the brain, chart 4, it is seen that when the lines representing the actual weight of water are contrasted with the dotted line, showing the amount of water necessary to maintain the percentage constant at its initial value, the former ascend less rapidly. Further inspection shows that the lines representing the increments of water as observed are slightly convex. This is true for both sexes. We will first consider in detail the relations as thus shown for the males.

A straight line drawn between the terminals for the male curve corresponds to an average of 73.6 per cent of water in the increments of weight after a brain weight of 0.35 grams. Since, however, the curve is slightly convex, it is better represented by two straight lines, one drawn from the initial entry to the entry above the brain weight of 1.05 grams, and the second from this latter to the final entry at 2.15 grains. The angle of the former line corresponds to 76.4 per cent, of water and that of the latter to 71.8 per cent.

From this it appears that the earlier increments of brain weight have a somewhat greater percentage of water than those acquired later.

It is to be noted however that the earlier period comes to an end when the animal weighs only 17 grams, and is about 15 days old (see chart 4) although by this age the very rapid growth of the brain in weight has been completed. (See Donaldson '08, plate III, chart 3.)

With slight differences, which are not significant, the relations here described for the males hold for the females also, but it is hardly necessary to give the determinations in detail.

Such are the general relations of the increase in the weight of water with the increase in brain weight. By these relations several facts are shown.

First. The proportion of water in the brain diminishes with


PERCENTAGE OF WATER


increasing brain weight; a fact already demonstrated previous tables and charts.


133

by the


WEIGHT OF WATER 17 r GMS.


CONSTANT PERCENTAGE BRAIN


WEIGHT OF WATER

BRAIN

ALBINO RAT

ACCORDING TO BRAIN WEIGHT


FEMALE


BRAIN WEIGHT GMS.


BODY WEIGHT GMS. AGE IN DAYS


1.7 1.8

I ) 200

i MALE


1.9


Chart 4


To show the absolute increase in the weight of water corresponding to the increase in brain weight. The data for the two sexes plotted separately. The first entry is for a brain weight of 0.35 grams. Below are given the body weights and the ages in days for the several brain weights. The dotted line indicates the amount of water which would be required to maintain the percentage at the initial value.

Second. The increments of brain weight are characterized by a continuous though small diminution in the percentage of water in the successive increments, the more rapid diminution occurring after the first fifteen days of life.

Third. After the rat has attained about fifteen days of age, the percentage of water in the increment of weight becomes


134


HENRY H. DONALDSON


approximately constant for the remainder of the life cycle, having an average value of 71.8 per cent. This value forms a limit towards which the percentage of water in the entire brain slowly falls.


WEIGHT OF WATER GMS.


WEIGHT OF WATER

SPINAL CORD

ALBINO RAT

ACCORDING TO

CORD WEIGHT


CONSTANT PERCENTAGE SPINAL CORD




i


SPINAL CORD WEIGHT i i i i i


.4



.5


.6


.7 Gms.


1 1 70 105



150


1 225


1 BODY 305 WEIGHT


Chart


5





To show the absolute increase in the weight of water corresponding to the increase in the weight of the spinal cord. The data for the two sexes are plotted separately. Below are given the body weights for the corresponding spinal cord weights. The dotted line indicates the amount of water which would be required to maintain the percentage at the initial value.


Fourth. During the period of most active medullation, i. e., from 20-100 days, the percentage of water in the increments of brain weight does not indicate that the medullary sheaths which are being rapidly formed, possess a percentage of water less than that of the axones on which they appear.

It follows from the foregoing that as the amount of water in the brain at any time after birth is the sum of the amount present at birth (a constant) plus the successive increments, the percentage of water will diminish most rapidly at first. As the brain becomes heavier, and the increments form a greater proportion of the total


PERCENTAGE OF WATER 135

weight, the rate at which the percentage of water diminishes will become slower and slower. At first glance it may be difficult to harmonize these data on the absolute weight of water with the rapid fall in the percentage of water as it appears in charts 1 and 2 based on body weight and on age. If, however, the precocious growth of the brain and spinal cord is recalled, a reference to chart 4 in which the body weight and the ages are given below the brain weights, will serve to make the matter clear.

Weight of water in the spinal cord. The foregoing relations as described for the brain hold true for the spinal cord of both sexes as well, with the difference that in the cord the percentage of water in the total increment from the first to the final entry is less than in the brain, being 68.3 per cent. The percentage of water in the increment during the first 15 days of life is on the average 70.4 and after that 67.9. The record in the case of the cord therefore is more nearly represented by a single straight line than in the case of the brain, but like conclusions can be drawn from the study of the data on the spinal cord as here presented.

Explanation of the change. It still remains to attempt an explanation of the course followed by the percentage of water through the life cycle, and ako to explain why even at birth the brain has more water than the cord, as well as why it shows a smaller fall in this percentage during the life cycle.

In the interests of such a general explanation, let us consider first the condition of the brain and the cord at birth.

In the albino rat at birth, both brain and spinal cord are unmedullated, and both are very watery. Both are composed of gray matter in the strict sense, and growing axones, also gray in color. The nerve elements are enmeshed in supporting tissues and vessels.

In the brain the probability is that the supporting tissues, as well as the vessels, form a slightly smaller fraction of the total mass than in the cord. Cell division in the brain is continued longer after birth than in the cord, while medullation in the brain begins later than in the cord, and is less rapid. During subsequent growth, medullation is most actively carried on from the age of about twenty to one hundred days.

Between birth and maturity the proportional increase in the


136 HENRY H. DONALDSON

weight of the brain is only about two fifths that of the spinal cord (Donaldson '08, p. 355 and p. 358) and at maturity the relative amount of white matter in the brain is much less than in the spinal cord (Donaldson and Davis '03; Watson '03). Such are the characteristics of these two portions of the central nervous system which are of interest to us in connection with the percentage of water.

Explanation of the greater percentage of water in the brain at birth. The greater percentage of water in the brain at birth may be connected with some of the facts just enumerated, namely, the lesser maturity of the brain, as indicated by the longer continuance of cell division, by the later onset of medullation, and by the lesser proportion of supporting tissues and other non-nervous constituents. All of these conditions would tend to give the brain a higher percentage of water.

During the subsequent growth, the slower diminution of the percentage of water in the brain is due to the fact that the relative increment of water is greater than in the case of the cord (see charts 4 and 5). This however is again no explanation and leaves the conditions which control the increment of water in each division of the system still to be described.

As can be seen from inspecting chart 4 it is possible to express the events taking place by assuming that the initial weight of material in the brain maintains its initial percentage of water and that each of the subsequent increments in weight from just after birth to old age has a relatively low and slightly diminishing percentage of water. Such a statement however is purely formal.

What probably takes place is this: Starting with a given percentage of water in the brain or cord, this percentage continually diminishes as the formed material becomes older— at the moment of formation, however, the young material subsequently added most probably has a relatively high percentage of water, and the percentage we obtain at any given age is therefore the mean of these several values. The rate at which the percentage is falling off at any moment, together with the general slowing of the growth process — requiring a longer and longer time to add the same increment of weight to the brain the older the brain be


PERCENTAGE OF WATER 137

comes— is so adjusted as to yield after the period of more rapid growth of the brain, the rather simple relations of a nearly constant weight of water for the same increment of total weight.

In this connection the analysis of the brain and cord should however be carried one step further. Both are composed of gray matter (substantia grisea) and axones, plus the supporting elements, the axones being more or less medullated according to locality and age. In the case of the rat, it has not been possible to study the changes in the percentage of water in the gray matter alone. We know however from a number of studies on man — on the cortical gray and the gray of the corpus striatum — that the change in the percentage of water in the substantia grisea with age, is much less — less than one-half — that in the axones (white matter). This has a bearing on the percentage of water in the brain as contrasted with the cord, because the brain has relatively less axone substance in it. Moreover the maturing of this substance is slower in the brain than in the cord. It is worthy of note as bearing on this last point that according to Watson ('03, p. 91 and 105) medullated fibers in the spinal cord of the rat are first found on the second day after birth, while in the cerebrum, they are not found until the eleventh day. At that age — eleven days — the percentage of water in the brain has fallen to that of the cord at the second day, and it thus appears that the medullation of axones begins in both divisions of the central nervous system when these have acquired the same percentage of water.

This suggestion, that the onset of medullation is closely related to the percentage of water in the axones, fits with the common observation that the fibers first medullated in any locality become the largest (because they have the longest time to grow aftei reaching the condition in which they can become medullated) and that in any nerve containing medullated and non-medullated fibers, it is the smaller (or younger) fibers which lack the sheath (Roughton '06). Also, as the portion of the axone nearest the cell body is the older, and hence would have the lower percentage of water, this should be the portion first medullated; a conclusion which fits with the observations.

It is hardly necessary to remark that these last two facts when


138 HENRY H. DONALDSON

interpreted in this way, constitute indirect evidence for the view that the axone is an outgrowth from the cell body.

The medullation process as such does not reduce the 'percentage of water. This statement, already made in the " fourth" conclusion on p. 19 is here repeated because there is a more or less widely diffused opinion that the medullary sheath is a structure containing less water than the axones, and that it is the addition of the myeline, as it appears in the medullary sheaths, which largely serves to reduce the percentage of water in the white substance, and thus in the entire mass of the central system. For this there is no evidence. Charts 4 and 5, exhibiting the increase in the weight of water with increasing brain weight and cord weight, show no changes in the increment of water which would warrant such an explanation. It appears most probable therefore that the medullary sheaths when first formed have approximately the percentage of water characteristic for the axones at the time of their myeli nation, and after that, in company with the axones, they undergo a slow but steady diminution in water content.

A few separate determinations of the percentage of water have been made by various investigators on the white and gray substances of man and other larger mammals. These show without exception that between birth and maturity there is a greater loss of water in the white substance than in the gray. In these cases of course the white substance at maturity is always medullated, and thus the results do not answer the question whether the formation of the medullary sheaths has contributed to the diminution in the percentage of water. That the axones previous to medullation do show a reduction in the percentage of water with advancing age, is indirectly indicated by my own observations in the following way:

At birth (i. e., 5 grams body weight) the average percentage of water in the rats' brains of both sexes is 87.8 per cent (table 1). At eleven days of age, as shown in chart 2, it is about 84.8 per cent, a loss of three units, yet it is not until after the eleventh day that medullation in the brain begins. The percentage of water in the brain therefore falls off before medullation begins, and the nerve substance, cell bodies and axones, are the portions in which this diminution chiefly occurs.


PERCENTAGE OF WATER 139

As there is every reason to think from what we know concerning the relatively small reduction in the gray substance that the percentage of water in the cell bodies in this case is not progressing more rapidly than it does in the entire brain, it follows that in the remaining nerve structure — the axones — the percentage of water is reduced at least an equal amount. Since however the diminution in the percentage of water is found to be much greater in the mature white substance than in the mature gray, it seems probable that the axones are subject to a more extensive reduction in the percentage of water than are the cell bodies. 3

From this it follows that the greater proportion of gray substance in the brain would tend to maintain in that organ a higher percentage of water at maturity, and the lesser proportion of gray in the cord, a lower percentage (see the measurements on the areas of the gray and white matter in the spinal cord as given by Watson, '03, p. 101).

But there is still one more peculiarity of the spinal cord which is important in this connection. This I have called (Donaldson '09, pp. 166-167) passive lengthening. The segments of the cord, especially in the thoracic region, undergo during growth a lengthening which is largely passive, and which does not imply any marked increase in the structural complexity of the cord, but serves mainly to keep the spinal nerves nearly opposite to their intervertebral foramina.

In the course of this lengthening, we have evidence that the volume of the gray substance is but slightly increased, while the proportions of the gray column are much modified in the sense that the diameter alters but slightly (it may even diminish) while the length is correspondingly increased (see the measurements on the areas of the gray and white matter in the spinal cord of the rat, Watson '03, p. 101, and Donaldson and Davis '03).

At the same time that this is occurring in the gray columns, the white tracts not ODly lengthen (passively) but also increase in the area of their cross sections, and thus at the end of any step

3 The question whether a growing fiber at any age of the animal becomes medullated as soon as its percentage of water falls below the value at which medullation first begins after birth, cannot at the moment be answered. It is conceivable however that with advancing age this critical point for medullation is lowered.


140 HENRY H. DONALDSON

in this transformation, we find a larger proportion of white substance than at the beginning. The white substance having a lower percentage of water than the gray, tends of course to bring down the general average. We know from previous studies that in the albino rat the weight (and length) of the spinal cord increases so long as the animal grows (Donaldson '09).

It is therefore the relative increase in the white substance due to this continuous passive lengthening — which is so marked in the cord — that can justifiably be held responsible for the more rapid decrease in the percentage of water in the cord after maturity.

In brief then, the more rapid diminution in the percentage of water in the cord up to maturity, and the greater rate of diminution after maturity, are due, aside from the excess of supporting tissues and vessels, to the greater amount of axone substance in the cord and the peculiar form of growth designated as passive lengthening.

General significance of this change. If we are correct in concluding that in the percentage of water we have a character correlated very closely with the age of the animal, and but slightly influenced by the conditions which modify general growth, it follows that this change must depend on processes intimately associated with the span of life or longevity of the animal concerned.

Broadly speaking, the changes in the percentage of water indicate progressive chemical modifications which take place in those constitutents of the cell that are most stable.

A comparison of the albino rat with man in respect to the percentage of water in the brain. In connection with such a comparison, I have examined the entire literature on the percentage of water in the nervous system. This literature needs to" be summarized, but for such a summary, this is not the occasion. Out of the data available, I have selected however the findings of Weisbach ('68) and of Koch ('09) to be used in the present instance, as from these we get the best series of determinations of the water in the human brain at different ages. The data from Weisbach are as follows:

He determined the percentage of water in each case for six different localities in the encephalon: (1) white substance (callosum); (2) gray substance (corpus striatum); (3) gyrus (white


PERCENTAGE OF WATER 141

and gray mixed); (4) cerebellum; (5) pons; (6) medulla oblongata.

For the percentage of water in the human encephalon at birth, the determinations for these several localities are averages from three male and five female newborn infants.

A series of tes+s applied by me to Weisbach's data for mature brains have shown that the percentage of water in the entire encephalon is approximately equal to the sum of four times that in the white substance (1) ; five times that in the gray substance (striatum) (2); and once that in the cerebellum (3) divided by 10.

This procedure gives for the percentage of water in the human brain at birth 88.34 per cent. The value thus obtained is probably nearly correct.

By the same procedure I obtained from Weisbach's data for children between three and fourteen years of age (2 males: 3 years and 8 years; 2 females: 4 years and 14 years) a mean value for the entire encephalon of 79 . 2 per cent at 9.5 years. Finally, Weisbach's records for 64 males and 17 females, 20-30 years of age, give a mean value of 77.0 per cent.

Turning now to the determinations of Koch ('09) we find his average determination of the percentage of water in five human encephala at maturity to be 77.8 per cent. In a female brain of two years, he gives the percentage of water in the cortex of hemispheres as 84.49 per cent and in the callosum as 76.45 per cent.

In a male of 19 years, the cortex was found to have 83.17 per cent and the callosum 69.67 per cent. This last case may be taken to represent the conditions at maturity. This being assumed, it was found that combining the determination for the cortex in the proportion of 603 times to 397 times of that for the white substance gave a mean percentage of 77.8, which is Koch's determination for the water in the entire encephalon at maturity. 4

Using the same proportions as those just given for the gray and white substances at maturity and applying them to the data for the brain at two years, mentioned above, we obtain as a mean

4 The proportional abundance of the gray and white substance in the encephalon is not to be inferred from the numbers given above. Each investigator has used more or less arbitrary criteria for the gray substance, and a treatment of the results of an author in the manner here followed, has a value for the determinations by that author alone.


142


HENRY H. DONALDSON


value for the percentage of water in the encephalon at two year? 81.1 per cent. Thus we are able to obtain approximate values for the percentage of water in the human encephalon at birth, two years, nine and one-half years and at maturity, twenty-five years.

TABLE 6.

Comparison of the percentage of water in the encephalon of man and the albino rat at

corresponding ages


MAN


RAT


Age Percentage of Years Water


Percentage of Age Water Days


Birth 2 years 9 . 5 years 25 years maturity


88.3 81.1 79.2

77.0


87.7 81.3 78.6

78.0


Birth

26 days

115 days

290 days


In order to compare these determinations, it is necessary to recall that the span of life in man is about thirty times as long as that in the rat, and if this relation holds throughout the life cycle, it follows that each determination for man is to be compared with that for the rat having one thirtieth the human age taken.

The data for the rat are based on the entries in table 2 giving the percentage of water according to age.

This table shows that when we compare brains of corresponding ages, the diminution in the percentage of water in the two forms has similar limits, and would be expressed by a curve of like form in both instances.

When we examine the records for other mammals, we find almost no determinations for the water in the encephalon at birth, but we do find determinations for this character at maturity, and the values are very similar to those for man and the rat. Remembering that the relative amount of white matter in the encephalon varies somewhat in different species, and must therefore modify this result, we reach the interesting conclusion that probably in all mammals we shall find approximately the same range in the percentage of water between birth and maturity, and that the loss of water in them occurs in the same manner but that the


PERCENTAGE OF WATER 143

time required for the successive steps is determined by the intensity of the growth process characteristic for each species. (Rubner '08 and '08a).

CONCLUSIONS

1. In the albino rat between birth and maturity, the percentage of water in the brain diminishes from 87.8 to 77.5 and in the spinal cord from 85.6 to 68.0. Table 1.

2. The progressive diminution of the percentage of water is a function of age and is not significantly modified by any conditions to which the animals have been thus far experimentally subjected.

3. The diminution in the percentage of water is most rapid during the first twenty-five days of life; the period at which the central nervous system is growing most actively.

4. The maturing of the axone substance is characterized by a greater diminution in the percentage of water than is the maturing of the gray substance.

5. Medullation begins when the percentage of water in the brain and cord has diminished to about 85.3 per cent (second day in the spinal cord; eleventh day, in the brain).

6. The process of medullation itself as indicated by the formation of the medullary sheaths, is not a controlling factor in reducing the percentage of water in the central nervous system.

7. The range and course of the diminution of the percentage of water in the brain are similar in man and in the albino rat. The rapidity of change agrees with the intensity of the growth processes in each of the two species, and is therefore about thirty times more rapid in the rat than in man. This point has not been tested for the spinal cord.

8. It is probable that the same limits in the percentage of water and the same course of diminution will be found to occur in other mammals.

9. The progressive diminution of the percentage of water in the central nervous system with advancing age, is to be regarded as an index of fundamental chemical processes, which take place in the more stable constituents the nerve cells.

These processes are but little modified by changes in the environ


144 HENRY H. DONALDSON

merit and taken all together constitute a series of reactions which express not only the intensity of the growth process in the nervous system, but also the span of life characteristic for any given species.

BIBLIOGRAPHY

Boughton T. H. The increase in the number and size of the medullated fibers 190G in the oculomotor nerve of the white rat and of the cat at different

ages. Journ. Comp. Neurol., vol. 16, no. 2, pp. 153-165. Donaldson, H. H. A. comparison of the white rat with man in respect to the

1906 growth of the entire body. Boas Memorial Volume, pp. 5-26. Donaldson, H. H. A comparison of the albino rat with man in respect to the

1908 growth of the brain and of the spinal cord. Journ. Comp. Neurol., vol. 18, no. 4, pp. 345-390.

Donaldson, H. H. On the relation of the body length to the body weight and to

1909 the weight of the brain and of the spinal cord in the albino rat (Mus norvegicus var. albus). Journ. Comp. Neurol., vol. 19, no. 2, pp. 155-167.

Donaldson, H. H. and Davis, D. J. A description of charts showing the areas

1903 of the cross sections of the human spinal cord at the level of each spinal nerve. Journ. Comp. Neurol., vol. 13, no. 1, pp. 19-40.

Hatai, S. The effect of partial starvation on the brain of the white rat. Am.

1904 J. of Physiol., vol. 12, no. 1, pp. 116-127.

HAtai, S. Effect of partial starvation followed by a return to normal diet, on

1907 the growth of the body and central nervous system of albino rats. Am. J. of Physiol., vol. 18, no. 3, pp. 309-320.

Hatai, S. Preliminary note on the size and condition of the central nervous

1908 system in albino rats experimentally stunted. J. of Comp. Neurol., vol. 18, no. 2, pp. 151-155.

Hatai, S. Note on the formulas used for calculating the weight of the brain

1909 in the albino rats. Journ. Comp. Neurol., vol. 19, no. 2, pp. 169173.

Koch, W. and Mann, S. A. A chemical study of the brain in healthy and diseased

1909 conditions, with especial reference to dementia prsecox. Mott's

Archives of Neurol, and Psychiatry, vol. 4, pp. 201-204. Rubner, Max. Probleme des Wachstums und der Lebensdauer. Wiener

1908 Medizinischen Wochenschrifl, nos. 11-13.

Rubner, Max. Das Wachstumsproblem und die Lebensdauer des Menschen

1908a und einiger Siiugetiere vom energetischen Standpunkt aus be trachtet. Archiv fur Hygiene, vol. 66, pp. 127-208. Watson, J. B. Animal education, 8° University of Chicago Press. Chicago.

1903 Watson, J. B. The effect of the bearing of young upon the body weight and the

1905 weight of the central nervous system of the female white rat. Journ. Comp. Neurol., vol. 15, no. 6, pp. 514-524.

Wlisbach A. Der Wassergehalt des Gehirns nach Alter, Geschlecht und Krank1868 heiten. Med. J alirbilcher , vol. 16, nos. 4 and 5, pp. 1-76.


PLEASURE, PAIN AND THE BEGINNINGS OF INTELLIGENCE

S. J. HOLMES

University of Wisconsin

The tendency of animals to repeat acts which result in pleasure and to discontinue or inhibit acts which bring them pain is a fundamental feature of behavior on the utility of which it would be superfluous to comment. But why do animals behave in this fortunate manner, and how did they come to acquire the faculty of so behaving? To our ordinary plain way of thinking it appears sufficient to say that a dog eats meat because he likes it, and that he runs away from the whip to avoid its painful incidence upon his integument. These acts are such natural and obvious things to do under the circumstances that to inquire why the animal does what it likes and avoids what is disagreeable may seem a sort of philosophic quibble which only a mind " debauched by learning" would think of indulging in. But a little consideration will show that we have here a real and very knotty problem, or rather set of problems, of the greatest importance to the student of genetic psychology.

There are few better illustrations of the modification of behavior through experiences of pleasure and pain than that afforded by the behavior of young chicks, which has been so well studied by Lloyd Morgan. A young chick when first hatched has the instinct to peck at all sorts of objects of about a certain size. If an object is a little too large the chick may hesitate. Should it venture to peck at the object and derive a pleasant taste from it the hesitation in the presence of similar objects becomes reduced and will finally disappear. If the chick in the course of its pecking seizes a caterpillar having a nauseous taste it is much less apt to seize a similar caterpillar a second time. The painful or un


146 S. J. HOLMES

pleasant experience it derives in some way inhibits further action towards that class of objects.

We have in this modification of instincts through the pleasurable or painful effects they produce the beginning of intelligence. The pecking, swallowing, and avoidance of certain objects are purely instinctive acts based on the chick's inherited organization. After its first experiences with pleasant or nasty caterpillars the chick is a different creature; it has learned by experience; and henceforth its acts, which at first were in a general way adaptive, become more perfectly adapted to its needs as the result of its learning. Instinct supplied the impetus to action and in a measure determined the direction of action, but intelligence refines upon the instinctive behavior and effects a closer adjustment to the environment.

In lower forms associations are formed as a rule with great slowness. Behavior is almost entirely instinctive, and the organism can be made to deviate from its stereotyped methods of action only with difficulty. It is probable that in low forms where associations of only the simplest kind can be established there is no association of ideas involved; and in fact there is no conclusive evidence of the existence of ideas even in animals quite high in the scale. Most animal learning consists in forming associations between certain sense experiences and certain actions which bring pleasure or pain. A common way of teaching an animal a trick is to try in various ways to induce it to perform the desired action and then to reward it by food or some other means of giving it pleasure. In this way the connection between the situation and the act is reinforced and the act follows more readily when the animal is placed a second time under the same conditions.

Consider the case of a cat placed in a box which can be opened by pressing down a lever or pulling a string, as in the experiments of Thorndike. If the cat is hungry and food is placed outside, the animal will probably make a vigorous effort to escape by clawing and biting in various parts of the enclosure, which are the usual instinctive methods employed in similar situations.. If the right movement is hit upon and the cat gets out and secures food, it will probably make its escape more readily than before


PLEASURE, PAIN AND INTELLIGENCE 147

when placed in the box a second time. After a number of trials the cat will come to make the right movements for escaping very soon after being placed in the box and its various useless random movements will be discontinued. The connection between the perception of the mechanism of escape in the box and the act necessary to gain its liberty comes to be more and more firmly established in the cat's brain with repeated experiences. The cat perceives a number of things in the box and performs a number of different acts but out of all these, associations are formed only between certain stimuli and those responses to them which bring pleasure to the animal.

Pleasure and pain therefore have apparently a fundamental connection with the development of intelligent responses out of instinctive activity. Were there not something to clinch or strengthen the connection between certain stimuli and the appropriate responses to them the organism might perform random movements till doomsday without being a whit better off. It is a problem therefore of fundamental importance to ascertain in what the mechanism of this ability to profit by experience essentially consists. It is not mere habit, not the mere making more permeable certain preformed connections in the brain. One act would then be just as apt to be followed up as another. Whether an act tends to be followed or not depends on what it brings to the organism. Apparently we have to do with a selective agency which preserves or repeats certain activities and rejects others on the basis of their results.

The importance of random movements lies in the fact that they offer opportunities for making favorable adjustments. For the development of intelligence they play a similar role to that of variations in the process of evolution. The animal that does the most exploration is the one most likely to hit upon new advantageous adjustments. In the same way intelligent adjustments as James has contended are favored by a multiplicity of instincts, especially if these instincts are of a contrary or conflicting nature, for now one and now another instinctive tendency may be reinforced in different conditions to which each may be adapted. Instinctive fear may be modified through experience so that it is


148 S. J. HOLMES

no longer attached to objects that are found to be harmless, while it may be intensified in relation to other objects that are found to be sources of injury. Where there is hesitation between the exercise of two instincts such as the tendency to pursue an animal as prey, and the instinctive fear which that animal may awaken, experience may quickly point out which proclivity is the more advantageous to follow. The pleasure-pain reaction enables an animal to select, so to speak, out of its stock of instinctive endowments those responses which are best adapted to the particular situations that confront it. It is a means of adapting instincts to new or inconstant conditions and thus of effecting a closer adaptation to the environment than that which would be possible by following purely congenital modes of response. The development of the pleasure-pain reaction marks one of the most important steps in the evolution of behavior, for the entire superstructure of intelligence in all its stages is based upon it, and it is not surprising that many writers regard it as an index of the beginning of consciousness, a point where a new entity is somehow mysteriously injected into the universe.

It is a general rule that what is pleasant is beneficial and what is painful is injurious; and, therefore, by following its desires and aversions an animal is guided in a tolerably safe course. Eating when hungry, drinking when thirsty, seeking warmth when cold, exercise when in a state of vigor, and rest when fatigued, all bring a state of satisfaction or pleasure. On the other hand, eating and drinking after a certain stage of repletion has been reached, or attaining too great a degree of warmth may be positively painful, the pain being correlated with carrying on these activities until they become injurious to the organism.

But it is well known that this correlation of the pleasant with the beneficial is not an absolute one. With complex creatures like ourselves with a multitude of different propensities and interests it is not infrequent that the pursuit of what is agreeable leads to all sorts of unfortunate consequences even of a purely physiological nature. In the lower animals where pleasure is a safer guide than among ourselves, what is pleasant is not always what is organically good. Poisonous articles may be eaten with appar


PLEASURE, PAIN AND INTELLIGENCE 149

ent relish and alcoholic liquors are readily imbibed even by such primitive creatures as bees and wasps upon their very first acquaintance with these intoxicants. But aside from exceptional cases pleasure in the animal world is a sufficiently good index of what is beneficial that under conditions which ordinarily present themselves it seldom leads to injurious courses of action.

The relation between the pleasant and the beneficial is, however, probably not a primary one, and it is not improbable that it represents a connection established by natural selection, as was first maintained by Herbert Spencer.

If the states of consciousness which a creature endeavors to maintain are the correlatives of injurious actions, and if the states of consciousness which it endeavors to expel are the correlatives of beneficial actions, it must quickly disappear through persistence in the injurious and avoidance of the beneficial. In other words, those races of beings only can have survived in which, on the average, agreeable or desired feelings went along with activities conducive to the maintenance of life, while disagreeable and habitually-a voided feelings went along with activities directly or indirectly destructive of life; and there must ever have been, other things equal, the most useful and long-continued survivals among races in which these adjustments of feelings to actions were the best, tending ever to bring about perfect adjustment.

This explanation which has become widely accepted leaves a fundamental question unanswered. It does not explain why certain acts are stamped in and certain others stamped out. Of the mechanism of this process, which is the real problem involved in the pleasure-pain reaction, we are as ignorant as before. The explanation means that animals which took pleasure in following acts that brought them benefit were preserved and those that did not behave in this manner were eliminated. But why does an animal tend to repeat an act that brings it pleasure and avoid one that produces pain? It seems so natural for creatures to behave in this way that the existence of any problem here is usually unsuspected, but this is the problem that confronts us when we endeavor to obtain a clear understanding of the way in which intelligence develops out of instinct.

In the pleasure-pain response we have two problems of a quite


150 S. J. HOLMES

different nature: (1) the problem of how behavior is modified by its results, and (2) the problem of why pleasure is associated with certain physiological activities such as securing movements and pain with others such as avoiding movements. The latter problem is one whose solution appears holpeless. If we accept the doctrines of psycho-physical parallelism in any of its forms, we must deny that psychical states are, strictly speaking, the causes of physical changes. Why then should pleasure be connected with one kind of activity and pain with another? Why not just the reverse? This problem is, I believe, insoluble, because it is a question of the relation of the physical and the psychical; it is of essentially the same nature as the question why one kind of retinal stimulation produces a sensation of red and another a sensation of green. Physical and psychical states are correlated in particular ways; this we accept as a matter of observed connection. But why a certain kind of brain vibration is associated with a state of consciousness we call a sensation of red instead of some other state is a question upon which we may intend our minds indefinitely without the least profit. If we adopt any other theory of the relation of mind and body we are in no way better off. If we have to do with a preordained connection of pleasure with certain physiological activities and pain with certain others, this connection is no more intelligible if we admit the interaction of psychical and physical states than it is under the theory of parallelism. We can only say that such is the observed relation of the phenomena. It may be regarded therefore as a piece of good luck that we are constituted so as to pursue pleasure and avoid pain. We might have been endowed with a fatal tendency to do just the reverse. Pain instead of pleasure would then have been the correlate of physical well being; those forms in which the painful corresponded with the organically good would have been preserved, the others would have perished; and thus there would have been established a correlation between what is sought and what is conducive to organic welfare, as there is now, but of a quite different kind.

There is another way of looking at the problem which avoids the difficulty we have mentioned ; and that is to suppose that pleasure


PLEASURE, PAIN AND INTELLIGENCE 151

and pain are such only through their motor correlations. We may say, not that pain is mysteriously associated with withdrawing activities, but that pain is the psychic state which accompanies such activities — that it is those activities that constitute it pain. Pleasure would thus become the feeling which is reached out for; pleasure is pleasure just because it is sought. Or to state the matter somewhat paradoxically, if we were so constituted as to seek pain, the pain would not be pain but pleasure. It is indeed difficult to define the difference between pleasure and pain without appealing to the motor accompaniments of these feelings. The following words from Spencer are significant in this connection :

If we substitute for the word Pleasure the equivalent phrase — a feeling which we seek to bring into consciousness and retain there, and if we substitute for the word Pain the equivalent phrase — a feeling which we seek to get out of consciousness and to keep out, etc.,

showing clearly that the qualities of these feelings are defined in terms of their motor accompaniments.

If we sought pain with quickened pulses and enthusiastic efforts, smiled when we obtained it, endeavored to keep it in consciousness as long as possible, and acted in all ways towards it as we now do towards pleasure, would it not then be pleasure? Such a conclusion, I must confess, jars against ones common sense habits of thought. Would it not seem absurd to say that red would be blue and blue red if the motor tendencies, whatever they may be, which these sensations awaken, were reversed? Would not the sensation of pain be the same if our nervous system were so constructed that excitations producing pain led to entirely different activities? How far the qualities of sensations are to be regarded as things standing as it were by themselves, just as atoms are supposed to possess certain qualities independently of their various combinations, may be open to question. It is a question for which fortunately our general discussion does not require an answer and we may leave it therefore to specialists in psychology and metaphysics. What we are concerned with at present is an explanation of the pleasure-pain process. If we


152 S. J. HOLMES

can explain this in physiological terms we can safely leave the preceding question to one side to be answered in whatever way it may.

Turning then to the problem of how behavior comes to be modified in adaptive ways by the pleasureable and painful experiences it brings to the animal, it is evident that it can be treated as a problem of physiology. We are dealing with a series of physiological reactions and how they come to be modified. We may assume that psychical states enter into the chain of causes and effects that make up an animal's behavior, but it is not clear that such an assumption throws the least light upon our problem, and it is open to serious objections on both scientific and metaphysical grounds. We shall therefore consider the question purely from the physiological standpoint. Viewed objectively we find that in an animal's behavior certain acts when once performed tend to be performed with greater readiness under similar conditions a second time, while other acts once performed tend under similar conditions to be inhibited. This problem of learning, Baldwin observes "is the most urgent, difficult and neglected question in the new genetic psychology." Spencer with his characteristic insight into fundamental problems has grappled with it and has attempted to give a physiological explanation. Pleasure, according to Spencer, is the concomitant of heightened nervous discharge; pain the concomitant of lessened discharge. In an animal with a diffuse discharge of its nervous energy resulting in random movements, some of these movements bring a heightened nervous discharge with its psychic accompaniment of pleasure. This tends to reinforce the movement that brought the increase of nervous energy and to cause it to be repeated. Responses resulting in pain tend on account of the diminution of nervous discharge that follows to be discontinued, and in this way the organism is kept repeating certain acts and avoiding others.

"Along with the concentrated discharge to particular muscles," says Spencer, "the ganglionic plexuses inevitably carry off a certain diffused discharge to the muscles at large, and this diffused discharge produces on them very variable results. Suppose, now, that in putting out its head to seize prey scarcely within reach, a creature has repeatedly failed. Sup


PLEASURE, PAIN AND INTELLIGENCE 153

pose that along with the group of motor actions approximately adapted to seize prey at this distance, the diffused discharge is, on some occasion, so distributed throughout the muscular system as to cause a slight forward movement of the body. Success will occur instead of failure ; and after success will immediately come certain pleasurable sensations with an accompanying draught of nervous energy towards the organs employed in eating, etc. That is to say, the lines of nervous communication through which the diffused discharge happened in this case to pass, have opened a new way to certain wide channels of escape; and, consequently, they have suddenly become lines through which a large quantity of molecular motion is drawn, and lines which are so rendered more permeable than before. On recurrence of the circumstances, these muscular movements that were followed by success are likely to be repeated : what was at first an accidental combination of motions will now be a combination having considerable probability."

Bain's view of learning is much like that of Spencer.

We suppose movements spontaneously begun, and accidentally causing pleasure; we then assume that with the pleasure there will be an increase of vital energy, in which increase the fortunate movements will share, and thereby increase the pleasure. Or, on the other hand, we suppose the spontaneous movements to give pain, and assume that, with the pain, there will be a decrease of energy, extending to the movements that cause the evil, and thereby providing a remedy. A few repetitions of the fortuitous concurrence of pleasure and a certain movement, will lead to the forging of an acquired connection, under the law of Retentiveness or Contiguity, so that, at an after time, the pleasure or its idea shall evoke the proper movement at once.

The theories of Bain and Spencer are discussed in detail by Baldwin, who, while differing from these writers in certain points which need not here be dwelt upon, adopts essentially the same view as regards the mechanism of reinforcement and inhibition. With Bain and Spencer, Baldwin assumes that

the pleasure resulting from the first accidentally adaptive movement, issues in a heightened nervous discharge toward the organs which made the movement, a discharge which finds its way to the same channels as before, and so makes it likely that the same movement will be repeated, the external conditions remaining the same Pleasure and


154 S. J. HOLMES

pain can he agents of accommodation and development only if the one pleasure, carry with it the phenomenon of "motor excess," and the other, pain, the reverse — probably some form of inhibition or of antagonistic contraction.

The theories of Spencer, Bain and Baldwin are physiological since they attempt to explain the modifications of behavior, not through the influence of certain states, but as the effect of the physiological conditions of which these states are the concomitants. The theories are all open to the objection that pleasure is by no means the constant concomitant of heightened nervous discharge. Laughing and crying are very similar in their physiological expression though they go along with very different psychic states. A child who burns his hands and writhes about in agony certainly manifests a heightened nervous discharge, but he shows no tendency to put his hands again into the fire. Another outreaching movement of the child brings his hands towards a pleasant degree of warmth. The movement tends to be repeated. The nervous discharge in the first case is much greater than in the second, but in both cases it goes to the arm, though along somewhat different nerves. It is obvious, I think, that we cannot account for the difference between the responses to pleasurable and painful stimuli on the basis of any quantitative difference in the discharges to the part affected. It is a matter of nervous connection rather than quantity of nervous energy.

Pain-giving stimuli, owing to the arrangement of an animal's reflex arcs, are generally followed by a withdrawing movement of the part stimulated, but that there is a tendency for the "increased energy of the pleasure process" to flow "into the channels of the movement associated with the pleasure' ' (that is, I take it, the movement which brings pleasure) is by no means evident. There is, I think, no primary tendency, as Spencer and Bain seem to think, for the nervous discharge to take the direction of the organ from which the pleasure is derived. Animals, it is true, move so as to bring an organ which is pleasantly stimulated again under the action of the stimulus, but this is often due to the discharge going mainly to a quite different part of the body, such as distant appendages, instead of the part directly affected.


PLEASURE, PAIN AND INTELLIGENCE 155

The theory of heightened nervous discharge as expounded by Spencer, Bain and Baldwin, fails to give us, I think, the desired explanation of the acquirement of individual accommodations, and one naturally turns to other theories of the psycho-physiology of pleasure and pain for light. Here, however, we are led into a veritable quagmire of psychological speculation, for there are few fields in which there are so many and so fundamental differences of opinion among competent psychologists.

It is quite generally agreed that there are specific pain sensations aroused by the stimulation of special nerve endings. The existence of specific pleasure sensations is much less widely accepted. Granting the existence of pain sensations, the relation of these to other forms of pain is by no means clear. Stumpf has made an elaborate attempt to show that all pleasurable and painful feelings are really sensations, a conclusion which he shares with a number of other psychologists. Wundt, Kulpe, Ebbinghaus, Titchener, and others, while admitting the existence of specific pain sensations contend that there is a class of feelings distinct from sensations, the affections. But among those who contend for affection as a distinct category of psychological phenomena, there is much lively controversy. Some contend, like Titchener, that " there are only two kinds or qualities of affection, pleasantness and unpleasantness." Wundt in his well known tridimensional theory of feeling, which has secured a small following, attempts to prove that feelings may differ in at least three pairs of contrasted attributes of which pleasantness and unpleasantness form one. Royce in his Psychology reduces these to two. The field of enquiry is one of peculiar difficulty and experiments in the hands of different investigators have yielded contradictory results. Brahn, for instance, finds pleasure and pain accompanied uniformly by certain variations of the pulse and reaches results confirmatory of Wundt's tridimensional theory. Titchener, Hayes and others, on the contrary, have reached results which the\ regard as clearly at variance with Wundt's doctrines.

There is no agreement among psychologists as regards the physiological expression of the pleasant and unpleasant. Fere, Leh


156 S. J. HOLMES

mann, Mentz, Zoneff and Meumann find that pleasant and disagreeable states are quite uniformly accompanied by certain characteristic physiological processes. Among the accompaniments of pleasurable feeling we have increased amplitude of heart beat, a slowing of the pulse, dilation of peripheral blood vessels and an increase in the rate of breathing, accompanied by a decrease in its depth. Unpleasantness on the other hand is said to go along with quickening of the pulse, contraction of the blood vessels, and slower and deeper respiration. Other investigators, however, fail to obtain such uniform results. Kelchner finds that agreeable tastes have an opposite effect on the pulse from that produced by sounds and colors and that the respiratory changes corresponding to agreeable and disagreeable stimuli are far from constant. Shepard in studying the effect of stimuli upon the peripheral circulation finds that 19 agreeable stimuli gave a fall of volume distinctly, while 4 gave a possible rise; 15 disagreeable stimuli gave a distinct fall, and 2 a possible rise. Agreeable smells were found to deepen the respiration and disagreeable ones to have the opposite effect.

A disagreeably exciting sound or a noise tends to deepen breathing and often makes it irregular also. Agreeably exciting stimuli at least as often increase as decrease the depth.

The volume of the brain (studied upon a person who had lost a portion of the skull in an accident) showed no constant relation to agreeable or disagreeable stimuli, both producing in general an increased volume and increased cerebral pulse.

Shields has reached similar conclusions in studying the effect of odors upon the circulation. Heliotrope and wood violet were enjoyed by the subject experimented on, but the volume of the arm diminished quite as often as increased during their application. Indol and skatol are unpleasant odors, but the volume of the arm frequently increased during the first few seconds of their application and then decreased. The effects of odors were very different with different people, and with the same person at different times, and the author concludes that


PLEASURE, PAIN AND INTELLIGENCE 157

the experiments give no support to the view that pleasant sensations are accompanied by a diminution of blood suppl} r to the brain, and unpleasant sensations by the reverse effect.

Angell and Thompson express themselves as 1 allows concerning the physiological concomitants of pleasant and unpleasant reactions :

It is in the case of the emotions, where the agreeable and disagreeable experiences are most intense, that we should expect to find the most marked and constant correspondence of agreeable states with one set of physiological processes and of disagreeable states with an antithetical set . if any such relationship existed. But our curves show not the slightest evidence of such an interconnection. None of the various factors involved, vaso-motor level, rate and amplitude of the pulse curve, position and emphasis of the dicrotic notch, or rate and amplitude of the breathing, changes uniformly in one direction for agreeable experiences,

and in opposite direction for disagreeable experiences

Almost all of our emotional experiences, whether agreeable or disagreeable, produced vaso-constrictions.

The results yielded by the study of affective states, by means of instruments for recording changes in circulation, pulse, and respiration and other physiological manifestations do not afford at pres ent a very encouraging outlook for the solution of our problem, for in so far as pleasurable and painful experiences are not associated with uniform outward expressions it is difficult to obtain clear evidence of the accompanying internal physiological states. It is commonly assumed that there is something in pleasure and pain or their physiological correlates that reinforces or inhibits, as the case may be, the responses from which these states result. What this something is and how it produces its effects are problems for which a satisfactory solution has not been offered. There has been a bewildering variety of theories of the nervous correlates of pleasure and pain and of pleasantness and unpleasantness but there has been little attempt to apply these theories to explain the mechanism of profiting by experience, and it is difficult to see how most of these theories would help us in regard to this matter even were the} established.


158 S. J. HOLMES

A new point of view in regard to our problem has been presented by Hobhouse in his Mind in Evolution. To illustrate this view let us recur to our chick. When a nasty caterpillar is seen for the first time the visual stimulus sets up a pecking reaction. This is followed by the stimulus of a bad taste which sets up various rejection movements, such as ejection of the food and wiping the bill. The order of events is

stimulus. . . .pecking. . . .bad taste. . . .rejection.

When the same kind of caterpillar is met with a second time the stimulus tends to elicit the rejection movements with which it has been associated instead of the movements of pecking. Is not the inhibition due to the fact that the stimulus has become associated with a response which is incongruous with the first? Movements of rejection and avoidance are incompatible with those of pecking and swallowing and it may therefore be unnecessary to look to any peculiarity of the physiological correlates of pain for an explanation of the inhibition of the original reaction. The stimulus becomes coupled with a new reflex arc; nervous energy is drained off in a new channel, and the future behavior becomes changed. If the taste is a very bad one, a great deal of energy is involved and the connection with the rejection response made very permeable and the rejection movement easily set up. If a person is confronted with a sight of some nauseating medicine he has recently taken, avoiding or rejection movements are set up, such as making a face or even retching movements of the stomach. Is it not these movements or attempts at movements that really inhibit the taking the medicine? This is evidenced by the chick described by Lloyd Morgan, which after an experience with a nast> caterpillar approached one a second time but stopped and wiped its bill and went away as if it actually repeated its first experience. Of course inhibition of the original response does not always involve contrary movements but there may be impulses to such movements which do not issue in action. The principal feature in the modification of action through painful experiences is the assimilation of impulses incongruous with the original one.


PLEASURE, PAIN AND INTELLIGENCE 159

In the reinforcement or stamping in of a reaction to a particular stimulus that brings pleasure, it certainly seems as if pleasure or its physiological correlate in some way serves to cement more firmly the association between the stimulus and the response. Let us consider, however, the case in which the chick pecks at a caterpillar which has a good taste; the presence of the caterpillar in the mouth excites the swallowing reflexes; in the presence of a similar caterpillar the pecking response is made more readily than before and whatever hesitation there may have been at first disappears. Is not the difference from the pain response due to the fact that there is an organic incompatibility between the first and second responses in the pain response, while there is an organic congruity or mutual reinforcement of these responses in the other? Pecking and swallowing form the normal elements of a chain reflex; when one part of the system is excited it tends to excite the rest, to increase the general tonus of all parts concerned in the reaction. Many reflexes instead of being mutually inhibitory, tend to reinforce one another's action. According to Sherrington,

When in the spinal animal the one fore foot is stimulated, flexion of the hind leg of the crossed side is often obtained. Stimulation of that hind foot itself also causes a like reflex of that limb. When these two are concurrently stimulated, the flexion movement is obtained more easily than from either singly. These widely separate reflex-arcs therefore reinforce one another in their action on the final common paths they possess in common. Similarly with certain reflex-arcs arising from the skin of the pinna of the crossed ear. In them excitation reinforces that of the just mentioned arcs from the fore foot and opposite hind foot.

The presence of savory food in a dog's mouth causes the secretion of saliva and the movements of chewing and swallowing, and the stomach at the same time may be stimulated to secrete gastric juice. These activities are organically associated and they are usually preceded bv seizing acts of various kinds. A particular object, then, which evokes the seizing response and which is of a character to set up these other reactions becomes more readily responded to again. The seizing reaction becomes assimilated to the other reactions which dispose of the food.


100 S. J. HOLMES

Let us illustrate this view by the results of some experiments on the crayfish. If a piece of meat is placed a short distance in front of a crayfish, the first response is usually a slight twitching of the outer ramus of the antennules ; this is followed by chewing movements of the mouth parts and restless movements of the legs; the small chelipeds are moved back and forth and grasping motions are made by the small pincers as if in the endeavor to find some object. These movements may be followed by walking, and exploring movements of the large chelipeds. The coordinated movements of the antennules, mouth parts and legs may be regarded as a complicated form of chain reflex. If now we apply a stimulus to any of the organs concerned, it tends to set up the reflexes in the rest. If a drop of meat juice be applied by means of a capillary pipette to the tip of a small cheliped the first response is usually a twitching movement in the chela followed by an exploring movement of the limb. This is followed by similar movements of the other chelipeds and chewing movements of the mouth parts, and these by the twitching of the outer ramus of the antennules. If the stimulus be applied to the maxillipeds the chewing movements of the mouth parts are followed by the movements of the antennules and the legs. Any reflex element in this chain of reflexes involved in food taking tends to set off all the others.

I have trained crayfish, by feeding them by hand, to come toward me to get meat. At first I would very slowly bring a piece of meat held in a fine forceps near the antennules. After the movements of the antennules and mouth parts the grasping movements of the chelipeds would result in securing the meat. After some trials I would not allow the meat to be pulled away from the forceps until the cra} r fish struggled awhile to secure it ; at the same time I moved my hand about so as to accustom the animal to my movements. There is a struggle between the instinct to flee from a large moving object and the instinct to secure a savory morsel which has been seized. With careful management the latter instinct may be made to predominate over the former and gradually the fear of one's movements becomes much reduced. The crayfish finally came to associate the approach of my hand with being fed and would rear up and hold out its large chelae,


PLEASURE, PAIN AND INTELLIGENCE '61

much as in the ordinary posture for defense. Crayfish and crabs often assume this attitude while alarmed and retreating from danger, but the crayfish would come toward my hand and moreover would react in this way only when hungry, so that the response was not a fear reaction but one showing rather the absence of fear. One individual would greet me as I entered my room in the morning by raising up its chelipeds and coming toward me, and it would follow me about as I went from one side of its inclosure to the other. When fed, however, it would manifest no further interest in my movements. I had come to mean food and was responded to, I fancy, much as the crayfish would respond to a small object of prey which it could approach without fear. A part of the crayfish's congenital endowments is the instinct to approach small moving objects, as it is to flee from large ones. This instinctive reaction to small objects normally precedes the movements of seizing and devouring, just as the latter are preceded, though more uniformly, by twitchings of the antennules; the two sets of processes are parts of a normally associated chain of events.

The appearance of a large moving object has been associated with a mode of response (food taking) which is incongruous with the fear response : the latter is therefore inhibited. The reaction proper to the small object normally pursued has become joined to the food taking activities and the large object is followed. The large object may mean food, but there must be supplied an impulse to go toward the object. This probably comes, not from a fiat of the creature's will, but from an already existing instinctive proclivity to pursue objects of possible interest or food.

According to the view here presented, whether a particular response to a stimulus tends to be repeated more readily or discontinued, depends not upon the peculiar physiological state which may be produced in the brain, but upon the kind of responses which the stimuli brought by the act call forth. If an outreaching reaction becomes coupled with a withdrawing response the result is inhibition. If the reaction, on the other hand, brings stimuli which produce congruent reactions the association formed with these latter reinforces the first reaction. The pleasure-pain


162 S. J. HOLMES

response then resolves itself into the formation of associations. Withdrawing and defensive responses are usually initiated by pain giving stimuli and the instinctive or random movement which brings a painful stimulus is inhibited under similar conditions in the future, not because 1 of the pain of its physiological correlate but because it comes to bo associated with a withdrawing or defensive and hence an incongruous or inhibitory reaction. Pleasure and pain thus interpreted have no mysterious power of stamping in or stamping out certain associations. Whether the result is reinforcement or inhibition depends on the way in which a reaction and the secondary responses resulting from the situation in which the organism is thereby brought, happen to harmonize. The step from instinct to intelligence viewed as a physiological process involves, therefore, no essentially new element beyond the well known physiological properties of the nervous system, and we are not committed to any particular hypothesis as to the physiological accompaniments of pleasure and pain, or pleasantness and unpleasantness, in order to understand how behavior may become adaptively modified. How far the interpretation given will enable us to explain the development of intelligence I do not pretend to say. It may break down m attempts to apply it to higher forms of learning, but it affords a useful working hypothesis and takes us a way, I think, towards the solution of our problem.

BIBLIOGRAPHY

Alechsiepp,N. Die Grundform dcr Gefuhle. Psych. Studien, Bd. 3, p. 156.

1907 Angell, J. R. and Thompson, H. B. A study of the relations between certain

1899 organic processes and consciousness. Psych. Rev., vol. 6, p. 32. Bain, A. The emotions and the will, 4th ed.

1899

1899 The senses and the intellect, 4th ed. Baldwin, J. M. Mental development in the child and the race, 2d ed.

1903

1902 Development and evolution, Berger, H. Ueber korperliche Ausserungen psychischer Zustande.

1904 Boggs, L. P. An experimental study of the physiological accompaniments of feel 1904 ing. Psych. Rev., vol. 11, p. 223.


PLEASURE, PAIN AND INTELLIGENCE 163

Brahn, Max. Experimentelle Beitrage zur Gufiihlslehre. Philos. Studien, Bd.

1903 18, p. 127.

Ebbinghaus, H. Grundziige der Psychologie.

1905 Feilchenfeld, H. Ueber das Wesen des Schmerzes. Zeit. fur Sinnesphysiol.,

1908 Bd. 42, p. 172.

1908 Ueber den Blendungsschmerz. I.e., p. 313. Fere, C. Sensation et Mouvement, Paris.

1887 Fite, W. The place of pleasure and pain in functional psychology. Psych. Rev.,

1903 vol. 10, p. 633.

Gilman, B. L. Syllabus of lectures on the psychology of pain and pleasure. Am.

1893 Jour. Psych., vol. 6, p. 1.

Goldscheider, A. Ueber den Schmerz. Berlin.

1894 Hayes, S. P. A study of the affective qualities. Am. Jour. Psych., vol. 17, p. 538.

1906 Herrick, C. L. Modern algedonic theories. Jour. Comp. New., vol 5, p .1.

1895 Hobhouse, L. T. Mind in evolution.

1901 James, W. The physical basis of emotions. Psych. Rev., vol. 1, p. 516.

1894 Kelchner, M. Die Abhangigkeit der Athem- und Pulsveranderungen vom Reiz

1905 und vom Gefuhl. Arch. ges. Psych., Bd. 5, p. 1.

Kulpe, O. Outlines of psychology.

1895 Lagerborg, R. Ueber die specifischen Ursachen der Unlust- und Lustgefuhle.

1906. Ska?id. Arch. f. Physiol, Bd. 18, p. 47.

1905 Das Gefiihlsproblem. Lehmann, A. Die Hauptgestze des menschlichen Gefuhlslebens.

1892

1899-1901 Korperliche Aiisserungen Psychischer Zustande.

1905 Elemente der Psychodynamik. Lipps, T. Vom Fiihlen, Wollen und Denken.

1902 Marshall, H. R. Pleasure, pain and aesthetics.

1894

1908 Pleasure-pain and sensation. Philos. Rev., vol. 1, p. 625.

1908 The methods of the naturalist and psychologist. Psych. Rev., vol. 15, p. 16. Meyer, M. The nervous correlate of pleasantness and unpleasantness. Psych.

1908 Rev., vol. 15, pp. 201 and 292.

Meynert, T. Psychiatrie.

1885 Morgan, C. L. Habit and instinct.

1896

1900 Animal behavior.


164 S. J. HOLMES

Munsterberg, H. Beitrage zur experimentellen Psychologie, IV.

1892 \ m)ejde, D. C. Die biologische Theorie der Lust und Unlust, I.

1908 Nichols, H. The origin of pleasure and pain. Philos. Rev., vol. r, pp. 403 and 518

1892

Pain nerves. Psych. Rev., vol. 2, p. 487 and vol. 3, p. 309 Orth, J. Gefiihl und Bewusstseinslage.

1903 Pikler, J. Das Grundgesetz alles neuro-psychischen Lebens.

1900 Richet, C. Etude biologique sur la douleur. Rev. Set., tme ser. t. 6, p. 225, and

1896 Jour, de Neurol, et d'Hypnot., t. 1.. pp. 396 and 416.

1902 '•Douleur" in Diet, de Physiol'., t. ">, p. 173. Royce, J. Outlines of psychology.

1903 Shepard, J. F. Organic changes and feeling. Am. Jour. Psych., vol. 17, p. 522.

1906 Sherrington, C. S. The integrative action of the nervous system. A < w York.

1906 Shields, T. E. Effect of odours, irritant vapours, etc., upon 1 he blood flow.

1S96 Jour. E.cpcr. Med., vol. I, p. 71.

Spencer, H. Principles of psychology, 2d ed.

1873 Stout, G. F. Analytical psychology.

1896 Stanley, H. M. Evolutionary psychology of feeling.

1895 Sttjmpp, C. Ueber Gefuhlsempfindungen. Zeit.f. Psych., Bd. 44, p. 1. .

1906 Thorndike, E. L. Animal intelligence. Monogr. Suppl. Psych. Rev., vol.2., Xo. 4.

1898 Titchener, E. B. An outline of psychology, new ed.

1900

1908 The psychology of feeling and attention. Tsisch, \Y. von. Der Schmerz. Zeit.f. Psych., Bd. '2li, p. 11.

1901 Wolley, Ft. Sensory affection and emotion. Psych. Rev., vol. 1 1. p. 329.

P. H 17 \\'r\i)T, W. Grundztige der physiologische Psychologie.

1903 Zonepf, P. und A 1 1 : ( \i w\. E. (Jeberden Begleiterscheinungen psychischer Vor 1903 gange in At hem und Puis. Philos. Studien., Bd. is, p. 1.


THE FORMATION OF HABITS AT HIGH SPEED 1

OTTO C. GLASER

From the Zoological Laboratory of the University of Michigan

WITH TWO FIGURES

CONTENTS

1 . Introduction 165

2. Methods 166

3. Genera! results 167

4. Specific results 171

5. The senses and the habit 179

6. The kinsesthetic chain 182

7. Summary and conclusion 183

8. Bibliography 184

INTRODUCTION

The method by which habits are ordinarily educed consists essentially in presenting a problem whose solution depends on the slow, and often painful suppression of irrelevant actions, and the survival of only those that count. The results so achieved are invaluable, but from the nature of the case difficult to verify. So much time is required that students of all classes are apt to be told in words, rather than by actual experiments, what has been accomplished in this interesting field. Habits formed slowly and gradually are not only incapable of quick demonstration, but the "slow method" leaves altogether untouched, a wide range of behavior. Animals do not always act slowly; they do not always overcome, with deliberation and care, the difficulties that block

1 Directly, as well as indirectly, I am indebted to Miss Frances J. Dunbar, and to Miss Nina Gage, for many of the results on which the present communication is based.

THE JOURNAL OF COMPARATIVE NEUROLOGY AND PSYCHOLOGY, VOL. 20, NO. 3.


166


OTTO C. GLASER


the road to food and comfort; indeed, in nature there are many problems whose solution must be accomplished at once with the utmost rapidity, and to the swift alone is the race.

METHODS

The animals used were white rats of different ages, and the general method, a modification of the time-honored labyrinth. Instead of the usual form, however, I constructed a zinc tank, 2 feet, 1 inch square, and 6 inches deep, and covered it with coarse-mesh zinc netting, in the exact center of which a circular opening, fitted with a cylindrical shoot, serves as an entrance for the animals. At each of the four corners the cover, which is firmly clamped to the sides of the tank, has a small hinged door that can be opened at will, or made fast, likewise by means of clamps.



Fig. 1. Tank. Photograph by Miss Frances Dunbar. Dimensions 2 feet 1 inch square: depth, (i inches; diameter of funnel. 4 inches; height of funnel, 6 inches; exits. ;; inches square.


THE FORMATION OF HABITS AT HIGH SPEED 167

During the experiments, the tank was filled, sometimes with warm, sometimes cold water; salt was added on some occasions, whereas on others additional stimuli were administered while the animals were falling through the shoot or after they had emerged through the open door. As there are four openings, any one of which may be made the correct one, and as the contents of the tank may be varied in many ways, and the experiences on entering and emerging complicated as much as one desires, the tank is in a real sense, labyrinthine, although simple in construction.

GENERAL RESULTS

It is needless to say that an inexperienced animal suddenly thrust into a tank of water makes strenuous efforts to escape. Under the conditions of the experiments it is not surprising that a high degree of variability should attach to the several attempts of an animal forced to undergo the experience of the tank half a dozen times in succession. Nevertheless there is an underlying regularity, for the time taken to escape in 83 per cent of the cases is less at the last attempt than at the first, whereas in 16 percent it is greater, and in only 1 per cent unchanged.

Time records, while the most convenient form of registration, are nevertheless not the only ways in which the formation of a habit manifests itself. Very much to the point in this connection are tracings of the actual pathway pursued in escaping. Six such graphic representations are given in fig. 2, and show conclusively that the first turn of the path that led to the first escape occurs, often much abbreviated, and regardless of advantage, in nearly all of the succeeding trials.

In experiments into which so many complicating factors enter a regular and machine-like progress toward perfection cannot be expected, and its failure to appear is clearly shown in series A and B. Records C and D, however, illustrate distinctly how the path followed was simplified until in the last trials it became to all intents and purposes a straight line. Hand in hand with this was a reduction in the amount of time taken to effect the escape.


168


OTTO C. GLASER


Two facts of considerable interest are to be read in these tracings. In series B, C, D, and E, the pathways are without exception dextral; in series E, on the other hand, all the turns made are sinistral. A directive factor seems to be operative, but analysis shows that it may not be simple. A considerable number of tests was made with blindfolded animals, and with young whose eyes were not yet open. The results in general indicate that certain individuals have a natural disposition to turn to the right, rather than the left, in swimming, whereas in others the reverse tendency is equally marked. I have made no experiments to determine whether this tendency is due to differences in the semi-circular


TRIALS



TIME

Fig. 2. Graphic representation of six sets of trials made by six individuals. The squares represent the tank, the incomplete corner, the open exit. Above the first series of records are figures indicating the number of each trial, whereas beneath the squares are the time records made in traversing the pathway indicated within each square. In every case the animal started from the central dot.


THE FORMATION OF HABITS AT HIGH SPEED 169

canals, or to inequalities in the right and left swimming musculature, or the neuro-muscular coordinations, or to a combination of these possible factors. Any one of them, or any combination, except of course a compensatory one, might be responsible for the fact that some animals naturally turn to the right, others to the left.

The determination of the true basis of this behavior would be very difficult because the repetition of the tests that bring it forth leads to the formation of a habit. That such becomes established, is, I think, sufficiently clear from the diagrams, as well as from the time records beneath them. Very interesting are series of which A is typical. In these, an animal, either because of its structure or its habit, turns consistently in one direction, in the present example, to the right. For some reason, exactly the reverse course was taken, with great success, at the fifth trial, and the sixth shows distinctly how this new departure complicated the course.

Series A illustrates a second fact, namely, that in many cases the pathway in succeeding trials becomes more complex, and the time consumed in escaping may increase. Nothing would be further from the truth than the conclusion that increase in either the complexity of the path or in the length of time, indicates progressive stupidity. The conditions of the experiments are such that a property analogous to cool-headedness in man, is at a premium. In many instances the animals became increasingly nervous, and lost their heads.

In problems w r hose solution can be effected, provided the animal has enough time and makes sufficient movements, this complication does not enter, but more than mere activity is necessary in a crisis. Animals which by luck or otherwise succeed early, behave much as they would in the face of "slow" problems; but animals that fail, or succeed only with difficulty, quickly become so handicapped by fear, and the useless activity characteristic of that state, that they fail more completely or succeed with much greater difficulty in later trials than they would in corresponding trials under more favorable conditions. While the experience of the tank brings out forcibly innate differences in the capacity to do the right thing quickly under stress, it is


170 OTTO C. GLASER

hardly adequate to measure real differences in the ability to learn. The individuals which are thrown out of commission, by the very nature of the problem itself, are simply out of the running; they are unable to compete, and their failure is no more to be ascribed to inferior intelligence than the failure of a blacksmith to win prizes at a swimming meet is to be ascribed to the superior strength of the other contestants.

Objectively, of course, the facts are that some animals fail, whereas others succeed. Success may come with progressive reduction in the complexity of the path, and in the time taken to traverse it, or with either of these elements separately; failure may result from the corresponding opposites. We can say with justice that the capacity to learn also expresses itself in one or more of these ways, but the obvious inference would not be the correct one. Should the chain of activities that we are interested in be inaugurated, our objective measurements would give us the information we want, but when some other chain is set up, the measurements show simply the extent to which the second disturbs the first chain. In other words, if the problems convert the animal into a non-learning mechanism, the movements which the individual performs throw no light on its learning capacity. What they do show is, that under the given circumstances, some animals improve, others do not, but the capacity of those that fail remains unknown. Even those that succeed do not divulge all. Strictly speaking, their capacity is shown to be not less than the records indicate, but actually it may be considerably more.

The animals that improve emphasize a practical question of some importance. There are plainly two ways of solving the tank problem; by increasing speed, and by decreasing the length of the pathway. Cases in which these two go hand in hand, or in which constancy of speed is offset by an abbreviation of the pathway, offer no difficulty; but what shall we say when increased complexity of the pathway is compensated for by heightened speed? If we limit ourselves entirely to the objective time measurements, such an individual may seem to improve or to hold its own; if, on the other hand, we study the pathway alone, the animal is clearly loosing in fitness. The sixth trial in series A is a concrete illustration. How shall such an individual be rated?


THE FORMATION OF HABITS AT HIGH SPEED 171

As the objective records are at variance, it seems at first sight arbitrary to fix on either one or the other, and to say this is a measure of the truth.

Series C and D are from animals naturally well fitted to cope with the problem presented; natural fitness, however, varies, and some individuals inaugurate an overpowering set of altogether irrelevant, interfering movements. Such animals, practically, are not competing. Between these extremes most of the other individuals take their places, for their irrelevancies are not sufficient to destroy all chances of success, though marked enough to affect the general averages. The problem is to get out of the tank as quickly as possible, and is solved both by the animal that reduces the complexity of its path, either with or without an increase of speed, and by the animal that compensates inferiority in one direction by superiority in another. The latter might even win, but if we were offering prizes, justice would demand one for improvement in speed, the other for improvement in form. As form in the end makes for speed, not in individual cases, but on the average, the time records may be adopted officially as a practical, thougn not necessarily complete, measure of fitness.

SPECIFIC RESULTS ADULT RATS

The first experiments were made with a tank differing slightly in size from the one described, and the animals, instead of being dropped into the water through a shoot that landed them in the exact center, were thrown in from one side, with a slight whirling movement. There was but one corner opening through which escape might be effected. The problem was thus essentially the same as the one described, and the differences in the records are mainly due to differences in the sizes of the two tanks, and to the way in which the animals were introduced. The whirling movement was adopted in order to insure that the effort to escape might not be begun in any except a chance direction. The results are tabulated below in self-explanatory form.


172


OTTO C. GLASER


ANIMAL I.



ANIMALS


ii.


ANIMAL III.


Trials


Time


Intervals


Trials


Time


Intervals


Trials


Time


Intervals


1


19.0"



1


14.0"



1


6.0"


1'30"


2


3.5"



2


11.0"



2


9.0"


l'lO"


3


2.0"


V


3


9.0"


1'


3


8.4"


1'30"


4


5.0"


15'


4


5.0"



4


26.0"


2'03"


5


3.5"



5


2.0"


15'


5


1.4"


2'00"


6


2.5"



6


2.9"



6


1.0"


Hours





7


17.0"



7

8

9

10

11

12


1.4" 1.2" 1.6" 2.4" 2.0" 1.0"


1'32" 1'20" 2'40" 0'40" 0'45"


Characterization, Animal I: At first many trial movements, then stopped investigating. In the end did not hurry out. Averages: first three trials, 8.16"; second three trials 3.66". Improvement, 55.1 per cent.

Characterization, Animal II: Accidental delay at second trial. Numerous extra movements in fourth. In seventh gave up trying to get out, hence this record is omitted in the averages. Averages : first three trials, 11.33"; second three trials, 3.33". Improvement, 70.8 per cent.

Characterization, Animal HI: Slight delay in starting watch at eighth trial. Averages: first six trials, 4.73"; second six trials, 1.60". Improvement, 66.3 per cent.


ANIMAL IV.


ANIMAL V. ANIMAL VI.


Trials Time


Intervals


Trials


Time


Intervals


Trials


Time


Intervals


1

2

3

4

5

6

7

8

9

10

11

12


57.4" 3.2" 8.2"

12.3" 8.2" 8.2" 9.3" 4.4" 4.3" 3.2" 3.1" 5.0"


30" 30" 30" 30" 30" 20" 30" 30" 30" 30" 30"


1 2 3 4 5

6

8

9

10

11

12


55.2" 7.1"

5.1"

6.3"

5.2"

15.3"

3.3"

6.0"

12.2"

19.2"

4.0"

4.0"


30" 30" 30" 30" 30" 20" 30" 30" 30" 30" 30"


1 .2 3 4 5

6

7

8

9

10

11

12


13.1"

18.1"

4.3"

5.1"

5.2"

3.3"

12.2"

13.1"

22.3"

8.0"

31.4"

4.3"


30" 30" 30" 30" 30" 20" 30" 30" 30" 30" 30"


THE FORMATION OF HABITS AT HIGH SPEED 173

Characterization, Animal IV: Animal quiet. Averages: first six trials, 16.25"; second six trials, 4.88." Improvement, 70 per cent.

Characterization, Animal V: Animal quiet. Averages: first six trials, 15.7"; second six trials, 8.11." Improvement, 48.4 per cent.

Characterization, Animal VI: Animal quiet in the first six trials, and apparently improving at a good rate. In the second set, however, it suddenly became "cranky" and did everything except the expected; it swam around aimlessly, or clawed the covering screen. Averages: first six trials, 8.18"; second six trials, 15.20". "Deterioration," 85.8 percent. This record is of necessity omitted in the later calculations.

Naturally there is much unevenness in these records, and at times the irregularities seem to obscure the evidence that a habit was formed, or to show that the exact opposite was established. Nevertheless there is a fundamental harmony beneath the discrepancies, and this is not destroyed even when the record of animal VI is admitted. For obvious reasons, however, it would be unfair to allow this animal to figure in calculations that involve the group of animals as a whole. On similar grounds I have excluded the 7th trial of animal II, for on that particular occasion this individual also very clearly did not attempt to escape. If with these modifications the results be taken as they stand, we can construct the following table for comparisons:




FIRST HALF RECORD



SECOND HALF RECORD




A verages




Averages




Seconds




Seconds


Animal


I


8.16..




3.66


Animal


II


11.33 .




3.33


Animal


Ill


4.73..




1.60


Animal


IV


16.25..




4.88


Animal


V

al average


15.70..




8.11



Gene


ral avi


Gener


11.23


srage 4.32


A glance at this record shows that experience, even when limited to very brief periods of intense activity, has its effect, and in the adults composing this group brought about on the average an increase of 6.89 seconds in speed, or an improvement of 62.1 per


174


OTTO C. GLASER


cent. Considering the unfavorable conditions for the formation of a habit which these experiments present, the result is very marked.

YOUNG RATS

That this conclusion is not a mistaken one, is indicated by a similar treatment of the records of very young animals that underwent the experience of the tank. The individuals whose performances are tabulated in detail below were all from the same litter, and were aged three and a half weeks.


ANIMAL A.



ANIMAL B.



Trials


Time


Intervals


Trials


Time


Intervals


1


9.0"


1'25"


1


5.0"


15"


2


4.0"


44"


2


13"


37"


3


5.0"


19"


3


14"


27"


4


2.5"


43"


4


10"


25"


5


5.0"


28"


5


4.5"


30"


6


2.0"


45"


6


5.0"


31"


7


3.0"


23"


7


7.0"


1'2"


8


1.3"


45"


8


3.5"


1'9"


9


1.5"


l'lO"


9


4.5"


l'S"


. 10


1.0"


1'15"


10


2.0"


58"


11


2.0"


56"


11


5.5"


32"


12


2.0"



12


6.0"



Characterization, Animal A: Animal between three and four weeks old; quick in its movements and not easily confused. Took a new route at the fifth trial, but at the sixth headed at once in the proper direction. Averages: first six trials, 4.58"; second six trials, 1.8". Improvement, 60.7 per cent.

Characterization, Animal B: Animal between three and four weeks old; quick but irregular and not to be counted on. In the twelfth trial ducked under just before emerging, and lost considerable time. Averages: first six trials, 8.56"; second six trials, 4.75". Improvement, 44.6 per cent.


Taking the figures as they stand, and comparing the average of the first half-record of each animal with the average of the second half-record, we get the following:


THE FORMATION OF HABITS AT HIGH SPEED


175


FIRST HALF RECORD SECOND HALF RECORD

Averages Averages

Seconds Seconds

Animal A 4.58 1 .80

Animal B 8 . 56 4 . 75

Animal C 13.75 4.80

Animal D 6.43 4.54

General average 8.33 General average 3.94

Here, too, is evidence that the behavior was modified by the "tank experience," but much more interesting is the fact that the changes undergone seem to harmonize very closely with the results secured by Watson ('03) under much more favorable conditions. The increase in the speed of the young animals was


ANIMAL C.


ANIMAL D.


Trials


Time


Intervals


Trials


Time


Intervals


1


8.5"


42"


1


10.5"


24"


2


33.0"


22"


2


8.5"


35"


3


19.0"


20"


3


5.8"


1'10"


4


9.0"


34"


4


5.8"


40"


5


9.0"


36"


5


6.0"


47"


6


4.0"


53"


6


2.0"


56"


7


3.0"


1'


7


7.0"


40"


8


3.0"


1'4"


8


3.0"


53"


9


3.0"


43"


9


3.3"


l'lO"


10


2.8"


17"


10


6.0"


58"


11


4.0"


1'28"


11


4.3"


42"


12


13. 0"



12


3.5"



Characterization, Animal C: Animal between tliree and four weeks old; very quick and nervous. In first trial explored all corners, became panic-stricken in the second, better in the third, but during the fourth lost much time clawing the wire screen. In the last half of the series the animal seemed to have found itself, and exhibited no semblance of fear. The length of the twelfth trial is due to having overshot the opening. Averages: first six trials, 13.75"; second six trials, 4.8". Improvement, 64.4 per cent.

Characterization, Animal D: Animal between three and four weeks old. Relatively slow and indifferent. Averages: first six trials, 6.43"; second six trials, 4.57". Improvement, 29.9 per cent.


176


OTTO C. GLASER


on the average 4.39 seconds, for the adults 6.89 seconds; the former represents an increase of 53 per cent in efficiency, whereas the adults improved 62.1 per cent.

DURATION OF THE HABIT

That a habit formed under stress during brief periods of intense activity, may endure from fifteen minutes to several hours, the records already given show plainly. Special tests were made, however, to determine, if possible, whether any traces of the heightened efficiency might be found after several days. For this purpose, animals A, B, and C, of the preceding group of young were used, and adults A, B, C, D, E, and F, upon whose records fig. 2 is based. In the present connection I shall distinguish the three young animals by small letters, the six adults by means of capitals.


AVERAGES


4/26 4/26

4/27

4/26

4/26 4/27 5/9 5/9

4/26 4/26

4/27


2.15 p.m. 3.28 p.m. 2.15 p.m.

2.30 p.m.

3.35 p.m.

2.30 p.m. 10.20 a.m. 10.45 a.m.

2.50 p.m. 3.43 p.m. 2.50 p.m.


Seconds 4.58 1.34 4.69

5.15 4.75 3.30 5.65 2.10

13.40 4.79 7.00


The adults A, B, C, D, E, and F, whose records on January 15th, 20th, 22d, and 28th follow, were all used prior to December 11th of the year before, but unfortunately the earlier experiments were not performed under exactly the same conditions as those made in January, and a comparison of records separated in time by more than a month cannot be made. The results of the series separated in time by five days, by two, and again by six, can, however, be safely compared.


THE FORMATION OF HABITS AT HIGH SPEED


177



DATE


DATE


DATE


DATE



Jan. 15


Jan. 20


Jan. 22


Jan. 2S



Seconds


Seconds


Seconds


Seconds


A


10.40


1.66


1.91


2.08


B


6.90


10.25


5.83


6.00


C


5.90


9.08


13.75


10.50


D


6.41


5.66


6.08


6.08


E


6.08


32.82


11.41


9.75


F


15.75


19.00


16.00


10.16


General average


8.58


13.08


9.16


7.33


The results from the young animals are not especially conclusive. Taking the records as they stand, however, it may be said that two out of the three seemed to show the effects of their experiences twenty-four hours afterward, and one of the animals, after twelve days, very quickly recovered and in the end actually bettered its previous best record. It is to be expected, of course, that individuals vary greatly in the length of time which their habits, whether formed slowly or rapidly, endure; and furthermore, there is no way in which one can be reasonably certain, except by the method of multiple instances, whether, in the long run, animals without previous experience might not, on the whole, make as good records as those made by experienced individuals. As far as the group as a whole is concerned, it does not seem to bear markedly one way or the other on this particular question, although it does show that the final records of animals b and c were better than their initial ones.

Considering the adults, not as individuals, but as a group, and comparing the averages of the four series, we get the results as tabulated. Individual differences notwithstanding, we may say that, on the whole, the second set of trials, five days after the first, was distinctly slower; the third, two days after the second, not as slow as the preceding, whereas the fourth, six days after the third, was the best of all. Of course, these statements have no bearing on specific cases. The record as given is sufficiently detailed to show how the relative fixity in one individual is balanced, and even discounted by the relative instability of another, or vice versa.


178


OTTO C. G LASER


INTRODUCTION OF ADDITIONAL STIMULI


The "tank experience" may be complicated by the introduction of additional factors, which, classified by their effects, may be called retarding and accelerating stimuli. Marked accelerations were produced by allowing the animals before reaching the water in the tank, to fall through a paper bag filled with water considerably warmer; also if, during their fall through the cylinder, the animals received an electric shock, the speed was increased. On the other hand, the addition of salt in small quantities to the water in the tank had, on the average, a depressing effect. The most marked, as well as constant effects, however, were go+ten by the system of desired rewards. Animals that had been without food for twenty-four hours, were dropped into disagreeably cold water, and immediately on escaping, were wrapped in warm towels and given a nibble of cheese. All these factors helped to make escape from the tank better worth while than it had been in any of the preceding experiments. The results, based on two full-sized adults, and on eight young, less than a year old, are tabulated below. As heretofore, the adults are labelled with capitals, the young with small letters.



FIRST


HALF


SECOND HALF



ANIMAL






IMPROVEMENT



Record


Time


Record


Time




Trials


Seconds


Trials


Seconds


Per Cent


AA


6


7 45 15.5


6 3


3.71

8.2


51


BB


3


48


aa


6


10.8


6


3.2


69


bb


6


13.03


6


5.26


60


cc


6


9.8 16.65


6

6


7.3

12.66


26


dd


6


24


ee


6


13.3


6


7.2


46


ff


6


15.9


6


6.3


61


gg


6


18.8


6


3.2


83


hh


6


8.3


6


3.7


56


Not only are the adults too few, but the number of trials made is too small to allow of satisfactory comparison with the earlier records made under simpler conditions. Neither AA nor BB


THE FORMATION OF HABITS AT HIGH SPEED 179

improved as much as the adults in former experiments, but this may have been because they were unusually slow. If a conclusion may be hazarded, it is that the additional stimuli had no effect or perhaps a slightly depressing one, though this is uncertain.

The young, on the other hand, improved on the average 62.5 per cent, as compared with an earlier record of 53 per cent. Animals cc and dd have been eliminated from these calculations on the ground of over-excitedness. The results might have been expected. There was an unmistakable effect on the young, and in two cases the additional discomforts and rewards produced so much activity, that these particular records were vitiated by it.

THE SENSES AND THE HABIT

The relation between the senses and the acquisition of habits in the white rat has been so thorough^ worked over by Watson ('07) that I have performed only a few experiments, the results of which I shall present for two reasons, first because they are corroborative, in spite of the differences in method; secondly, because they show that the various senses can be eliminated without resorting to the extirpation, or destruction of the sense organs. Under the conditions of the experiments, hearing, smell, sight, and touch, either singly or in any combination, may conceivably furnish the sensual basis on which the habit rests.

HEARING

As the associations were formed while one of the four exits was open, it may have been that the movements of the animals were influenced by the differences in sound intensity or quality due to the unblocked passage. Such differences, if they exist, must be very small, for the closed openings are blocked simply by wire netting, but as rats are known to perceive minute sound differences, the point was well worth testing.

Associations were established in the usual way, after which


180 OTTO C. G LASER

the exit to which the animal had been trained, was closed and one or several of the others were opened. In every case the individuals continued to go to the exit from which they had previously escaped. Of course this test could not be continued indefinitely, as failure to escape from the tank, resulted in new exploratory movements that in the end led the animals to one of the other openings. Frequent visits to the old corner are characteristic of these secondary trials.

In another experiment, the association was established amid general noise, so great just above the center of the tank, that probably there were no differences at any of the corners. When the animals, after a brief rest, were placed again in the tank, they were allowed to make their second series of trials in absolute silence. In all cases they swam to the right place with no more variation in time or in the complexity of the pathway than is found under ordinary conditions.

A third test was made by allowing the association to establish itself while a noise was made at one of the other opened escapes. These noises, purposely not loud enough to frighten the animals, but certainly marked enough to be audible, seemed to have no effect whatever.

SMELL

To determine the part played by the sense of smell, the exit left open was thoroughly perfumed with cheese. After the association was established, the tank was turned in such a way that the perfumed opening was diagonally opposite its original position, whereas another exit, unperfumed, occupied geographically the place previously held by the escape to which the animal had been trained. Variations of this plan, such as holding a piece of cheese, first in one corner, then in another, were also tried, but in each case the animals swam in the direction that led to the earlier escapes. So far as these tests of the importance of smell are concerned, they give only negative results.


THE FORMATION OF HABITS AT HIGH SPEED


181


SIGHT

The experiments on sight bear simply on such rays as we ourselves are able to perceive, and were performed in a carefully constructed and thoroughly efficient photographic dark room. From the subjective standpoint, the animals were allowed to swim in the dark, but whether what we call dark is in reality dark to a rat, is another question.

Two sets of animals were used, two adults, and two youngones less than a year old. Each animal was given four sets of six trials. The averages follow:


ANIMAL


DATE


TRIALS


AVERAGES





Seconds


A


11/11

11/11


6

6


11.03


A


17.10


A


11/12 11/13


6 6


10.40


A


12.73


B


11/12


6


13.40


B


11/12


6


24.96


B


11/13


6


0.42


B !


11/13


6


11.76


a


11/11


6


18.23


a


11/12


6


7 05


a


11/12


6


6.40


a


11/13 11/13


6 6


2.00


b


90.00


b


11/13


6


50.24


b


11/14


6


31.56


b


11/14


6


10.90


If the first two series of A are compared with the last two, it will be found that this animal remained practically constant in speed; if a similar comparison be made for B, we find that this individual increased its speed on the average by 9.28"; whereas the records of a and b show increases in efficiency in each of the last three sets. Undoubtedly, then, the tank problem can


THE JCURNAL OF COMPARATIVE NEUROLOGY AND PSYCHOLOGY, VOL. 20. NO. ,?.


182 OTTO C. GLASER

be solved in the dark, and the solution, may with repetition, become a habit, but the evidence does not show that the sense of sight plays no part under other circumstances. Watson's negative evidence on this point seems very good, but the maze problem and the tank problem differ so much, that what is true in the solution of the one, is not of necessity true in the solution of the other.

TOUCH

Touch in the ordinary sense, is practically eliminated by the nature of the problem. The actual solution takes place in a medium in which the animals float. There are no solid bodies to be touched; no differences in homogenity in any direction, that might guide the individuals to the correct opening. The same thing applies to the temperature sense. It is possible, however, that touch and temperature, singly or together, affect the result, but if they do, it is not in the usual way, but as the initiating stimuli of other activities.

THE KINESTHETIC CHAIN

With hearing, smell, sight and touch, either eliminated, or shown to be not essential, the question of how the habit becomes established, naturally arises, Watson ('07) and others have presented a large body of evidence suggesting that the kinesthetic sense may be the controlling factor in the behavior of the white rat. My own experiments seem to bear similar interpretation, but if the facts corroborated by the tank method are due to a kinesthetic sense, it follows that this needs investigation.

Granted that a relation exists between the objective phenomena which we call a habit, and the inferred basal kinesthesia, the inference that the relation is definite, seems just; for the habit itself is definite, and capable not only of exact measurement, but also of modification. It follows that in the emergence of a habit, the internal basis also is modified, for the habit itself is modified behavior. In other words, kinesthesia is easier to understand if


THE FORMATION OF HABITS AT HIGH SPEED 183

thought of as a sequence of states which will be repeated, provided the physiological state of the animal remains favorable, whenever the stimuli that started it in the first place retmr.

It is here that any sensations associated with the beginning of a solution may come into play. Unfortunately, I have no results to offer on this subject at present, but the standpoint itself may not be without value. If kinsesthesia is indeed a special element in the sum total of internal conditions, is in fact a chain, based on the physiological, and expressed objectively in the habit, the answer to a difficult problem may be a little easier to find than heretofore.

SUMMARY AND CONCLUSIONS

My purposes in presenting these results have been to show the adequacy of the tank as an instrument for studying animal behavior; to show that our usual methods leave out of consideration a large range of interesting activities; and finally to demonstrate that behavior, limited to brief periods of very intense exertion under stress, may lead to the formation of habits.

Incidentally, the results seem to corroborate, as far as they go, some of the conclusions set forth in Watson's splendid work, and wherever at variance, are so probably from the differences between the problem of the maze, and the problem of the tank. In certain ways, the latter seems well suited for studying the "direction sense" or whatever it is that in the absence of sight, hearing, smell, and touch, enables the animals, not only to solve the problem, but to improve in efficiency. If kinsesthesia plays an important role, if indeed it is by this means that the animals sense direction, and if furthermore, it is a sequence of states, dependent on the physiological condition of the animal, and on certain initiating stimuli, then by variations in the last two categories, one should get changes in kinsesthesia, which in turn would be registered by corresponding differences in the objective habit.


Accepted by The Wistar Institute of Anatomy and Biology, March 18, 1910. Printed July 7, 1910.


184 OTTO C. GLASER

BIBLIOGRAPHY

Watson, John B. Animal Education. The University of Chicago Contributions 1903 to Philosophy, vol. 4, no. 2.

Watson, John B. Kinesthetic and Organic Sensations, etc. Psychological 1907 Review, Monograph Supplements, vol. 8.

Zoological Laboratory, University of Michigan. January 13th, 1910.


ON THE MEDIAN ANTERIOR CEREBRAL ARTERY AS FOUND AMONG THE INSANE

I. W. BLACKBURN

From the Government Hospital for the Insane, Washington, D. C.

WITH SIX FIGURES

The subject of anomalies of the cerebral arteries among the insane has received considerable attention in the past, especially noteworthy being the article by Bullen in the Journal of Mental Science, volume 36, 1890, 1 but unfortunately comparative statistics are meagre. In 1907 I made a study of the conditions of development of the encephalic arteries in 220 consecutive cases of mental diseases, making a comparison with the studies of Windle 3 in 200 cases of those presumably sane. The results of this comparison seemed to show a decided predominance, in general, of anomalous conditions among the mentally diseased. 3

In the study above mentioned' being chiefly concerned with the anomalies of the circle of Willis, sufficiently careful studies of other arteries were not made, and as a result, the anomalous vessel which forms the subject of this paper was not as frequently found as subsequent observations show is probably the fact. This artery is found of course among the sane as well as the insane. My studies, however, are based upon the examination of 400 consecutive cases of mental disease examined with special reference to this vessel. In all of the text books of anatomy at my com 'Bullen. Post mortem Examination of the Brain, etc. Journal of Mental Science, vol. 36, 1890.

2 Windle. On the Circle of Willis. Reports of the British Medical Association, for 1887. New York Medical Journal, vol. 2, 1888, and Journal of Anatomy and Psysiology, 1887-1888.

2 Blackburn. Anomalies of the Encephalic Arteries among the Insane. Journal of Comparative Neurology and Psychology, vol. 17, no. 6, 1907.


186 I. W. BLACKBURN

mand I have either found no reference to the artery at all, or a mere mention of it, as by Cunningham, as a third anterior cerebral artery sometimes present, or by Quain, as found in 4.5 per cent of cases. 4

In an article in the Journal of Nervous and Mental Disease, volume 12, no. 3, July, 1885, under the heading, "On a seldom described Artery (Arteria Termatica) etc," Professor Wilder describes the artery and gives it the name of Arteria Termatica, from the place of location and chief distribution in his cases. Professor Wilder thus describes the vessel which he found in a large percentage of human brains examined :

It usually soon divides into a right and left portion which supply respectively the cinerea forming the surface of the triangular area ventrad of the rostrum on either side, and then extend around the genu to the dorsal aspect of the callosum.

In another place in the same paper, he describes its origin, as most frequently from the place of junction of the precerebrals, the pre-communicant being absent, an important observation in connection with the present subject. In some respects most of my cases of this artery have varied somewhat from Wilder' s description, being as a rule larger, and following more closely the description of the median anterior cerebral artery as described by Windle, who says,

It passes along the longitudinal fissure for two-thirds of the length of the callosum, and divides into two branches, supplying the opposite surfaces of the hemispheres.

Windle found this artery present in 9 of the 200 cases examined, and as his description of it is unqualified, it may be concluded that most of his cases were of the larger-sized vessels, perhaps more properly named the median anterior cerebral artery, though I have no doubt that Wilder, Windle and I are describing the same vessel .

4 Quain's Anatomy, 1892. Other text books of Anatomy.


MEDIAN ANTERIOR CEREBRAL ARTERY 187

In a typical median anterior cerebral artery, the description is as follows :

The vessel arises from the anterior communicating artery, usually near the middle; curves upward, and in close proximity to the lamina terminalis, around the genu of the callosum, lying close to this body; and at about the middle or at the junction of the anterior two-thirds with the posterior third of the callosum it divides into two nearly equal branches which are distributed to the paracentral lobules and the quadrate lobules of the opposite hemispheres. In the course of the main trunk and the principal branches of the vessel, small and unimportant branches are given off to the lamina terminalis, the callosum, and to the adjoining gyri, but these ultimate branches have not been worked out, and are perhaps relatively unimportant. A large majority of these aberrant vessels have this origin and distribution, though occasionally the origin may be nearer to one anterior cerebral artery than the other. It may even seem to originate from one of these vessels near to the junction with the communicans, or what is still more important for our study, it not infrequently arises from the end of the junction of the two anterior cerebrals when these are fused and the anterior communicating artery as such is absent.

Differences in the size and distribution of the main branches are occasionally met with; sometimes one branch may be wanting, and the vessel is then distributed to one side only. In the small or undeveloped vessels as those described by Wilder, the branches are all small and irregular, and are sent mainly to the lamina terminalis, the rostrum callosi, and the gyri in the immediate vicinity. 5

In an article by Grunbaum and Sherrington in Brain, 1902,

'I find that on carefully removing the basal arteries and floating them out in water, often a long, slender, arteria termatica is unexpectedly revealed. This leads me to suspect that arteries of this type are quite frequently present but are undiscovered; probably the percentage of median anterior cerebral arteries might be nearly doubled if careful search were thus made. Whether these arteries are vestigial or reversionary, it is certain that in the majority of instances no such vessel arises from the anterior communicating artery, and that no corresponding vessel is mentioned among the branches of the artery by most anatomists.


188 I. W. BLACKBURN

"Note on the arterial supply of the Brain in Anthropoid Apes," these observers state:

Among features of human character pertaining to the anthropoid apes, we find one not hitherto recorded, namely the existence in the cerebral arterial supply of a Cir cuius Willisii resembling that of man. In these highest apes, as in man, an anterior communicating artery quite frequently completes the circulus in front;

and, quoting Parsons, they add :

In all of the lower mammals which I have examined, including the Platyrrhine monkeys, Cercopithecus, etc., I find no anterior communicating artery.

In all of the lower forms of mammals instead of the pair of anterior cerebral arteries, there is but a single azygos vessel, which in its course forward gives off branches to the right and left hemispheres respectively. These observers examined six brains of chimpanzees, and one orang-outang. Of the six chimpanzees, in five the circulus had the human type. In one of these two anterior cerebral trunks were united or fused into a common trunk about four mm. in length. In the orang-outang also the circulus was completed by an anterior communicating artery, and in one of the six chimpanzees there was a single azygos anterior cerebral artery as in the lower mammalian forms. This work, which is in accord with the writer's limited experience, shows that the existence of an azygos vessel in place of the anterior communicating artery and the two anterior cerebrals is the rule in the lower mammals, and that the form found in the higher anthropoid apes and in man is a phylogenous development while it also suggests that in view of the frequency of anomalous forms in the anterior cerebrals, that this phylogeny is as yet unstable.

In these arteries, anomalies of development are by all odds more frequent than in any other sets of vessels. Windle in 200 cases found 8 of fusion, 9 median anterior cerebral arteries, and a large number of anomalies of the anterior communicating artery; in my own 220 cases, fusion of the vessels was found in 7 cases, in


MEDIAN ANTERIOR CEREBRAL ARTERY


189



Fig. 1. a — e. Developmental types of the anterior system of cerebral arteries. Fig. 2. a — e. Reversionary types of the anterior cerebral arteries. Fig. 3. a — e. Types of anomalies of the anterior cerebral arteries.


190 I. W. BLACKBURN

10 the right artery was very small, and the opposite carotid artery sent the blood supply to the two hemispheres through an enlarged anterior communicating artery, and in 6 cases the left artery was the smaller, and the main blood supply came from the right carotid system; in a large percentage of cases, anomalies of the communicans were found. It is possible that in many of the apparent anomalies of the anterior cerebral arteries with disproportion in size of the trunks beyond the anterior communicating artery, the larger vessel may be a partial reversion to the primitive type ; that is, a median anterior cerebral unusual only in place of origin. In fact, it is not at all uncommon for the larger of the two trunks to send branches to the mesial surfaces of both hemispheres.

It may then be granted that the lower types are being developed in the phylogeny of the races, and such being the case, reversion to the primitive type may now and then be expected. It has seemed to the writer that the development and reversions might be shown graphically in the accompanying drawings, nearly all of which are drawn from actual specimens. In Fig. 1, a represents the primitive form as met with in the lower mammals; b shows the fusion of the anterior cerebrals with the remnant of the azygos vessel at the end of the junction; c a later stage of the development, with the remnant of the azygos vessel; d shows the shortening of the place of junction, and e shows a fully developed anterior communicating artery. Fig. 2 shows the occasional reversions met with ; a gives the type of arteria termatica of Wilder; b gives the type of termatic artery which arises from the end of the short fusion of the two vessels; c shows the most common form of median anterior cerebral artery; d, that in which the anomalous artery is the largest of the three, and e, the almost complete reversion to the lower mammalian type; e, however, has not actually been seen by the writer, while d is common. 6

It is suggested by these forms that in the higher phylogenetic development of the brain, the greater size of the frontal lobes has

Quain says: "The two arteries have also been seen united in a single trunk, which runs in the longitudinal fissure, giving off branches to both hemispheres." The writer has recently found one of these vessels, hence the series may be regarded as complete.


MEDIAN ANTEKIOR CEREBRAL ARTERY


191


necessitated a development of two anterior lateral branches of the azygos vessel into the two lateral anterior cerebral arteries, and that with this has progressively gone on a shortening of the fusion of the two vessels and an atrophy of the azygos terminal, until finally we find the human type of this vessel, the anterior communicating artery, and the two lateral cerebrals. The fact that we occasionally meet with a median vessel, with fusion of


Bifurcation and left paracentral branch of M. A. C. A.


Branch to left Precuneus



Median Ant. Cerebral A.


Right Ant. Cer. A.


Left Ant. Cer. A.


Fig. 4. Photograph of a dissection showing the two lateral anterior cerebral arteries and a large median anterior cerebral artery with a typical distribution.


the two anterior cerebrals, and other anomalies of this set of vessels, I think may be accounted for on the principle of reversions and anomalies so frequently found among the insane.

The median anterior cerebral artery has been found in all forms of mental diseases, possibly a little more frequently in forms of dementia. In 400 cases of mental disease, 42 instances of the abnormal vessel were found; 11 were in senile dementia, 7 in chronic


192 I. W. BLACKBURN

dementia, 5 in epileptic dementia, 5 in arteriosclerotic dementia, 4 in paresis, 3 in chronic melancholia, 2 in dementia praecox, and 1 case each in manic-depressive insanity, organic dementia, acute delirious mania, acute mania, and idiocy. Twenty-five of these cases were in white males, 11 in colored males, 4 in colored females and 2 in white females. The disproportion between the cases examined of each color and sex makes this of relatively little value.

T do not see that there can be any special relationship between the mental disease per se and the presence of this anomalous vessel, but on the supervention of arterial diseases some very interesting cerebral conditions might be caused. In most instances, cerebral arterial anomalies are compensated for before the develop


Fig. o. Diagram of the Circle of Willis with a typical median anterior cerebral artery.


ment of the psychosis, the anomaly being then only indicative of ill development.

In these days of bold operations upon the vessels and even on structures at the base of the brain, it might be well for the surgeon to bear in mind the existence of these anomalies. In fact the main object of the paper • of Grunbaum and Sherrington, quoted from above, was to call attention to the surgical relations of the inequalities of cross anastomosis of the circulation in anomalies of this anterior arterial system. Illustrative of these dangers may be mentioned one of my own cases in which thrombosis of the intracranial portion of the carotid artery resulted in death of the


MEDIAN ANTERIOR CEREBRAL ARTERY


193


whole left hemisphere on account of small size and sclerotic obstruction of the anterior communicating artery and the left posterior cerebral artery.



Fig. 6. Photograph of the base of the brain .showing the origin of a large median anterior cerebral artery.


The conclusions which I think we may reach with certainty are as follows:

1. There is a gradual development of the lower mammalian form of azygos anterior cerebral artery into the perfect Circulus Willisii of the human type.

2. The frequent variations of this system of vessels suggest instability of ontogeny, and of phylogeny.

3. The anomalies of these arteries together with the presence


194 I. W. BLACKBURN

of the median anterior cerebral artery are frequent among the insane.

4. That in keeping with other anomalies of the brain among the insane it is probable that many of these variations are in the direction of reversion to the primitive type, and that it is reasonable that the frequently found arteria termatica is one of these reversions or survivals such as are not uncommon in other organs.


Accepted by the Wtstar Institute of Anatomy and Biology, March 24, 1010. Printed July 7, 1910.


CONTRIBUTIONS FROM THE ZOOLOGICAL LABORATORY OF THE. MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE. E. L. MARK, Director. No. 20S.


DEGENERATION IN THE GANGLION CELLS OF THE CRAYFISH CAMBARUS BARTONII GIR.

HANSFORD MacCURDY

WITH NINE FIGURES

INTRODUCTION

Much of our knowledge of the changes occurring in the central roots of nerves and their ganglia after the nerve trunks have been severed has accumulated during the last two decades. In the earlier observations and experiments, attention was directed chiefly to the nerve roots and their related ganglia. From various causes, including the complex character of the nerve centers, the earlier investigations did not include the ganglion cells. Only comparatively recently has attention been directed specially to the ganglion cells and the changes occurring in them.

It was suggested to me by Professor G. H. Parker that the large nerve cells of the abdominal ganglia of the eastern crayfish, Cambarus bartonii, would afford favorable material for the study of the changes in the ganglion cells after their nerve fibers had been severed. Inasmuch as the investigations hitherto reported have been on the nerve cells of vertebrates, the additional purpose of extending our knowledge to an invertebrated animal would also be served.

As it is well known, Waller held that only those parts of the nerve fibers degenerate which are separated from their nerve centers. For many years this view prevailed. In time, however, from observations made in cases where limbs had been amputated which revealed an altered condition of the nerve roots and their ganglia, doubts arose as to whether Waller was correct in limiting the changes to the peripheral parts of the affected nerves. The


196 HANSFORD MACCURDY

observations of Friedlander und Krause ('86) showed that alterations in the roots of the severed nerve were considerable. In long-standing cases of amputations, these authors reported alterations of medullated fibers, and a reduction in the size of the bundles of fibers. These observations led to experiments on animals for more direct evidence on the questions involved. Homen ('90) performed experiments by cutting nerves in a number of dogs and found that the cells of the related ganglia were much reduced in size. He attributed this reduction to atrophy of the ganglion cells and their nerve fibers. Krause ('87) expressed the opinion that true retrograde degeneration is identical with Wallerian degeneration. Marinesco ('92) found that on cutting the spinal nerves marked changes occurred in the dorso-lateral group of cells in the spinal cord, and stated that the number of these cells was much less in animals operated upon than in the case of normal animals. He also claimed that the remaining cells showed more or less atrophy. In experiments in which the ganglion nodosum was concerned, he (Marinesco, '98) described a period of restoration after the period of depression.

Fleming ('97), operating on dogs and rabbits, measured the nuclei of the affected nerve cells and at the end of 12 days found them to be smaller than the nuclei of normal nerve cells. At the end of eighteen weeks, he found still greater differences in the size of the affected and normal nuclei and the noted some atrophy and the disappearance of some cells. He also described differences in the size, position, and arrangement of the chromatic elements in the cells. In experiments on rabbits. Van Gehuchten ('97) found the majority of the ganglion cells of the ganglion nodosum degenerated. Similar results were obtained from the dog by Kosaka und Yagita ('05). Koster ('03) gave an account of a series of experiments on cats, dogs and rabbits in which the sciatic nerve was cut near its exit from the vertebral canal. He described differences in the tigroid bodies of the spinal ganglion cells. In some of the cells a partial restoration occurred, while in others complete degeneration was found. A modification of the tigroid substance of the cell protoplasm was found to take place four to six days after the operation, while degeneration occurred only


GANGLION CELLS OF THE CRAYFISH 197

after a much longer interval of time. Kleist ('04) used in his experiments half-grown cats and rabbits and found degeneration of the spinal ganglion cells of the upper cervical and lower thoracic nerves which had been severed. He further described some of the cells which did not degenerate as having undergone distinct atrophy, while others, he concluded, returned to a normal condition, since he found more cells at the end of four months than were seen at the end of ten days. Van Gehuchten ('03) found degeneration taking place centrally and declared it to be a true degeneration, i.e., the same as that which takes place peripherally. He also pointed out the importance of this process in tracing the course of nerves.

Ranson ('06) operated on white rats by cutting the second cervical nerve and found simple atrophy and true degeneration in the ventral and dorsal roots, the spinal nerve ganglia, and the spinal cord. An apparently variable number of ventral-horn cells disappear as a result of degeneration, likewise a considerable and constant number of spinal ganglion cells. To avoid septic infection of the parts studied, Ranson used methods which should satisfy every requirement.

It is seen that in a majority of cases changes are said to occur in the central parts of the nervous system after nerves have been severed. These changes consist in a reduction in the size of the fibers and the ganglia, and, in fewer cases, in the disappearance of the nerve fibers and nerve cells. Perhaps the mos,t convincingevidence of the disappearance of the ganglion cells is that found in the results of Ranson in the disappearance of one-half of the nerve cells in the spinal ganglion after cutting the second cervical nerve in the white rat.

EXPERIMENTAL ON THE FIFTH ABDOMINAL GANGLION OF THE

CRAYFISH

Retzius ('90) long ago described the structure of the abdominal ganglia of Astacus fluviatalis. His work renders unnecessary a complete description of these ganglia in Cambarus, the form used

THE JOURNAL OF COMPARATIVE NEUROLOGY AND PSYCHOLOGY, VOL. 20, NO. 3.


198 HANSFORD MACCURDY

in these experiments. The fifth abdominal ganglion was selected as the most suitable for study, because its position rendered it accessible for operative purposes and its structure appeared reasonably simple as to the number and distribution of the large ganglion cells. In this ganglion, in its normal condition, are found a few relatively large ganglion cells whose positions are fairly constant and whose identity can be determined with certainty, at least in most cases. Of these large cells a few lie well forward in the ganglion, near its ventro-lateral surfaces, forming a group on each side of the ventral line. In each of these groups there are a few easily recognized cells larger than the others. These lie in the anterior third of the ganglion. Immediately posterior to each anterior group of cells, though not distinctly separated from them, is another group of cells, which lies for the most part beneath the roots of the large lateral nerves. In this region are three or four nerve cells which are usually the largest found anywhere in the ganglion. A transverse section passing through the roots of the lateral nerves usually cuts through one or more of these large ganglion cells. In the section, these cells lie near the ventral surface of the ganglion and to the right or left. It is thus seen that the identification of the particular cells is reasonably certain. Posterior to the group of cells just described, and closely connected with it by the smaller cells, lie the remaining large ganglion cells which form the posterior part of the ganglion. As in the other parts of the ganglion, the cells of the posterior group vary in diameter, and individual cells may be identified by their position as well as by their size. It is to the large ganglion cells in the positions which have just been pointed out that attention is particularly directed later.

The smaller cells were less suited to the requirements of the experiments than those just described because of the difficulty with which their finer structure could be determined. Of the other abdominal ganglia only the fourth and sixth need be mentioned. While the fourth ganglion was not studied as fully as the fifth, it was clear that the large cells were arranged in it in much the same way as in the fifth, and that the disappearance of any cell or cells from this ganglion could be easily recognized. The sixth


GANGLION CELLS OF THE CRAYFISH 199

ganglion is larger than the other two and contains a greater number of ganglion cells, whose arrangement was different from that in the other ganglia. No attempt was made to determine any particular arrangement of the cells in this ganglion. On this ganglion the general effects of the operation were observed as well as the degeneration and disappearance of its cells without respect to their particular position.

In all these ganglia, each large ganglion cell has a single nerve fiber proceeding from it, which may be traced in favorable sections far enough to determine its course some distance through the ganglion. Retzius ('90) has shown that the fibers from some of these cells pass into the connectives and from others into the lateral nerves. These fibers are non-medullated, having only the sheath of Schwann. The fibers also show very characteristic nerve fibrillse. These fibrillse are best seen in the large fibers when special methods of fixation are employed. They may also be seen very distinctly in the axis-cylinder within the ganglion cells, where they extend partly around the cell nucleus, though separated from it by a certain amount of the protoplasm of the cell. It is evident from what has been said that these large ganglion cells with their fibers extending outward in the manner described, afford relatively simple conditions for experiments, in which the effects on the ganglion cells following the cutting of their fibers could be readily observed.

It was not known how well the animals could endure the injuries incident to the necessary operations, such as the effects due to shock and the interference with the ventral blood sinus. Throughout the series of experiments, however, little difficulty was experienced from either of these sources. There was but little loss of blood, shock effects soon passed away, and the wounds healed with greater promptness than was anticipated. Aside from the loss of movement of the fifth abdominal segment and those posterior to it, the animals operated upon differed from normal individuals only in that they were slightly less active.

In operating, an incision was made with a sharp lance through the integument well toward the right side of the ventral surface of the abdomen between the fourth and fifth stern ites. By insert


200 HANSFORD MacCURDY

ing the lance, or a fine-pointed pair of scissors, the connective was severed about midway between the fourth and the fifth ganglion. Through the same incision the instrument was passed backwards in a line nearly parallel to the long axis of the body, severing the lateral nerves on the right side of the fifth ganglion. Another incision similar to the first was made to the left of the median line between the fifth and sixth sternites and through it the connective between the fifth and sixth ganglion and the left lateral nerves of the fifth ganglion were severed. A thin coat of celloidin was then applied to the wounds. Thus the connectives, both in front of and behind the fifth ganglion, and the lateral nerves on each side of it were cut with the least possible amount of injury to the tissues, and the ganglion was thus isolated as regards its nerve connections. The animals were then numbered for purposes of identification, placed in an aquarium, and propery cared for until they were taken for study. Only those animlals in which the wounds healed readily and which showed no evidence of infection, etc., were selected for final preparation and study.

At desired intervals after the operation, individual animals were killed, and the abdominal ganglia from the third to the sixth inclusive were removed together by cutting along the ventral surface through the integument and sternites on each side of the median line, and carefully removing the ventral wall of the body with the nerves and ganglia in situ. The entire piece was kept straight by attaching it to a glass rod and in this position it was immersed at once in the killing fluid. Before clearing, the nerve elements were carefully removed from their natural position on the body wall and transferred together through absolute alcohol and xylol and imbedded in paraffin.

Since the cells of the isolated and the normal ganglia were finally to be carefully compared, it was necessary to give them, as nearly as possible, the same treatment. To secure this equality, one normal individual and one individual which had been operated on, were killed and prepared together in the manner described and given parallel treatment throughout.

Two methods of treatment were used. One series was prepared according to the Nissl method for staining the tigroid substance,


GANGLION CELLS OF THE CRAYFISH 201

and another was treated with vom Rath's picro-aceto-platinoosmic fluid.

In using the Nissl method, the ganglia were put for forty-eight hours in 95% alcohol, then dehydrated, cleared in xylol, imbedded in paraffin, cut in sections 10m. thick, and stained in toluidine blue. In order to secure perfectly parallel treatment of the cells in the staining, washing, and subsequent dehydration, serial sections of the isolated ganglion of a normal individual and of an individual operated upon were placed in alternate rows on the same slide. Preparations were made in this way from materials killed at intervals of from two to five days covering a period in all of thirty-three days after the operation.

The large nerve cells of the normal ganglia prepared by the Nissl method presented a characteristic appearance (fig. 1). Fibrillse in the axis cylinder were usually visible, though not all the cells revealed them. That they were not always seen, is believed to be due to this method, which, as is well known, is not a wholly satisfactory one for demonstrating fibrillse. The Nissl flakes" were present, though somewhat less distinct than in the ganglion cells of vertebrated animals. In some cases there appeared, more or less distinctly, small centers, which stained deeply and from which radiated irregularly threads of protoplasm. These threads or strands connected with other similar centers or faded out in the surrounding cytoplasm. These small centers or granules with the network of radiating threads are confined almost wholly to regions adjoining the axis-cylinder area. The nucleus exhibited a full rounded form and possessed a reticular structure with a very distinct and deeply stained nucleolus. Occasionally two nucleoli were found in one nucleus. Sometimes a slight shrinking of the cytoplasm next the nucleus was observed.

With the state of the normal ganglion cells, are to be contrasted the conditions found in the corresponding ganglion cells from a fifth ganglion (fig. 2) which had been nervously isolated about twelve days, but otherwise had received identical treatment. The cells from the isolated ganglia showed distinct alterations. The nerve fibrillse in the axis-cylinder area had disappeared and the Nissl flakes had become finely granular or had disappeared.


202 HANSFORD MACCURDY

The reticulate structure was lost, and the cytoplasm had shrivelled away from the outer parts of the cell. The nucleus also shared in these changes. The nuclear membrane had suffered a complete collapse, while the nuclear network had disappeared. The nucleolus apparently was more resistant than the other parts, and only after longer periods than that shown in figure 2 did it finally fail to stain and ultimately disappeared.

Degeneration may be practically complete at the end of two weeks, or it may be prolonged through a much greater period. In addition to the changes in the large ganglion cells, man}^ of the smaller cells undergo corresponding changes and disappear. The ganglion is much reduced in size because of these changes. Preparations made three to five days after the operation usually showed but little change. After seven to fourteen days the changes were quite evident . In some cases cells had lost their distinguishable structures eighteen days after the nerves had been severed, while other cells in the same ganglion were apparently only slightly affected. The disappearance of individual cells in the midst of cells which still remained apparently little affected is regarded as strong evidence of true degeneration.

On examining the results obtained in the first series of experiments, it became evident that, although degeneration could be demonstrated, the Nissl method had revealed only a part of the changes which had occurred within the cells. Some method which would show the changes in the finer structure of the cells and one which would serve as a check on the previous experiments was, therefore, sought. Since the picro-aceto-platino-osmic fluid of vom Rath is known to demonstrate admirably the finer struct jres of these ganglion cells in their normal condition, it was thought that this treatment might also show the changes occurring in connection with the loss of function and consequent degeneration. Accordingly the ganglia were removed, in the manner described for the Nissl method, from an individual in which the fifth ganglion had been isolated twenty-nine days, and were immersed in vom Rath's fluid for a period of forty-eight hours, washed in water, dehydrated, cleared in xylol, imbedded in paraffine and cut into sections 10/j. thick. Examination of the material revealed nerve


GANGLION CELLS OF THE CRAYFISH 203

cells in various stages of disintegration in the fourth, fifth and sixth ganglia. This unmistakably demonstrated the usefulness of this method for detecting degenerating cells. In the fourth ganglion, certain large nerve cells were almost wholly disintegrated, while adjoining nerve cells were apparently normal. This is considered significant, since this ganglion had only the connective between it and the fifth severed. A similar condition was found in the sixth ganglion, which had only its anterior nerve connections severed. The fifth ganglion showed more extensive changes, which will be referred to later.

By this method preparations were made of material in situ at intervals of three or four days over a period of thirty-eight days from the time of operation. Corresponding normal ganglia were given treatment at the same time and mounted in parallel series with the isolated ganglia in a majority of the experiments. In later experiments parallel series were deemed unnecessary because of uniformity in results.

Were it not needed for comparison, it would be unnecessary to give even a brief description of the normal ganglion cell prepared according to the vom Rath method. The action of the fluid on the nerve cells is such that the various cell structures are well differentiated in black and white. The fibrillae (fig. 3) are very clearly seen in the axis cylinder within the cell as very fine dark lines. They are frequently also seen in the protoplasm of the cell surrounding the nucleus. Numerous flake-like bodies, apparently having some relation to the Nissl bodies, lie in the cytoplasm, arranged concentrically around the nucleus. This concentric arrangement of parts around the nucleus is a condition which will be referred to in a future paragraph. The nucleus shows the characteristic nuclear membrane, network, and nucleolus.

In ganglia which had been isolated five to seven days, some of the large ganglion cells had begun to show alterations in their structure. The large cells in the posterior part of the ganglion were among the first to become affected. Some of the smaller cells in this part showed on-coming changes somewhat earlier. In the large cells the fibrillae (fig. 4) became somewhat nodular and tortuous in their course. The limits of the axis-cylinder area in


204 HANSFORD MacCURDY


EXPLANATION OF FIGURES

All figures were drawn with the aid of the camera lucida and are magnified 375 diameters. Figures 1 and 2 are from preparations made by the Nissl method; the others are from preparations made by the vom Rath method; all are from the fifth abdominal ganglion of Cambarus bartonii.

1. Large ganglion cell, normal; Nissl method.

2. Large ganglion cell from ganglion isolated twelve days; nucleus collapsing, axis-cylinder area reduced, fibrillse indistinct.

3. Large ganglion cell showing normal condition of axis cylinder, fibrilla?, nucleus, and "flakes;" vom Rath method.

4. Large ganglion cell from ganglion isolated twelve days; tortuous and nodular fibrillse. Compare with figs. 2 and 3.

5. Ganglion cell from ganglion isolated fifteen days; nucleus and axis cylinder reduced, fibrillse breaking up, cytoplasm granular.

6. Ganglion cell from ganglion isolated twenty-six days; nucleus collapsed, fibrillae absent, cytoplasm granular.

7. Ganglion cell from ganglia isolated twenty-eight days ; nucleus and nucleolus disappearing, cytoplasm granular.

8. Ganglion cell from ganglion isolated thirty days; nucleus and nucleolus disintegrating (seen in only a few cells), cytoplasm faintly granular.

9. Ganglion cell from ganglion isolated twenty-nine days; nucleolus has disappeared; cell-contents faintly granular and apparently very fluid.


GANGLION CELLS OF THE CRAYFISH Hansford MacCurdy


PLATE I




fi#


¥


,












m









THE JOURNAL OF COMPARATIVE NEUROLOGY AND PSYCHOLOGY— VOL. 20 NO. 3


GANGLION CELLS OF THE CRAYFISH 205

the cells were less clearly denned, and the normal concentric arrangement of the parts about the nucleus was more or less disturbed. Usually in twelve to fourteen days after the operation these changes became very evident. At this stage the fibrillae had become more nodular and tortuous in appearance and the axis-cylinder area was usually lost in the increased disturbance in the arrangement of the cytoplasm of the cells. In addition to these alterations in the organization, a marked change was seen in the results of the action of the vom Rath fluid on the various parts of the cells, in that they did not become black, as in the case of normal cells, but assumed a somewhat yellowish color, instead. This chromatolysis is one of the most characteristic alterations observed in these cells and indicates chemical changes in their constitution. This appeared early and increased as degeneration advanced. This change is undoubtedly the chromatolysis demonstrated by the Nissl method and observed in the spinal-ganglion cells of dogs by Lugaro ('87), and in cats, dogs and rabbits by Koster ('03). In eighteen to twenty days the fibrillae and axis cylinder usually had entirely disappeared (figs. 5, 6), and the concentric arrangement about the nucleus had entirely broken up. The flake-like bodies in the cytoplasm had given place to fine granules. The nucleus (fig. 6) was usually collapsed and the nucleolus had also undergone chromatolysis. The later stages (fig. 7) , which were usually found at eighteen to thirty days after the operation, were marked by the gradual loss of the various parts of the cell. The fine granules visible earlier in the cytoplasm disappeared, the nucleoplasm shrank, and the nuclear membrane disintegrated. The nucleolus remained slightly longer than the other organs of the cell and was seen sometimes to disintegrate (fig. 8) and sometimes to grow fainter and fainter until it became invisible (fig. 9). The order of these events rarely varied.

The number of cells in the fifth ganglia which were affected varied in different cases. This depended in part on the time that had elapsed since the operation, and no doubt in part on slightly varying conditions of the state of the animal. In no case did all the ganglion cells in any ganglion disappear, and in no case did degeneration fail to occur in a considerable number of cells within


206 HANSFORD MACCTJRDY

the limits of the experiments. Three individuals in which the fifth ganglion had been isolated were not killed until ninety days after the operation. These three presented a condition of the ganglia and ganglion cells very similar to that found in ganglia at the end of thirty-nine days. The parts , of the ganglion which had undergone degeneration were somewhat more clearly distinguished from the remaining portion. There was some evidence that partial regeneration had occurred in these ganglia, but this was not conclusive.

THE FOURTH AND THE SIXTH ABDOMINAL GANGLIA

In these experiments, as already stated, the connective between the fourth and fifth abdominal ganglia was severed a few millimeters posterior to the fourth ganglion ; some cells in this ganglion might, therefore, be expected to degenerate. In all cases the fourth ganglion was prepared in the same manner as the fifth and this included the posterior end of the connective. An examination of these ganglia revealed degenerating cells in their anterior parts on both right and left sides, and a few cells in their middle portions on each side near the ventral surface. These were most clearly seen twenty to thirty days after the connective had been severed. In these cases the nerve cells in the posterior part of the ganglion were but little influenced. In certain cases of longer standing some of these posterior cells suffered partial degeneration.

The sixth abdominal ganglion was prepared in the same way as the fourth and fifth. Degeneration of a considerable number of nerve cells in the anterior parts of this ganglion was found in all cases twenty or more days after the connective had been severed. A smaller number of nerve cells in a similar condition of disintegration were found near the ventral surface in the sections through the middle and posterior portions of the ganglion. The beginning of degeneration in this ganglion was found in preparations made twelve to fourteen days after the operation.

The degenerating ends of the connectives were included in the preparations and mounted in series with the related ganglion.


GANGLION CELLS OF THE CRAYFISH 207

The nerves showed a more advanced state of degeneration at and near the cut surfaces than was found at points more remote from these surfaces.

DISCUSSION

It has been found in the experiments described in this paper that the ganglion cells in the fourth, fifth and sixth abdominal ganglia of Cambarus bartonii undergo distinct structural changes when the connectives and lateral nerves are severed. A larger number of cells is affected in the fifth ganglion than in the others ; this ganglion, from the nature of the operation, had more of its nerve fibers cut than either of the other two. That these changes were due to the cutting of the nerve is beyond reasonable doubt , since they were found only in those ganglia which had their connectives severed, and were found in every such case. That the changes are those connected with true degeneration finds support in the fact that they resemble in their essen t ial points the histological changes in the corresponding parts of the nerve fibers and cells of vertebrated animals as described by Monkeberg und Bethe ('99), Kleist ('04), and others. In all cases chromatolysis is reported to have occurred soon after the nerves were severed, which indicates important chemical changes in the nerve cells. Following this were characteristic alterations of the nerve fibrillae and other structures of the cell, resulting finally in the 'destruction of the cell itself. Kleist ('04) stated that two to six days after the sciatic nerve in dogs, cats and rabbits had been severed, alterations occurred in the tigroid bodies of the spinal ganglion cells. In ten days vacuolar degeneration and shrinking of the cells had occurred. It has been shown in this paper that alterations in the tigroid bodies of the ganglion cells of Cambarus were found five to seven days after the connectives had been cut, and that in ten to twelve days the fibrillae were distinctly altered and a shrinking of the cytoplasm and nucleus had occurred.

In regard to the time required for complete degeneration to take place, experimental results vary widely. Fleming ('97) observed disintegrating nerve cells in six weeks, but many more in eighteen weeks. Koster ('03) found that only relatively few cells had de


208 HANSFORD MacCURDY

generated in three months, but a large number had disappeared after nine months. Ranson ('06) observed no further change after two months, and for this reason concluded that degeneration was not progressive. In the present series of experiments some cells had lost all distinguishable structures in twenty-one days, and a greater number was found in this condition twenty-eight to thirtynine days after the injury. It is, therefore, reasonable to believe that degeneration takes place slowly in some cases and more rapidly in others. These differences are best explained by the differences in the animals used and in the nature and the conditions of the experiments.

The results of this series of experiments furnish some additional support to the view that the continued life of the neurone depends upon the performance of the normal functions of all its parts.


SUMMARY

1. Observable structural changes occur in many of the ganglion cells of the fifth abdominal ganglia of Cambarus three to seven days after the connectives are severed between the fourth and fifth, and the fifth and sixth ganglia.

2. A smaller number of ganglion cells in the fourth and sixth ganglia likewise degenerate.

3. Complete degeneration of many of the cells occurred twenty eight to thirty-nine days after the nerve fibers were severed. The length of time required for degeneration varies.

4. The histological changes accompanying degeneration in these nerve fibers and nerve cells are similar to those which have been described by others for the nerve fibers and nerve cells of vertebrates, excepting those w T hich pertain to the medullary sheath, which is absent in Cambarus.


Accepted by the Wistar Institute of Anatom3 - and Biology, April 17, 1910. Printed July S, 1910.


GANGLION CELLS OF THE CRAYFISH 209


BIBLIOGRAPHY

Fleming, R. A. The effect of "ascending degeneration" on the nerve cells of

1897 the ganglia, on the posterior nerve roots and the anterior cornua of the cord. Edinburgh Med. Jour., n.s., vol. 1, pp. 174-182, 279-288, pi. 11.

Friedlander, C, und Kraxjse, F. Ueber Veranderungen der Nerven und des

1886 Ruckenmarks nach Amputationen. Fortschr. d. Med., Bd. 4, pp. 749-764, Taf. 7.

Homen, E. A. Veranderungen des Nervensystems nach Amputationen. Beitrdge 1890 z. path. Anat. u. allg. Path., Bd. 8, pp. 304-351, Taf. 18-19.

Kleist, K. Experimentell-anatomische Untersuchungen iiber die Beziehungen

1904 der hinteren Riickenmarkswurzeln zu den Spinalganglien. Arch, f. path. Anat. u. Physiol., Bd. 175, pp. 382-407, Taf. 9.

Kosaka. K., und Yagita, K. Experimentelle Untersuchungen iiber den Ursprung

1905 des N. Vagus. (Quoted from Ranson 1906.)

Koster, G. Ueber die verschiedene biologische Werthigkeit der hinteren Wurzel 1903 und der sensiblen peripheren Nerven. Neurol. Centralbl, Bd. 22,

pp. 1093-1102.

Krause, F. Ueber aufsteigende und absteigende Nervendegenerationen. Arch.

1887 /. Anat. u. Physiol, Physiol. Abt., Jahrg. 1887, pp. 370-376.

Ltjgaro, E. Sulle alterazioni delle cellule nervose dei gangli spinali. Rivista di 1887 pathol. nerv. e. ment., 1896, no. 12, p. 457.

Marinesco, G. Ueber die durch Amputation hervorgerufenen Veranderungen 1892 der Nerven und des Ruckenmarks. Berliner klin. Wochensc'r.,

Bd. 39, p. 988.

1898 Sur les phenomenes de separation dans les centres nerveux apres la section des nerfs peripheriques. Presse med., annee, 6, semestre 2, pp. 201-206.

Monkeberg, G., und Bethe, A. Die Degeneration der markhaltigen Nerven 1899 fasern der Wirbelthiere unter hauptsachlicher Beriicksichtigung des Verhaltens der Primitivfibrillen. Arch. f. mikr. Anat., Bd. 54, pp. 135-183, Taf. 8-9.

Ranson, S. W. Retrograde degeneration in the spinal nerves. Jour. Comp.

1906 Neurol, and Psychol, vol. 16, no. 4, pp. 265-293.


210 HANSFORD MACCURDY

Retzius, G. Zur Kenntniss des Nervensystems der Crustaceen. Biol. Unters.,

1890 X. I'.. Bd. 1. pp. 1 50. Taf. 1-14.

Van Gehtjchten, A. L'anatomie tine de la cellule nerveuse. La Cellule, tome.

is'..: 13, pp. 312-390, pi. 1.

1903 La (lefieueresccnce dite retrograde OU ilefienereseence Wallerienn

Lndirecte. I.> Vivraxe, vol. .">. p. 1. (Quoted from Hanson, 1906.)


The development of the sympathetic nervous system in mammals

Albert Kuntz From the Laboratories of A nimal Biology of the State University of Towa

With Eighteen Figures

CONTENTS

I. Introduction 213

II. Historical, Survey 214

III. Methods of Investigation 218

IV. Observations 219

1 Sympathetic trunks 219

a. Early development 219

h. Cell migration 222

c. Later development 229

d. Nature of migrating cells 230

2 . Prevertebral plexuses 233

a. Develo ment 233

6. Histogenetic relationships 234

3. Vagal sympathetic plexuses 235

a. Introductory 235

b. Myenteric and submucous plexuses 236

c. Pulmonary plexuses 239

c/. Cardiac plexus 239

e. Cell migration along the vagi 241

V . Discussion of Results, and Conclusions 246

(i. Migration of medullary cells 246

h. The neurilemma 248

c. Sympathetic excitatory and sympathetic sensory neurones 249

il. A wider application of Schaper's conception 351

e. Relation of the sympathetic to the central nervous system 251

/'. Functional relations -^

V.'L Summary •• 253

Bi bliography 256


212 ALBERT KUNTZ

• I. INTRODUCTION

The present investigation of the development of the sympathetic nervous system in mammals was carried on in the laboratories of Animal Biology of the State University of Iowa, under the direction of Prof. Gilbert L. Houser.

Although much excellent work has been done on the development of the sympathetic nervous system, our knowledge concerning the sympathetic neurones and the relation of the sympathetic to the central nervous system is still very meager. Our newer conceptions of nerve-components and of the functional divisions of the peripheral nervous system call for a re-investigation of the development of the sympathetic system m order to bring this division of the nervous system into harmony with established facts.

The present investigation was undertaken in order to further exact knowledge concerning the histogenesis of the sympathetic system, to establish the histogenetic relationships between the sympathetic neurones and the neurones in the central nervous system, and to correlate the sympathetic system with the other functional divisions of the nervous system. The most iniportant results achieved pertain to increased knowledge concerning the histogenesis of the sympathetic system and its relation to the central nervous system, and to the fact that the cardiac plexus and the sympathetic plexuses in the walls of the visceral organs are not derived from the sympathetic trunks, as has hitherto been supposed, but have their origin in nervous elements which migrate from the vagus ganglia and the walls of the hind-bniin along the fibers of the vagi. During the progress of the work, two preliminary papers were published (see Bibliography).

It is a real pleasure to express my deep sense of obligation to Prof. Houser for his many helpful suggestions and for the inspiration afforded by the constant enthusiastic interest manifested by him during the progress of this investigation. I desire also to express my indebtedness to Dr. F. A. Stromsten for many valuable suggestions in technique.


SYMPATHETIC SYSTEM IN MAMMALS . 213

II. HISTORICAL SURVEY

The earliest observations on development of the sympathetic nervous system are those of Remak ('47). That pioneer among the investigators of the sympathetic system described the anlagen of the sympathetic trunks in the chick as ganglionic enlargements on the communicating rami, situated at their point of deviation from the spinal nerves. He believed that the cells composing these ganglionic enlargements are derived from preformed elements arising in the mesoderm. This view of the mesodermal origin of the sympathetic nervous system held undisputed sway for more than two decades, and has found advocates, among whom may be mentioned Paterson ('91), in more recent times.

The work of Balfour ('77) marks the beginning of our modern conception of the ectodermal origin of the sympathetic nervous system. According to his observations on the selachians, the anlagen of the sympathetic trunks arise as simple enlargements on the spinal nerve-trunks. Subsequently, these enlargements advance toward the aorta, each, however, retaining connection with its respective nerve by a fibrous branch which becomes the communicating ramus. These ganglionic enlargements are at first independent of each other, but become united later by longitudinal commissures. These observations on the selachians were substantiated by Onodi ('86), Van Wijhe ('89), and Hoffmann ('99).

Schenck and Birdsall (78) extended the conception of Balfour, somewhat modified, to the higher vertebrates. Tracing the development of the sympathetic trunks in birds and mammals, they found that before the anlagen of the sympathetic trunks appear, the spinal ganglia are not sharply limited distally. Groups of cells become detached from the distal ends of the spinal ganglia and advance far into the spinal nerve-trunks. These cells, they believe, constitute the anlagen of the sympathetic trunks, but they have given no clear conception of the process by which these cells are transferred from the spinal ganglia to their new location in the sympathetic anlagen.

Kolliker ('97) adopted the doctrine of Balfour and attempted to extend it to the peripheral sympathetic plexuses. In the ab THE JOURNAL OF COMPARATIVE NEUROLOGY AND PSYCHOLOGY, VOL. 20, NO. 3


214 ALBERT KUNTZ

sense of confirmatory evidence, he set up the hypothesis that the peripheral sympathetic plexuses arise as cellular offshoots from the cerebro-spinal ganglia.

Onodi ('86) finally established the cerebro-spinal origin of the sympathetic trunks and the prevertebral plexuses for all vertebrates. He believed that the cells at the distal ends of the spinal ganglia are forced to advance farther peripherally by the pressure exerted by the newly formed elements back of them. He could not, however, derive the sympathetic trunks and the peripheral sympathetic plexuses from the same source because he found no cellular connections between these two complexes. He believed it necessary, therefore, to cling to the doctrine of Remak with regard to the peripheral sympathetic plexuses, and derive them from the mesoderm.

His ('90) introduced a new factor in the development of the sympathetic nervous system. In a human embryo 6.9 mm. in length, he observed cells migrating from the spinal ganglia. These he described as germinal cells which break through the motor roots of the spinal nerves and migrate in swarms toward the future location of the sympathetic trunks.

Pushing on the way indicated by his father, His, Jr., ('91) traced the origin of the entire sympathetic system to the spinal ganglia. He described cell-swarms in the chick similar to those described by His, Sr., in the human embryo. These cell-swarms become detached from the spinal ganglia, break through the motor roots of the spinal nerves, and migrate toward the dorso-lateral surfaces of the aorta, where they become aggregated into cellgroups which constitute the anlagen of the sympathetic trunks.

From these aggregates, cells proceed round the aorta until the latter is surrounded ventrally by a complete ring of sympathetic cells. This ring gives rise to new cell-swarms which migrate farther peripheral^ and become the anlagen of the peripheral sympathetic plexuses, including the sympathetic plexuses in the walls of the digestive tube and the sympathetic components related to the vagi.

In his later researches on embryos of the chick, His, Jr., ('97) found that the earliest anlagen of the sympathetic system arise about the beginning of the fourth day of incubation as a pair of longitudinal cell-columns lying along the sides of the dorsal sur


SYMPATHETIC SYSTEM IN MAMMALS 215

face of the aorta. These are the beginnings of the primary sympathetic trunks. About the beginning of the sixth day the anlagen of the secondary, or permanent, sympathetic trunks arise as a series of cell-aggregates situated just median to the ventral roots of the spinal nerves. The cells giving rise to the primary sympathetic trunks migrate thither from the spinal ganglia, along the spinal nerves and the communicating rami. The anlagen of the secondary sympathetic trunks are separated from the spinal ganglia only by the fibers of the ventral nerve-roots. Neuroblasts may be found in the ventral nerve-roots, caught apparently in migration from the spinal to the sympathetic ganglia. After the appearance of the secondary trunks, the primary sympathetic trunks become resolved into the various ganglia and nerves of the prevertebral and the peripheral sympathetic plexuses. This view was adopted by Lillie ('08).

Marshall ('93) found that the anlagen of the sympathetic trunks in embryos of the frog and the chick arise "as a series of outgrowths from certain of the cranial and all of the spinal nerves." These outgrowths develop into ganglionic enlargements which become connected later by longitudinal commissures. These findings agree essentially with Balfour's observations on the selachians, but differ very materially from the findings of later observers for amphibians and birds.

In his later researches on the urodeles, Hoffmann ('02) found conditions of development differing widely from those in selachians. In this type the anlagen of the sympathetic trunks arise as scattered cells along the sides of the dorsal surface of the aorta, some of which are connected with the communicating rami by slender protoplasmic processes.

In his work on the common toad, Jones ('05) pointed out a notable difference in the development of the anterior and the posterior regions of the sympathetic trunks. In the region anterior to the second spinal nerve, they arise from cells scattered in the mesoderm. This agrees essentially with the findings of Hoffmann. In the region posterior to the second spinal nerve, ridges of cells appear along the sides of the aorta. The cells at the tops of these ridges become differentiated to form the sympathetic trunks. These findings have not been substantiated by other observers.


216 ALBERT KUNTZ

Kohn ('05, '07) describes the anlagen of the sympathetic trunks in the rabbit as a pair of columns of cell-aggregates arising along the sides of the dorsal surface of the aorta. Similar cells are found in intermediate positions between these cell-aggregates and the spinal nerves, in the paths later occupied by the fibers of the communicating rami. According to Kohn, the sympathetic anlagen are composed of cells which arise by the division of elements which have not migrated thither, but were differentiated in situ in the spinal nerves. Embryonal neurocytes deviate from the course of the spinal nerves toward the aorta. By division they yield a syncytial cellular communicating ramus which extends toward the aorta. Cell-groups become separated from its distal end and give rise to the cell-aggregates of the sympathetic anlagen.

According to Neumayer's observations on embryos of Lacerta (spec?) and the chick ('06), the anlagen of the sympathetic trunks arise as short cellular outgrowths on the spinal nerves which early develop ganglionic enlargements at their distal ends, which become united later by longitudinal cellular commissures. Neumayer, like Kohn, traces the origin of the sympathetic system directly to the spinal nerves. He is of the opinion that in all vertebrates the sympathetic anlagen arise from cells which are to be regarded as the offspring of the dorsal and the ventral nerve-roots and are differentiated in situ, like the cells of the spinal ganglia and the fibers of the nerve-roots.

The work of Froriep ('07) on embryos of Torpedo and of the rabbit, marks a decided advance in our knowledge of the histogenesis of the sympathetic nervous system. He succeeded in tracing medullary cells peripherally along the ventral roots of the spinal nerves. These cells he formerly interpreted as elements which give rise to the neurilemma. After Harrison C04) showed experimentally that in amphibians the cells giving rise to the neurilemma of both the sensory and the motor fibers have their origin in the neural crest, Froriep concluded that the cells migrating peripherally in the ventral nerve-roots, either alone or with cells which wander out from the spinal ganglia, give rise to the sympathetic nervous system. In his summary he expresses the opin


SYMPATHETIC SYSTEM IN MAMMALS 217

ion that all the sympathetic neurones in the sympathetic trunks as well as in the prevertebral and the peripheral sympathetic plexuses have their origin in the ventral half of the neural tube.

Held ('09) and Marcus ('09) have recently taken exception to Froriep's conclusions. Held has attempted to show, for the entire vertebrate series, that the cells present in the motor nerve-roots play no part in the development of the sympathetic system. He still regards the sympathetic system as an offshoot from the spinal ganglia. Marcus has attempted to show that the cell-groups which Froriep observed in the ventral roots of the spinal nerves do not wander out from the neural tube, but migrate thither from the neural crest. In early stages of embryos of Torpedo, he has observed cell-chains connecting the neural crest with the cellaggregates in the ventral nerve-roots. He concludes, therefore, that the neural crest represents the sole source of all the cells giving rise to sympathetic neurones.

This brief review of the literature has shown that the advocates of the theory of the ectodermal origin of the sympathetic nervous system agree in tracing the origin of the cells giving rise to the sympathetic anlagen to the cerebro-spinal system. There is a wide difference of opinion, however, concerning the immediate source and the histogenesis of these cells.

Two views have been prevalent among the older investigators. Onodi advanced the idea that the cells at the distal ends of the spinal ganglia are forced to advance farther peripherally by the pressure exerted by the newly formed elements back of them. In this manner cell-groups become constricted off from the spinal ganglia and give rise to the anlagen of the sympathetic trunks. His, His, Jr., and some of the later writers have traced the origin of the cells giving rise to the sympathetic anlagen to the spinal ganglia, but have accounted for the transfer of these elements from the spinal ganglia to their new location by active migration, either directly through the mesenchyme or along the paths of the spinal nerves and the communicating rami.

The difference between these two views may be accounted for in part by fundamental differences in the morphogenesis of the sympathetic nervous system in the various classes of vertebrates.


218 ALBERT KUNTZ

In the selachians the anlagen of the sympathetic trunks arise as ganglionic enlargements on the spinal nerves (Balfour, Van Wijhe, Hoffmann). In the amphibians fibers are present in the communicating rami before the anlagen of the sympathetic trunks appear. The latter arise along the sides of the dorsal surface of the aorta (Hoffmann, Neumayer). In Lacerta (spec?), a reptilian type, the sympathetic anlagen arise as short cellular outgrowths on the spinal nerves, which early show ganglionic enlargements at their distal ends (Neumayer). In birds the primary sympathetic trunks arise as a pair of cell-columns lying along the sides of the dorsal surface of the aorta. These early give way to the secondary sympathetic trunks which arise as cellaggregates just median to the ventral roots of the spinal nerves (His, Jr. , Lillie) . In mammals the sympathetic trunks arise as a pair of cell-columns lying along the sides of the dorsal surface of the aorta (Paterson, His, Jr., Kohn).

With these morphogenetic differences in mind, it is apparent that the view of Onodi was based primarily on the selachians, while the theory of the active migration of sympathetic cells finds its basis primarily in birds and mammals.

Kohn and Neumayer have rejected both these views. They admit of no active cell migration, but trace the origin of the sympathetic system to elements which arise in situ in the spinal nerves, and account for the multiplication of cells along the paths of the communicating rami and in the sympathetic anlagen by local cell division. Their views, however, are obviously influenced by their conception of the neurone and their allegiance to the theory of local differentiation and the multicellular nature of nerve-fibers.

III. METHODS OF INVESTIGATION

The following observations are based on embryos of the pig. Several embryos of the cat and a goodly number of embryos of the chick were at my disposal and were used for checking results. Embryos of the pig were found to be more desirable than embryos of the cat, because the cells in the former are comparatively larger and appear to be less crowded.


SYMPATHETIC SYSTEM IN MAMMALS 219

Various methods of technique were employed, but the ironha?matoxylin method was found to be most satisfactory. The embryos were fixed in Zenker's fluid, chrom-aceto-formaldehyde, or chrom-oxalic acid. The sections were usually cut a thickness of 10 microns. It was found most satisfactory to over-stain in hematoxylin, then to differentiate in the iron-alum bath until the color had almost disappeared except from the denser tissues, and to counter-stain lightly with orange-G. The degree to which the haematoxylin shall be differentiated in order to obtain the best results must be learned by experience.

The results obtained from one lot of embryos may be of interest to students of special technique. A few young embryos taken from the laboratory collection were sectioned and stained. These had been kept in 10 per cent formaldehyde for some ten years. When stained by the iron-hsematoxylin method, the nuclei of the nerve cells took a much deeper and more solid stain than in fresh material. These preparations were, therefore, found very serviceable in tracing migrant nervous elements.

The silver reduction method was used with good results in the later stages.

IV. OBSERVATIONS I. SYMPATHETIC TRUNKS

(a) Early development. — The earliest traces of the sympathetic trunks appear in the thoracic region of embryos of the pig about 6 mm. in length, as small cell-aggregates lying along the sides of the dorsal surface of the aorta. The spinal nerves in the thoracic region have already become fibrous and extend peripherally to a point a little beyond the dorsal level of the aorta. Fibrous communicating rami are not present as yet, and the sympathetic anlagen are apparently independent of the spinal nerves.

In embryos 7 mm. in length, the anlagen of the sympathetic trunks may be traced throughout the thoracic and the dorsal region. The cell-aggregates have become larger, and because of the strong curvature of the embryo they are brought into such


220


ALBERT KUNTZ


close proximity with each other that the entire anlage of the sympathetic trunk appears as a continuous cell-column. This cell-column is not of uniform diameter, but in transverse sections traces of it are not wanting in any section in the thoracic and the dorsal region. Fibers are not present as yet either in the anlagen of the sympathetic trunks or in the communicating rami. The sympathetic anlagen are essentially cellular. The cells are closely aggregated and many of them present delicate protoplasmic processes, or are included in small syncytia. These structures are not very apparent in transverse sections, but in sagittal or frontal sections the anlagen of the sympathetic trunks present the appearance of a loose-meshed cellular network.



., a 3 supr: rues. ))v.a.


Fig. A. Diagrammatic transverse section of an embryo 12 mm. in length through the suprarenal bodies, X 50.


SYMPATHETIC SYSTEM IN MAMMALS


221



Fig. 1. Section showing the path of the communicating ramus in an embryo 9 in mm. length, X 210. (See reference letters below.)


REFERENCE LETTERS OF FIGURES

(Except Fig. 16)

All of the figures were drawn with the aid of the camera lucida or the projection lantern. A uniform scale of magnification was not adopted, but the scale of diameters of the drawing as reproduced is given in the description of each figure.


a. i.e. -Accompanying indifferent cells.

ao. -Aorta.

b. rec. -Branch of recurrent nerve.

ca.-Carotid artery.

car.n.-Cardiac nerves.

car. b. -Anlagen of cardiac plexus.

c.r. -Communicating ramus.

cm. d.n.r. -Cells migrating into dorsal nerve-root.

c.m.pv. -Cells migrating from sympathetic trunks into prevertebral plexuses.

c.m.v.r. -Cells migrating into ventral nerve-root .

c.m.vag.r. -Cells migrating into vagus rootlets.

(/.//./•.-Dorsal nerve-root.

e.l.m. -External limiting membrane.

f.pv. -Fibers extending into prevertebral plexuses.

g.c. -Germinal cells of His.

■i.c.c.r. -Indifferent cells in communicating ramus.


i.l.m. -Internal limiting membrane. mes.-Mesentery. m.r ./.-Motor root-fibers. w.s.p.-Anlage of myenteric and submucous plexuses, nb. -Neuroblasts. oe. -(Esophagus.

o. rec. n. -Origin of recurrent nerve. /). <7. -Pulmonary artery. pv. a. -Anlagen of prevertebral plexuses. rec. n. -Recurrent nerve. sp.g. -Spinal ganglion. sp.7i. -Spinal nerve. s.r./.-Sensory root-fibers. supr. -Suprarenal bodies. sy. -Anlagen of sympathetic trunks. /.-Trachea, ■fag. -Vagus trunk.*. vag. b. -Branch of vagus nerve. vag. r. -Rootlets of vagus nerve. v.n.r. -Ventral nerve-root.


222


ALBERT KUNTZ


The spinal nerves are composed of bundles of parallel fibers accompanied by numerous cells which/as will be shown presently, are obviously of medullary and ganglionic origin. These cells are present both at the surface of the bundles and among the growing fibers in the interior of the nerve-trunk. They are easily distinguished from the cells of the surrounding mesenchyme by their larger size and the characteristic chromatin structure of their nuclei.

In embryos 9 mm. in length, the anlagen of the sympathetic trunks may be traced from the cervical to the sacral region. The


@ c §


9 9




9 f • K t , * * • #£ § § i ^

4 % + .



XT


Fig. 2. Section showing the path of the communicating ramus in an embryo 10 mm. in length, X 200.


cells have become more numerous and more closely aggregated. Fibers are present in the communicating rami, but do not yet extend into the anlagen of the sympathetic trunks (fig. 1, c.r.). "Accompanying" cells are present all along the spinal nerves and the communicating rami. At the tips of the growing rami, cells appear to become detached and to wander into the anlagen of the sympathetic trunks in advance of the growing fibers (fig. 1, i.e. c.r.).

In embryos 10 mm. in length, the fibers of the communicating


SYMPATHETIC SYSTEM IN MAMMALS


223


rami extend into the anlagen of the sympathetic trunks (fig. 2, c.r.). It may be noted in passing that in some embryos 10 mm. in length the number of cells accompanying the fibers of the spinal nerves has materially decreased, while in other embryos of the same length, the "accompanying" cells are equally as numerous as in the preceding stages. We may conclude, therefore, that



i «.*••• • • • $J % \ -V.V.7


Fig 3. Transverse section of the dorsal part of the neural tube of an embryo 7 mm. in length, showing cells migrating into the dorsal nerve-root, X 275.

there is a notable degree of individual variation in the rate of development.

(b) Cell migration. — In a recent paper 1 the writer has described the migration of nervous elements from the. neural tube and the spinal ganglia, along the spinal nerves and the communicating rami, into the anlagen of the sympathetic trunks. The earliest

1 A contribution to the histogenesis of the sympathetic nervous system. Anatomical Record, vol. 3, no. 8, pp. 458-465.


224


ALBERT KUNTZ



CLLGr


Fig. 4. Transverse section of the neural tube and the anlage of the sympathetic trunk of an embryo 11 mm. in length, X 125.


evidence of the migration of medullary cells from the neural tube is found in embryos about 4.5 mm. in length. At this stage the neural crest is not yet differentiated into ganglia, but appears as an inconspicuous ridge spreading laterally from the median dorsal line of the neural tube. It is so inconspicuous indeed that in many sections it may be distinguished only under favorable conditions. In a few instances the fibers of the ventral nerve-roots have penetrated the walls of the neural tube and are accompanied


SYMPATHETIC SYSTEM IN MAMMALS


225


by medullary cells which have broken through the external limiting membrane.

In embryos 7 mm. in length, the spinal ganglia are distinct, but are not completely formed as yet, and have receded but a short distance from the point at which the fibers of the dorsal nerveroots enter the neural tube. In transverse sections, numerous breaches may be observed in the external limiting membrane in



<£■ TL IS


ciiivr. fnrf s ub


Fig. 5. Transverse section of the ventral part of the neural tube of an embryo 9mm. in length, showing cells migrating into the ventral nerve-root, x 250.


the region of the dorsal nerve-roots. Rows of cells practically touching each other end to end may be traced from the mantle layer, through these breaches, into the proximal parts of the dorsal nerve-roots (fig. 3, c.m.d.n.r.). Further evidence for the migration of medullary cells into the dorsal nerve-roots is presented by the fact that in many sections where no breaches occur, cells are crowded close to the external limiting membrane in this region. In embryos 9 mm. and over in length, this area is always


226


ALBERT KUNTZ


occupied by the fibers of the dorsal nerve-roots, and cells are rarely found among them (fig. 4, d.n.r.). Migration of medullary cells into the dorsal nerve-roots evidently ceases before the 9 mm. stage is reached. It is probably only a transient process which takes part in the development of the cerebro-spinal ganglia.

In transverse sections of embryos from 6 to 11 mm. in length, breaches occur frequently in the external limiting membrane in the region of the ventral nerve-roots. Medullary cells migrate in considerable numbers, through these breaches, into the ventral nerve-roots (fig. 5, c.m.v.r.) . This fact has recently been observed by Carpenter and Main ('07). I have been able to substantiate


sprri



LCC.T

i


• 9


f



Fig. 6. Section showing the path of the communicating ramus in an embryo 7 mm. in length, X 190.


their observation that cells may be observed "just inside the external limiting membrane, in an intermediate position half in and half out of the neural tube, and in the base of the ventral nerve-root just outside the external limiting membrane."

The orientation of the cells in the neural tube is such, during the period of migration, that the cells which migrate into the dorsal and the ventral nerve-roots seem to have their origin in more or less distinct regions. In the dorsal region most of the migrating cells move quite directly outward from the dorsal zone, while others tend dorso-laterally in slight curves from regions ventral to the dorsal nerve-roots. In the ventral region some of the


SYMPATHETIC SYSTEM IN MAMMALS 227

migrant cells move quite directly outward from the ventral zone, while others tend ventro-laterally from the region in which later the lateral horn of the gray matter arises.

These migrant medullary cells, with similar elements which wander out from the spinal ganglia, migrate peripherally along the fibers of the spinal nerves. In transverse sections of embryos 7 or 8 mm. in length, the spinal ganglia are not sharply limited distally. Cells become separated from their distal ends and migrate peripherally along the fibers of the sensory roots. There is no recognizable difference between the cells which wander down from the spinal ganglia and those which migrate out from the neural tube. It is impossible, therefore, to distinguish the cells which migrate from the neural tube along the ventral nerveroots from those which become separated from the spinal ganglia, after they have advanced beyond the point of union of the sensory and motor roots. These "accompanying" cells are present in the spinal nerve-trunks as far as the latter may be traced. Fibers are not present as yet in the communicating rami, but at a point a little above the level of the aorta, cells, either singly or in small groups, deviate from the course of the spinal nerves nearly at right angles and migrate through the mesenchyme toward the dorso-lateral surfaces of the aorta, along the paths later occupied by the fibers of the communicating rami (fig. 6, i.c.c.r.). These findings do not differ essentially from those of Froriep. They differ from those of the older investigators primarily in the fact that cells migrate peripherally not only from the spinal ganglia but also from the neural tube along the fibers of the ventral nerve-roots.

It is important to note at this point that the cells accompanying the fibers of the spinal nerves actually migrate peripherally. Fig. 7 indicates schematically the course and the direction of the cells which migrate along the spinal nerves and the communicating rami into the sympathetic anlagen. Fig. 8 is designed to show approximately the relative number of " accompanying " cells present in the spinal nerves in successive stages during the period of migration, and also the relative number remaining in the nerve-trunks after migration has ceased. The figures in the hori


228


ALBERT KUNTZ


zontal line represent the lengths of the embryos in mm. ; the figures in the vertical line indicate the number of cells present in a given length of longitudinal sections of the spinal nerves, as they appear in transverse sections of the embryos, taken at random between the point of union of the sensory and the motor roots and the origin of the communicating rami. Embryos which seemed to be most normal in their development were selected, and the curve is based on the averages of ten independent counts. This curve



Fig. 7. Diagrammatic transverse section of an embryo 9 or 10 mm. in length. The arrows indicate the course and the direction of the cells migrating from the neural tube and the spinal ganglia into the sympathetic anlagen. ao., Aorta, c.r., Path of communicating ramus, d.n.r., Dorsal nerve-root, pv.a., Anlagen of prevertebral plexuses, sp.g., Spinal ganglion, sp.n.. Spinal nerve. supr., Suprarenal bodies, sy., Sympathetic trunks, v.n.r., Ventral nerve-root. ir.b., Wolffian body.


indicates that the rate of migration reaches its maximum in embryos 9 mm. in length, and that migration practically ceases when a length of 13 mm. is attained. It also indicates that a relatively small but fairly constant number of cells remains in the spinal nerves after migration has ceased.

As already indicated in reviewing the literature, Kohn and Neumayer have attempted to account for the cells giving rise to the sympathetic nervous system, by local differentiation of elements


SYMPATHETIC SYSTEM IN MAMMALS


229


already present in the sensory and the motor nerve-roots. This view seems to be quite generally accepted by the advocates of the theory of local differentiation and the multicellular nature of nerve-fibers. These two conceptions obviously go hand in hand. It is not the writer's purpose to discuss the nature of nerve-fibers. Suffice it to say that in the light of the recent investigations of Cajal, Harrison, and others, the neurone theory




Fig 8. Curve designed to indicate the relative rate of migration of cells from the neural tube and the spinal ganglia along the spinal nerves, in successive stages of development. For explanation see text.


seems to be firmly established. On the other hand, if the " accompanying" cells do not migrate peripherally we cannot account for the rapid decrease in the number of such cells present in the proximal part of the spinal nerves, which, as shown in fig. 8, takes place in embryos from 9 to 13 mm. in length. Mitotic figures occur occasionally in the nerve-roots as well as in the nerve-trunks and in the communicating rami. Doubtless, many

THE JOURNAL OF COMPARATIVE NEUROLOGY AND PSYCHOLOGY, VOL. 20, NO. 3.


230 ALBERT KUNTZ

of the "accompanying" cells arise by the mitotic division of " indifferent" cells along the course of migration, but these mitotic figures are by no means sufficiently numerous to account for the multitudes of cells which take part in the development of the sympathetic trunks. Furthermore, I have observed cells along the spinal nerves and the communicating rami, which, as will be shown later, are neuroblasts. Such cells were recently described by Cajal ('08) in the spinal nerves, and the communicating rami in embryos of the chick. According to Cajal, "these neuroblasts do not correspond to the neurocytes of Kohn, but to the real motor cells in the neural tube." These facts are incompatible with the theory of local differentiation.

(c) Later development. — In embryos 12 mm. in length, the anlagen of the sympathetic trunks are rapidly becoming fibrous. They still appear as continuous cell-columns showing little evidence of their future segmental character. The earliest fibers of the longitudinal commissures, therefore, do not grow out through the inter-gangliar spaces, but the cells become aggregated into distinct ganglia after the sympathetic trunks have become fibrous. " Accompanying" cells are still present along the spinal nerves and the communicating rami, but they are notably fewer than in the preceding stages. Cells may still be observed migrating from the neural tube and the spinal ganglia, but such migration probably does not continue far beyond this stage. In embryos over 12 mm. in length, the motor niduli are sharply limited, and medullary cells are rarely seen along the fibers of the ventral nerveroots as they traverse the marginal veil. The spinal ganglia are also becoming more sharpty limited distally, and cells no longer become separated from them. The later development of the sympathetic trunks consists in progressive changes and growth of the elements already present.

(d) Nature of migrating cells. — In his excellent work on the earliest differentiations in the central nervous system, Schaper ('97) made a most thorough and detailed study of the cells which arise by the mitotic division of the "germinal" cells (Keimzellen) of His. These cells were originally described by His as cells


SYMPATHETIC SYSTEM iN MAMMALS


231



-'UK.


Fig. 9


Fiq. 10


Fi g. 9 . Diagram illustrating the developmental relationships of the neuroblasts and the embryonic supporting cells in the neural tube of mammalian embryos (modified from Schaper, '97). Elongated dotted cells = ependymal cells; large circular cells with crosses = germinal cells of His ; plain rounded cells = indifferent cells; rounded cells with dotted crosses = indifferent cells which undergo further division by mitosis; rounded cells with dots in center = neuroglia cells; block cells = nerve cells.

Fig. 10. Ganglion cells, neuroblasts, and indifferent cells, X 1100. a., Cells in the spinal ganglia, b., Neuroblasts in the ventral nerve-roots, c, Neuroblasts in the spinal nerve trunks, d., Bundles of fibers with accompanying indifferent cells, from the spinal nerves, e., Neuroblasts in the communicating ramus. /., Neuroblasts in the anlagen of the sympathetic trunks.


of ectodermal origin undergoing mitotic division near the internal limiting membrane of the embryonic neural tube, giving rise to cells which develop into neuroblasts. Schaper has shown that the cells arising by the mitotic division of the "germinal" cells of His do not all develop into neuroblasts. They are cells of an indifferent character. In the lower vertebrates they are transformed either into neuroblasts or into embryonic supporting cells. In the higher vertebrates, however, many of these " indifferent" cells retain a capacity for further propagation by di


232 ALBERT KUNTZ

vision and give rise to new generations of "indifferent" cells which may develop either into neuroblasts or into embryonic supporting cells. The accompanying figure (fig. 9), modified from Schaper, has been introduced to illustrate the developmental relationships of the neurones and the supporting elements in the embryonic nervous system. According to Schaper's original description, the "indifferent" cells are characterized by large rounded nuclei showing a delicate chromatin structure, and very little cytoplasm. The "neuroblasts" are characterized by large rounded nuclei showing little structure in the interior except a well defined nucleolus, and a larger cytoplasmic body which is early drawn out to a point at one side.

The great majority of the cells migrating from the neural tube and the spinal ganglia, along the spinal nerves and the communicating rami in embryos of the pig, answer to the description of the "indifferent" cells of Schaper. When observed in the motor niduli or in the distal ends of the spinal ganglia, their nuclei usually appear more or less rounded in outline and show a very delicate chromatin structure. The cytoplasm is so meager that it can be observed only under the most favorable conditions. As these cells migrate peripherally they assume a more elongated form. In the ventral nerve-roots many of them have assumed their maximum elongation soon after they have left the neural tube. In the lines of cells which may be observed migrating out of the neural tube, the inner ones are often nearly circular in outline, the outer ones are greatly elongated, while those in intermediate positions show varying degrees of elongation. This elongation cannot be accounted for mechanically, as by the squeezing through a narrow aperture in the external limitingmembrane. These apertures are usually broad enough to permit free passage of the cells. It is probably due to more subtle forces which are operative in the process of migration. Furthermore, it is by no means certain that such a change of shape always takes place. Rounded "indifferent" cells are sometimes observed far distally along the course of migration, while elongated cells are present in the motor niduli and in the distal ends of the spinal ganglia.


SYMPATHETIC SYSTEM TN MAMMALS 233

Among the "indifferent" cells of Schaper, cells are occasionally found which are characterized by large rounded or elongated nuclei showing a well defined nucleolus and very little chromatin structure, and a considerable quantity of cytoplasm which is usually drawn out to a point at one side (fig. 10). These cells are obviously the "neuroblasts" of Schaper. They are few in number, but occur all along the path of migration of the sympathetic cells. I have observed them in the ventral nerve-roots both inside and outside the external limiting membrane, in the spinal nerves, in the communicating rami, and in the anlagen of the sympathetic trunks.

The histogenetic relationships of the cells taking part in the development of the sympathetic trunks will be considered further in section V. The facts of importance at this point are that cells which are endowed with a capacity to develop into neurones, migrate peripherally from the neural tube and the spinal ganglia, and that some of these cells migrate into the anlagen of the sympathetic trunks. These facts establish a direct genetic relationship between the sympathetic and the central nervous system. We are not to suppose, however, that all the cells taking part in the development of the sympathetic trunks actually migrate as such from their sources in the neural tube and the spinal ganglia. Doubtless, many arise by the mitotic division of "indifferent" cells along the course of migration. The sources of these migrating elements are, therefore, sufficient to account for all the cells which take part in the development of the sympathetic trunks and the sympathetic plexuses genetically related to them.

II. PREVERTEBRAL PLEXUSES

(a) Development. — In embryos 10 mm. in length, the anlagen of the prevertebral plexuses may be recognized as small cellaggregates lying along the ventro-lateral surfaces of the aorta in the dorsal and the lumbar regions. In these regions the sympathetic trunks are not sharply limited ventrally. Cells become separated from their ventral margins and migrate ventrally along the sides of the aorta (fig. 11, c.m.pv.). In the region of the


234


ALBERT KUNTZ


suprarenals, such cells have descended as far as the mesial surfaces of these bodies which, at this time, appear as compact cellcolumns more or less circular in transverse section, lying parallel with the aorta a short distance from its ventro-lateral surfaces (fig. 11, supr.). Posterior to the suprarenals, small cell-aggregates are present all along the ventro-lateral surfaces of the aorta as far as the origin of the iliac arteries. It is impossible, in this stage, to determine the limits of the anlagen of the several sym



CMi/i/


4*I»




msm



\mp\c


<ines.


Fig. 11. Transverse section through the sympathetic trunks and the suprarenal bodies of an embryo 9 mm. in length, X 150.


pathetic plexuses in this region. The only distinction between the cell-aggregates which constitute the anlagen of the cceliac, the renal, the abdominal aortic, and the hypogastric plexuses consists in a difference in degree of development. Development proceeds somewhat more rapidly in the anterior than in the posterior region. In transverse sections through this region, traces of one or the other of these plexuses appear in nearly every section.


SYMPATHETIC SYSTEM IN MAMMALS 235

In embryos 13 mm. in length, the cell-aggregates lying along the ventro-lateral surfaces of the aorta have become more pronounced. The greatest development occurs in the region of the suprarenals. At a few points, fibers may be traced from the sympathetic trunks into the anlagen of the prevertebral plexuses (fig. 12, f.pv.). In the region of the coeliac plexus, fibers have advanced farther peripherally and may be traced for a short distance into the mesentery. The anlagen of the abdominal aortic plexus have developed into a loose network which completely surrounds the aorta ventrally.

In embryos 16 mm. in length, the prevertebral plexuses are becoming more distinct and more fibrous. The sympathetic trunks are more sharply limited ventrally, except in the region of the suprarenals. At this point there is still a continuous line of sympathetic cells extending from the sympathetic trunks into the cell-aggregates associated with the suprarenals.

(b) Histogenetic relationships. — The cells constituting the anlagen of the prevertebral plexuses show all the characters of the cells present in the sympathetic trunks. Continuous lines of cells may be traced from the latter into the former. The prevertebral plexuses, therefore, stand in direct genetic relationship to the sympathetic trunks. Mitotic figures occur occasionally along the courses of migration from the sympathetic trunks as well as in the anlagen of the prevertebral plexuses. Doubtless, a goodly number of the cells taking part in the development of the prevertebral plexuses arise by the mitotic division of "indifferent" cells along the courses of migration. The development of the prevertebral plexuses is, therefore, entirely analogous with the development of the sympathetic trunks.

III. VAGAL SYMPATHETIC PLEXUSES

(a) Introductory. — Under the term vagal sympathetic plexuses, we shall consider those plexuses, usually regarded as sympathetic, which are directly related to the vagi; viz., the cardiac plexus and the sympathetic plexuses in the walls of the visceral organs.


236


ALBERT KUNTZ




•:• V


» P




\nes.


73va.


Fig. 12. Transverse section through the sympathetic trunks and the anlagen of the cceliac plexus of an embryo 12 mm. in length, X 150.

Our knowledge concerning the development of the sympathetic plexuses related to the vagi is very limited. The older workers generally gave little attention to the peripheral sympathetic plexuses. Onodi ('86), though he traced the origin of the sympathetic trunks and the prevertebral plexuses to the spinal ganglia, could not derive the peripheral sympathetic plexuses from the same source, because he found no cellular connections between the latter and the sympathetic trunks. He believed it necessary, therefore, to cling to the doctrine of Remak ('47) with regard to the peripheral sympathetic plexuses and derive them from the mesoderm. His. Jr., ('91) traced the origin of the peripheral sympathetic plexuses, including the sympathetic plexuses in the walls of the digestive tube and the sympathetic components related to the vagi, to cell-swarms which migrate peripherally from the anlagen of the sympathetic trunks.


SYMPATHETIC SYSTEM IN MAMMALS 237

Later writers have generally assumed that the cardiac plexus and the sympathetic plexuses in the walls of the visceral organs have their origin in the sympathetic trunks, but the course of their development has not been made clear. The literature bearing on this point is conspicuously meager.

My own observations, as indicated in a recent paper, 2 have shown that the cardiac plexus and the sympathetic plexuses in the walls of the visceral organs do not owe their origin to the sympathetic trunks as has hitherto been supposed, but that the}- arise from cells which migrate from the vagus ganglia and the walls of the hind-brain along the fibers of the vagi.

Because the origin of these plexuses is distinct and -eparate from the origin of the sympathetic trunks and the sympathetic plexuses described above as prevertebral plexuses, they cannot properly be characterized as prevertebral sympathetic plexuses. In view of their relation to the vagi I have chosen to designate them as vagal sympathetic plexuses. The term "vagal sympathetic" is a departure from the established nomenclature, but inasmuch as there is no good collective term which could be applied to the cardiac plexus and the sympathetic plexuses in the walls of the visceral organs, it has seemed well, for the sake of clearness, to employ a new term.

(b) Myenteric and submucous plexuses. — In transverse sections of embryos 6 and 7 mm. in length, in the region of the cesophagus, the vagus trunks appear as large bundles of loosely aggregated fibers accompanied by numerous rounded or elongated cells. These cells, which, as will be shown later, are of medullary and ganglionic origin, are easily distinguished from the cells of the surrounding mesenchyme by their larger size and the characteristic chromatin structure of their nuclei. Many of them appear to become separated from the nerve-trunks and to wander into the walls of theoesophagus until the latter is completely surrounded by these migrant cells. In a few sections short fibers are seen to bend from the vagus trunks toward the cesophagus (fig. 13, vag.


2 The role of the vagi in the development of the sympathetic nervous system. Anatomischer Anzeiger, Bd. 35. no. 15. 16. pp. 381-390.


238 ALBERT KUNTZ

b.). From the tips of these growing fibers, cells pass in well defined paths into the walls of the oesophagus. It is probable that most of the cells which become separated from the vagi wander out along the fibers of these growing branches. The cells which have wandered into the walls of the oesophagus are not arranged in well defined rings as yet, but are loosely scattered in the tissues (fig. 13, m.s.p.).

The fibers of the vagi do not yet extend beyond the region of the heart. In transverse sections through the stomach, the paths of the vagus branches are indicated by the presence of numerous


rasp

ox ft 6 ft J*&\



Fig. 13. Transverse section through the oesophagus and the vagus trunks in an embryo 7 mm. in length. X 160.


cells like those described above. These cells show a tendency to spread until they have completely surrounded the walls of the stomach. Similar cells are found scattered in the walls of the intestine as far as the latter can be traced. Thus, it appears that, having once become established in the anterior region of the digestive tube, these cells migrate posteriorly along its course.

That these migrant nervous elements found in the walls of the digestive tube have migrated thither from the vagus trunks cannot be doubted. There is no difficulty in tracing cells from the


SYMPATHETIC SYSTEM IN MAMMALS 239

tips of the growing branches of the vagi into the walls of thecesophagus. Furthermore, it is impossible to trace cells from any other source. There is no evidence as yet of the migration of cells from the sympathetic trunks or from the prevertebral plexuses toward the walls of the digestive tube. Neither cellular nor fibrous connections occur between the sympathetic trunks or the prevertebral plexuses and the sympathetic plexuses in the walls of the digestive tube until the latter have become well established.

In transverse sections of embryos 9 mm. in length, there is no evidence of cells wandering from the vagus trunks toward the oesophagus except along the fibers of the growing branches. These courses are still plainly visible. The migrant cells in the walls of the oesophagus have become arranged in more definite rings, and none are found scattered in the surrounding tissue. Numerous cells still accompany the fibers of the vagi all along their course and seem to escape freely at their growing tips.

In embryos 12 mm. in length, the number of cells in the proximal part of the vagus trunks has materially decreased. Most of those still remaining probably subserve a supporting function. The more distal parts still contain numerous cells. It is probable, however, that the migration of cells along the vagi does not continue far beyond this stage. In the region just anterior to the stomach, the vagus trunks have broken up into a loose network which is the beginning of the oesophageal plexus. Vagus fibers still accompanied by numerous cells may now be traced along the lesser curvature of the stomach. The anlagen of the coeliac plexus are well established, but there are no fibrous connections as yet between them and the anlagen of the sympathetic plexuses in the walls of the digestive tube.

In embryos 16 mm. in length, the vagus trunks as well as their branches, many of which have established connections with the myenteric and the submucous plexuses, are apparently free from migrating cells. In the walls of the oesophagus, the cells which have wandered in are aggregated into more or less distinct groups arranged in two broken rings. The myenteric and the submucous plexuses are thus becoming distinct. A similar arrangement, though less definite, is apparent also in the walls of the intestine.


240 ALBEKT KUNTZ

Fibrous connections have become established with the sympathetic trunks as well as with the cceliac and the hypogastric plexuses. It is interesting to note that all these sympathetic nerves still contain numerous "accompanying" cells which are apparently migrating peripherally along their fibers. It is probable, therefore, that cells wander down from the sympathetic trunks into the myenteric and the submucous plexuses after these fibrous connections are established.

(c) Pulmonary plexuses. — In transverse sections of embryos 6 or 7 mm. in length, some of the cells which wander from the vagus trunks toward the oesophagus, in the region of the bifurcation of the trachea, are carried out along the anterior and the dorsal surfaces of the bronchi. These cells obviously give rise to the anlagen of the pulmonary plexuses.

(d) Cardiac plexus. — The first unmistakable evidence of ganglia pertaining to the cardiac plexus is found in embryos about 12 mm. in length. In transverse sections through the anterior region of the heart, small groups of nervous elements are observed ventral to the trachea (fig. 14, car.p.), a few of which have penetrated deep into the angle between the aorta and the pulmonary artery. These cell-aggregates constitute the anlagen of the earliest ganglia of the cardiac plexus. They are without fibrous connections as yet, but a few short fibrous branches are seen to arise from the vagus trunks and the left recurrent nerve, which extend toward the heart (fig. 14, earn). These are obviously the earliest cardiac nerves. Their fibers are still loosely aggregated and are accompanied by numerous cells, some of which appear to escape at the tips of the nerves and to migrate toward the cardiac ganglia in advance of the growing fibers. Nerves cannot be traced as yet from the sympathetic trunks toward the heart, and there is no evidence of the migration of cells from the sympathetic trunks into the anlagen of the cardiac plexus.

In embryos 16 mm. in length, branches of the vagi as well as cardiac nerves having their origin in the sympathetic trunks may be traced into the ganglia of the cardiac plexus. Here again it is interesting to note that while the branches of the vagi are apparently free from migrating cells, the cardiac nerves having their


SYMPATHETIC SYSTEM IN MAMMALS


241


vaqb.


ft \ CO. ft

o


\: Q


msp.




.-brec.




- :


A i + '-:\



-■ ; ■-£?;$



Fig. 14. Transverse section through the oesophagus, the vagus trunks, and the anlagen of the cardiac plexus in an embryo 13 mm. in length, X 100.


origin in the sympathetic trunks still contain numerous " accompanying' 7 cells which are apparently migrating peripherally along their fibers. It is probable, therefore, that the cardiac plexus also receives cells from the sympathetic trunks after the sympathetic cardiac nerves have become established.

This stage in the pig may be compared with the human embryo 10.2 mm. in length described by His, Jr. ('91). He also observed that, in this stage, the branches of the vagi are comparatively free from cells, while the cardiac nerves having their origin in the sympathetic trunks contain many cellular elements. He con


242 ALBERT CUNTZ

eluded, therefore, that the ganglia of the cardiac plexus are composed exclusively of cells which have migrated thither from the sympathetic trunks.

The above observations prove conclusively that the earliest anlagen of the cardiac plexus in the pig arise from cells which migrate thither from the vagus trunks. This is probably true for all mammals. In the human embryo of His, Jr., referred to above, the cardiac plexus already had fibrous connections with both the vagi and the sympathetic trunks. The anlagen of the cardiac plexus would probably have been found considerably earlier.

My observations on the later development of the cardiac plexus in the pig do not differ essentially from those of His, Jr., 3 on the human embryo, except that the earliest cardiac nerves having their origin in the sympathetic trunks are less intimately associated with the vagi, and enter the cardiac plexus independently. This fact was also observed by His, Jr., in embryos of the cat.

(e) Cell migration along the vagi. — In sections taken at right angles to the axis of the trunk, in the head region of embryos 9 or 10 mm. in length, medullary cells may be observed migrating from the walls of the hind-brain into the rootlets of the vagus and the spinal accessory nerves (fig. 15, c.m.vag.r.). That these cells wander out in considerable numbers cannot be doubted. In many sections medullary cells are observed drawn out into coneshaped heaps in the nerve-rootlets as they traverse the marginal veil. Occasionally one of these cells is observed half in and half out of the neural tube, and many are present in the nerve-rootlets just outside the external limiting membrane.

In sagittal sections the entire vagus trunk is seen to contain many of these " accompanying" cells which are apparently migrating peripherally. The ganglion of the trunk is, at this stage, a somewhat irregular oval or elliptical body which is not sharply limited distally. Cells appear to become separated from its distal end and to wander peripherally along the vagus trunk. Mi


3 Abhdl. Math-physischen Classe d. Konigl. Sachs. Gesell. d. Wiss. Bd. 8, Leipzig 1891.


SYMPATHETIC SYSTEM IN MAMMALS


243


totic figures occur frequently in the ganglion of the trunk and occasionally all along the vagus nerve.

The course and the direction of the cells migrating peripherally along the fibers of the vagi are indicated by the arrows in fig. 16. That these cells actually migrate peripherally cannot be doubted. The number of "accompanying" cells present in the vagus trunks increases rapidly until a maximum number is reached in embryos 9 or 10 mm. in length; then it decreases rapidly until the embryos


\T ! •

i


nb.


6 *%





vfli/r-, ; f



Fig. 15. Section through the rootlets of the vagus nerve in an embryo 10 mm. in length, taken at right angles to the axis of the trunk, X 270.

have attained a length of about 13 mm., when only a relatively small number of cells remains distributed along the nerve-fibers. These phenomena can be explained on no other ground. Again, the preparations studied show figures of cells escaping from the growing branches of the vagi into the anlagenof the cardiac plexus and the sympathetic plexuses in the walls of the visceral organs which are perfectly clear, and can be interpreted only to mean that these are the cells which give rise to the vagal sympathetic plexuses.

The majority of the cells migrating peripherally along the fibers of the vagi are characterized by large rounded or elongated nuclei showing a delicate chromatin structure, and very little cytoplasm.


244 ALBERT KUNTZ

These are obviously the "indifferent" cells of Schaper. Among these, other cells occasionally are found which are characterized by large rounded or elongated nuclei showing a well defined nucleolus and very little chromatin structure, and a larger cytoplasmic body which is usually drawn out to a point at one side (fig. 17). These cells are obviously the " neuroblasts" of Schaper. From this description it is obvious that the cells which migrate from



Fig 16. Diagram designed to show the relation of the vagi to the vagal sympathetic plexuses. The arrows indicate the course and the direction of the cells migrating from the walls of the hind-brain and the vagus ganglia into the anlagen of the vagal sympathetic plexuses.

a. a., Aortic arch, aw., Auricle, b. c, Buccal cavity, br., Bronchi, ear. p., Anlagen of cardiac plexus, cb.l., Cerebellum. /.&., Fore-brain, g.t., Ganglion of the trunk. m. b., Mid-brain, m.ob., Medulla oblongata, n A., Neural tube, oe., (Esophagus. oe.p., (Esophageal plexus, p. a.. Pulmonary artery. St., Stomach, t., Trachea. ven., Ventricle, ven. IV., Fourth ventricle, v.t., Vagus trunk. 10., Roots of vagus nerve. 11., Roots of spinal accessory nerve. 12., Roots of hypoglossal nerve. r.J., First cervical nerve.

the vagus ganglia and the walls of the hind-brain along the vagi are cells of the same character as those which migrate from the neural tube and the spinal ganglia along the spinal nerves.

The above observations prove conclusively that the myenteric and the submucous plexuses, the pulmonary plexuses, and the


SYMPATHETIC SYSTEM IN MAMMALS 245

cardiac plexus have a common origin which is distinct and separate from the origin of the sympathetic trunks. They arise from cells which have their origin in the vagus ganglia and the walls of the hind-brain. As in the case of the sympathetic trunks, however, we are not to suppose that all the cells taking part in the development of the vagal sympathetic plexuses actually migrate as such from their sources in the cerebro-spinal nervous system. Doubtless, many arise by the mitotic division of "indifferent" cells along the course of migration and in the anlagen of these plexuses. The vagus ganglia and the walls of the hindbrain, therefore, constitute a source which is sufficient to account



Fig. 17. Neuroblasts and indifferent cells located in the vagi and the ganglia of the trunk, X 1100.

a., Neuroblast in the vagus rootlets, b., Neuroblasts in the vagus trunks, c, Neuroblasts in the ganglia of the trunk. (/., Bundles of fibers with accompanying indifferent cells, from the vagus trunks.

for all the cells which take part in the early development of the vagal sympathetic plexuses. Migrant cells cannot be traced from the sympathetic trunks into the anlagen of these plexuses until the nerves connecting the latter with the sympathetic trunks are present. At this time the vagal sympathetic plexuses are well established, and the great majority of the cells taking part in their development are already present. We may conclude, therefore, that the nerves entering the vagal sympathetic plexuses

THE JOURNAL OF COMPARATIVE NEUROLOGY AND PSYCHOLOGY, VOL. 20, NO. 3.


246 ALBERT KUNTZ

from the sympathetic trunks represent later connections, and play only a secondary part in their development.

These conclusions differ widely from the views hitherto generally accepted concerning the development of the cardiac plexus and the sympathetic plexuses in the walls of the visceral organs, but the facts on which they are based are perfectly clear. Furthermore, they obviate certain difficulties which arise in any attempt to derive these plexuses from the sympathetic trunks. The anlagen of the sympathetic plexuses in the walls of the visceral organs are present before any traces of the prevertebral plexuses or of cells migrating peripherally from the sympathetic trunks are found. It is obvious, therefore, that the vagal sympathetic plexuses cannot be derived from the sympathetic trunks. These findings also give the vagi an importance in the development of the sympathetic nervous system which has hitherto been unrecognized, but which is in complete harmony with other known facts.

V. DISCUSSION OF RESULTS, AND CONCLUSIONS

(a) Migration of medullary cells. — Neurological literature contains frequent allusions to the migration of medullary cells ever since the time of Balfour (75). That pioneer among the investigators of the histogenesis of nerve-forming elements observed cells which he believed to be nervous elements, migrating from the embryonic neural tube in elasmobranchs. These observations were substantiated by Beard ('88) and Dohrn ('88, '91). Herrick ('93) observed medullary elements migrating from the motor niduli into the ventral roots of the spinal nerves in amphibians, reptiles, and mammals. Ganglion cells have also been found occasionally in the motor nerve-roots in adult animals. Such cells were observed by Freud (78) in the ventral roots of the spinal nerves in Petromyzon, and by Schafer ('81) and Kolliker ('94) in the ventral roots of the spinal nerves in the cat. Thompson ('87) described cells which he interpreted as degenerating ganglion cells, in the third and fourth cranial nerves in man.

More recent investigators have frequently observed migrant


SYMPATHETIC SYSTEM IN MAMMALS 247

medullary cells and have variously interpreted them. Harrison ('01) has shown that in the salmon the spinal ganglia arise from cells which migrate out from the dorsal region of the neural tube. He also observed medullary cells migrating into the ventral roots of the spinal nerves, and suggested the possibility that tnese cells may migrate peripherally into the ganglia of the sympathetic trunks and there give rise to sympathetic excitatory neurones. Bardeen ('03) observed that in mammalian embryos a certain number of cells migrate from the neural tube and the spinal ganglia along the fibers of the spinal nerves. He suggested that these cells take part in the development of the neurilemma. He believes, however, with Vignal ('83) and Gurwitsch ('00), that in mammals the neurilemma is derived largely from the mesoderm. Neal ('03) described medullary cells in the ventral roots of the spinal nerves in Squalus acanthias, and expressed the opinion that they take part in the development of the neurilemma. Carpenter ('06) has shown that in embryos of the chick medullary cells which he recognizes as the " indifferent" cells of Schaper, migrate from the walls of the mid-brain along the fibers of the abducent and the oculomotor nerves. According to Carpenter most of these cells become distributed along the nerve-trunks and may be recognized as the cells which give rise to the neurilemma. In the oculomotor nerve, however, some of these " indifferent" cells migrate farther peripherally and give rise to neurones in the ciliary ganglion. Carpenter and Main ('07) are of the opinion that some of the medullary cells which they observed migrating into the ventral roots of the spinal nerves in embryos of the pig become cells of the neurilemma and there subserve a supporting function similar to that of the neuroglia cells in the central nervous system. Cajal ('08) described elements which he recognizes as nerve cells in the bipolar phase, in the ventral roots of the spinal nerves and certain of the cranial nerves in the chick.

Although the advocates of the theory of local differentiation and the multicellular nature of nerve-fibers reject the theory of the migration of nervous elements, the results of recent researches are so convincing that we must accept the migration of medullary

THE JOURNAL OF COMPARATIVE NEUROLOGY AND PSYCHOLOGY, VOL. 20, NO. 3.


248 ALBERT KUNTZ

cells as a fact. The present series of observations shows, moreover that the migration of medullary elements plays a far more important role in the development of the peripheral nervous system than has hitherto been admitted. Direct observations have shown that medullary cells migrate into the ventral roots of the spinal nerves and into the roots of several of the cranial nerves. The present observations have further shown that such cells migrate also into the dorsal roots of the spinal nerves and into the roots of the vagus and the spinal accessory nerves. I have also observed medullary cells migrating into the semilunar ganglia. Furthermore, it has been shown that some of the cells which migrate peripherally from the neural tube and the cerebrospinal ganglia give rise to the sympathetic nervous system.

(b) The neurilemma. — An extended discussion of the development of the neurilemma is beyond the scope of this paper. Inasmuch, however, as the histogenesis of the neurilemma is so intimately related to the histogenesis of the sympathetic neurones, its origin may be considered briefly at this point. As the "indiffent " cells migrate peripherally from the neural tube and the spinal ganglia, they migrate not only into the anlagen of the sympathetic trunks but also along the growing fibers beyond the origin of the communicating rami. These cells as well as the cells which, as has been shown, remain distributed along the nerve-trunks after migration has ceased, obviously take part in the development of the neurilemma. They cannot be accounted for in any other way.

Not a few of the more recent investigators, including the advocates of the theory of local differentiation and the multicellular nature of nerve-fibers, are of the opinion that the neurilemma is of ectodermal origin. We agree with the advocates of local differentiation on this point, but we must disagree with them as to the manner in which the cells giving rise to the neurilemma arise and are distributed along the nerve-fibers. It is significant that Kolliker ('05), though formerly of the opinion that the neurilemma is of mesodermal origin, came to the conclusion, in his last research, that some of the cells which wander out from the spinal ganglia give rise to the neurilemma of the sensory fibers, and that the neurilemma is everywhere of ectodermal origin. Carpenter ('06) has


SYMPATHETIC SYSTEM IN MAMMALS 249

shown that migrant medullary cells actually develop into cells of the neurilemma in the abducent and the oculomotor nerves in the chick.

Proof of the medullary and the ganglionic origin of the cells giving rise to the neurilemma was difficult only until it could be demonstrated that cells actually migrate peripherally from the neural tube and the cerebro-spinal ganglia along both the spinal and the cranial nerves. The present series of observations presents conclusive evidence on this point. We may here repeat what has already been stated with regard to the cells taking part in the development of the sympathetic system. We are not to suppose that all the cells taking part in the development of the neurilemma actually migrate as such from the neural tube and the cerebro-spinal ganglia. Doubtless, many arise by the mitotic division of "indifferent" cells along the course of migration. According to this interpretation, the cells of the neurilemma are homologous with the neuroglia cells in the central nervous system.

(c) Sympathetic excitatory and sympathetic sensory neurones. — The problem of the histogenetic relationships of the sympathetic excitatory and the sympathetic sensory neurones presents peculiar difficulties. The presence of sympathetic sensory neurones in the sympathetic trunks and prevertebral plexuses has not been demonstrated. Froriep, like Langley, Kolliker, and P. Schultz, denies the existence of sympathetic sensory neurones entirely. There can be little doubt, however, that sympathetic sensory neurones are present in the sympathetic plexuses in the walls of the digestive tube. According to Bayliss and Starling ('99), the peristaltic contractions of the small intestine are true coordinated reflexes carried out by the local nervous mechanism (myenteric plexus). The later experimental work of Cannon ('06) and of Auer ('10) lends support to this view by showing that the peristaltic contractions of the stomach and the intestine may be carried on more or less regularly for a considerable length of time after both the vagi and the splanchnic nerves have been severed. These phenomena seem to indicate the existence of true sensory neurones in the sympathetic plexuses in the walls of the digestive tube.


250 ALBERT KUN1Z

Froriep's conclusion that the sympathetic neurones have their origin in the ventral half of the neural tube and migrate out along the fibers of the ventral roots of the spinal nerves is probably correct with regard to the neurones in the sympathetic trunks and the prevertebral plexuses. I have shown, however, that the vagal sympathetic plexuses arise from cells which migrate peripherally along the fibers of the vagi. If, as experimental evidence indicates, so ne of these plexuses contain sensory neurones, it is probable that these arise from cells which migrate from the vagus ganglia. While it is impossible by direct observation to trace either sympathetic excitatory or sympathetic sensory elements back to their specific source in the cerebro-spinal nervous system, the facts at our command warrant the conclusion that the sympathetic excitatory neurones arise from cells which migrate from the neural tube along the fibers of the motor nerve roots, while the sympathetic sensory neurones, wherever such neurones exist, arise from cells which migrate from the cerebro-spinal ganglia.

The nervous elements in the neural crest obviously have the same origin as those which remain within the neural tube; they are the descendents of the "germinal" cells of His. Retzius has shown that in amphioxus sensory neurones are found lying near the internal limiting membrane lining the slit-like central canal, some of which send their dendrites out to the skin. In the fishes also a relatively large number of cells remaining within the neural tube give rise to sensory fibers which run to the skin. In embryos of the salmon, according to Harrison, cells become separated from the neural tube and, migrating ventrally, give rise to the spinal ganglia. In embryos of the pig, as already indicated, medullary cells migrate from the dorsal region of the neural tube into the dorsal nerve-roots. All these facts suggest that the cerebro-spinal ganglia have arisen from cells which originally lay within the neural tube, and indicate the common origin of all sensory and motor neurones.

The orientation of the cells in the neural tube, during the period of migration, is such that the cells which wander into the dorsal nerve-roots seem to have their origin in the dorsal part of the neural tube, while those which migrate into the ventral nerve


SYMPATHETIC SYSTEM IN MAMMALS 251

roots wander out from the ventral zone and from the region in which later the lateral horn of the gray matter arises. This also is in full accord with the conditions in the adult nervous system. The neurones in the cerebro-spinal ganglia, as far as is known, are sensory in character, while the cells whose axones constitute the fibers of the motor nerve-roots are located in the ventral part of the neural tube. Furthermore, the investigations of Kohnstamm ('00) render the existence of efferent fibers in the dorsal nerve-roots of the higher vertebrates extremely doubtful. Inasmuch as nervous elements which have the capacity to develop into neurones migrate peripherally along both the sensory and the motor nerve-roots, we are driven to the conclusion that the sympathetic excitatory elements migrate from the neural tube along the fibers of the motor nerve roots, while the sympathetic sensory neurones, wherever such neurones exist, arise from cells which wander out from the cerebro-spinal ganglia. This interpretation makes the sympathetic neurones entirely homologous with the efferent and the afferent components of the other functional divisions of the peripheral nervous system.

(d) A wider application of Schaper's conception. — As has been shown in the preceding pages, the cells which migrate peripherally from the neural tube and the cerebro-spinal ganglia have a common origin; they are the descendants of the "germinal" cells of His; viz., the " indifferent " cells and the "neuroblasts" of Schaper. Therefore, Schaper's conception of the developmental relationships of the neurones and the supporting elements in the central nervous system may be extended to the sympathetic neurones and the cells of the neurilemma.

(e) The relation of the sympathetic to the central nervous system. — In the light of the present investigation, the sympathetic system bears a direct genetic relationship to the central nervous system. The cells giving rise to the sympathetic trunks, and the prevertebral plexuses migrate peripherally along the spinal nerves, while those giving rise to the vagal sympathetic plexuses migrate peripherally along the vagi. The cells giving rise to the sympathetic neurones, however, all have the same genetic relationships; they are the descendants of the "germinal" cells of His.


252 ALBERT KUNTZ

Therefore, the sympathetic neurones are homologous with the neurones in the central nervous system.

The sympathetic system is not a nervous mechanism separate from the central nervous system, but the nervous system is a unit of which the sympathetic system is a part homologous with the other functional divisions. It may be looked upon as an accession to the nervous system which has arisen comparatively late in the evolution of vertebrates, in response to an increasing demand for a nervous mechanism of a lower order, which might assume the direct control of the purely vegetative functions.

(/) Functional relations. — The reader will, undoubtedly ask what bearing the facts set forth in the preceding pages may have on physiological and psychological problems involving the sympathetic system. This question we cannot hope to answer at present. We may, however, offer a few suggestions which have presented themselves during the progress of this investigation.

Our knowledge concerning the functional relations and the physiological activities of the sympathetic system is very limited. Nor could we hope for much progress in this direction as long as the developmental relationships of the sympathetic to the central nervous system were not definitely known. The fact that the sympathetic system is homologous with the other functional divisions of the nervous system lends a new aspect to the entire problem. The fact, however, that the vagal sympathetic plexuses have their origin in the hind-brain and the vagus ganglia will probably be of even greater physiological and psychological importance. This fact indicates a close relationship between the lower centers of the brain and the innervation of the heart and the visceral organs. The suggestion is here ventured that in this relationship will probably be found the basis of certain physiological and psychological problems involving the digestive functions and the action of the heart, which have hitherto been obscure.

Here is a field for investigation which challenges the attention of both the student of physiology and the student of psychology. It is beset with the* greatest difficulties, but promises to be fruitful of the most far-reaching results.


SYMPATHETIC SYSTEM IN MAMMALS 253

VI. SUMMARY

1. The sympathetic trunks arise as a pair of cell-columns lying along the sides of the dorsal surface of the aorta. In the early stages, medullary cells migrate from the neural tube into the dorsal and the ventral nerve-roots. The cells which migrate into the ventral nerve-roots, with similar cells which wander down from the spinal ganglia, migrate peripherally along the spinal nerves. Some of these cells deviate from the course of the spinal nerves and, migrating along the paths of the communicating rami, give rise to the sympathetic trunks. These findings differ materially from those of the earlier investigators. They agree essentially with the findings of Froriep.

2. The prevertebral plexuses arise as cell-aggregates lying along the ventro-lateral aspects of the aorta in the posterior region of the body. They are derived directly from the sympathetic trunks.

3. The cardiac plexus and the sympathetic plexuses in the walls of the visceral organs are not derived from the sympathetic trunks, as has hitherto been supposed, but have their origin in nervous elements which migrate from the hind-brain and the vagus ganglia along the fibers of the vagi. In view of the relation of these plexuses to the vagi, the author has chosen to designate them as " vagal sympathetic" plexuses. These findings give the vagi an importance in the development of the sympathetic system which has hitherto been unrecognized.

4. The cells migrating peripherally from the cerebro-spinal system along the spinal nerves and the vagi are the descendants of the germinal" cells of His; viz., the "in different" cells and the "neuroblasts" of Schaper. Therefore they are homologous with the cells giving rise to the neurones and the supporting elements in the central nervous system.

5. The cells migrating peripherally along the spinal nerves and the vagi do not all take part in the development of the sympathetic system. Some become distributed along the nervefibers and give rise to the neurilemma. Therefore, the cells of


254 ALBERT KUNTZ

the neurilemma are homologous with the neuroglia cells in the central nervous system.

6. The cells taking part in the development of the sympathetic nervous system and the neurilemma do not all actually migrate as such from their sources in the cerebro-spinal system. Doubtless, many arise by the mitotic division of "indifferent" cells along the course of migration.

7. The existence of sympathetic sensory neurones in the sympathetic trunks and the prevertebral plexuses has not been demonstrated. Experimental evidence, however, indicates the presence of sympathetic sensory neurones in the sympathetic plexuses in the walls of the digestive tube. While it is impossible, by direct observation, to trace either sympathetic excitatory or sympathetic sensory elements back to their specific source in the cerebro-spinal nervous system, indirect embryological and anatomical evidence warrants the conclusion that the sympathetic excitatory neurones arise from cells which migrate from the neural tube along the fibers of the motor nerve-roots, while the sympathetic sensory neurones, wherever such neurones exist, arise from cells which migrate from the cerebro-spinal ganglia. This interpretation makes the sympathetic neurones homologous with the afferent and the efferent components of the other functional divisions of the peripheral nervous system.

8. Inasmuch as the cells migrating peripherally from the cerebro-spinal nervous system are the "indifferent" cells and the "neuroblasts'" of Schaper, Schaper's conception of the developmental relations of the neurones and the supporting elements in the central nervous system, may be extended to the sympathetic neurones and the cells of the neurilemma.

9. The nervous system is a unit of which the sympathetic system is a part homologous with the other functional divisions. The sympathetic system may be looked upon as an accession to the nervous system, which has arisen comparatively late in the evolution of vertebrates in response to the conditions of the vegetative life.


SYMPATHETIC SYSTEM IN MAMMALS 255

10. The fact that the sympathetic system is homologous with the other functional divisions of the nervous system lends a new aspect to the problems involving its functional relations. The fact that the vagal sympathetic plexuses have their origin in the hind-brain and the vagus ganglia will, doubtless, have an important bearing on certain physiological and psychological problems involving the heart action and the digestive functions.


Accepted by the VVistar Institute of Anatomy and Biology, April 21, 1910. Printed July 8, 1910.


256 ALBERT KUNTZ

BIBLIOGRAPHY

Auer, J. The effect of severing the vagi or the splanchnics or both upon gastric 1910 motility in rabbits. Amer. Jour, of Physiol., vol. 25, pp. 334-344.

Balfour, F. M. The development of elasmobranch fishes. Development of 1877 the spinal nerves and the sympathetic nervous system. Jour, of

Anat. and Physiol., vol. 11. 1880 Comparative embryology, vol. 2, pp. 384-386.

Bardeen, C. R. The growth and histogenesis of the cerebro-spinal nerves in

1903 mammals. Amer. Jour, of Anat., vol. 2, pp. 231-257.

Bayliss and Starling. The movements and innervation of the small intestine.

1899 Jour, of Physiol., vol. 24, pp. 99-143.

Beard, J. Morphological studies. 2. The development of the peripheral nervous 1888 system of vertebrates. Parti. — Elasmobranchii and aves. Quart.

Jour, of Micr. Sci., n. s., vol. 29, pp. 153-227. Cajal, S. R. Nouvelles observations sur revolution des neuroblastes, avec 1908 quelques remarques sur l'hypothese neurogenetique de Hensen Held. Anatomischer Anzeiger, Bd. 22, pp. 1-25 and 65-78. Cannon, W. B. The motor activities of the stomach and small intestine after 1906 splanchnic and vagus section. Amer. Jour, of Phsiol., vol. 17,

pp. 429-442. Carpenter, F. W., The development of the oculomotor nerve, the ciliary gan 1906 glion, and the abducent nerve in the chick. Bull. Mus. Comp. Zobl., Harvard College, vol. 48, pp. 141-228.

Carpenter, F. W., and Main, R. C. The migration of medullary cells into th.

1907 ventral nerve-roots of pig embryos. Anatomischer Anzeiger, Bd 31, pp. 303-306.

Dohrn, A. Studien zur Urgeschichte des Wirbelthierkorpers. Nr. 17. Nerven1891 faser und Ganglienzelle. Histogenetische Untersuchungen.

Mitteil. d. Zoolog. Station zu Neapel, Bd. 10. 1901 Studien zur Urgeschichte des Wirbelthierkorpers. Die Schwannschen Kerne, ihre Herkunft und Bedeutung. Erwiderung an A. v. Kolliker. Mitteil. d. Zoolog. Station zu Neapel, Bd. 15. Froriep, A. Die Entwickelung und Bau des autonomen Nervensystems. Medi 1907 zinisch-naturwiss. Archiv, Bd. 1, pp. 301-321.

Gurwitsch, A. Die Histogenese der Schwannschen Scheide. Archiv f. Anat.

1900 u. Physiol., Jahrg. 1900, anat. Abt., pp. 85-94.

Harrison, R. G. Ueber die Histogenese des peripheren Nervensystems bei Salmo salar. Archiv f. mikr. Anat., Bd. 57, pp. 354-444.

1904 Neue Versuche und Beobachtungen tiber die Entwickelung der peri pheren Nerven der Wirbeltiere. Sitzungsber. d. Niederrhein. Ges. f. Nat. u. Heilkunde zu Bonn, 1904, B., S., 55. 1906 Further experiments on the development of peripheral nerves. Amer. Jour, of Anat., vol. 5, pp. 121-131.


SYMPATHETIC SYSTEM IN MAMMALS 257

His, W. Ueber die Anfange des peripheren Nervensystems. Archiv f. Anat. u1879 Physiol., anat. Abt., pp. 455-482.

1888 Zur Geschichte des Gehirns sowie der centralen und peripherischen

Nervenbauen beim menschlichen Embryo. Abhld. d. Sachs. Gesell. Wiss., Math-phys. Classe, Leipzig, Bd. 14, pp. 339-392.

1889 Die Neuroblasten und deren Entstehung im embryonalen Mark.

Archiv f. Anat. u. Physiol., Jahrg. anat. Abt., pp. 249-300. His, Jr., W. Die Entwickelung des Herznervensystems bei Wirbelthieren1891 Abhdl. Math-physischen Classe d. Konigl. Sdchsischen Gesell. d Wiss., Bd. 8, Leipzig, 1897 Ueber die Entwickelung des Bauchsympathicus beim Hiihnchen und Menschen. Archiv Anat. u. Entwg., Supplement. Hoffmann, C. K. Zur Entwickelung des Sympathicus. I. Die Entwickelungs1900 geschichte des Sympathicus bei den Selachiern (Acanthias vul garis), Verhdl. d. Konigl. Akad. v. Wetensch. te Amsterdam. (Tweede Sectie), Deel vii. 1902 Zur Entwickelung des Sympathicus. II. Die Entwickelungsgeschichte des Sympathicus bei den Urodelen. Verhdl. d. Konigl. Akad. v. Wetensch. te Amsterdam, Deel viii. Huber, G. Carl. Lectures on the sympathetic nervous system. Jour. Comp.

1897 Neurol, and Psychol., vol. 7, pp. 73-145.

Johnston, J. B. The nervous system of vertebrates. Philadelphia.

1906

Jones, W. C. Notes on the development of the sympathetic nervous system in the

1905 common toad. Jour. Comp. Neurol, and Psychol., vol. 15, pp. 113 131.

Kohn, A. Ueber die Entwickelung des peripheren Nervensystems. Verhdl. d.

1905 anat. Gesell. auf d. 19 Versamml. in Gemf. Erganzungshefl zum 27.

Bd. d. Anat. Anz., Jena. 1907 Ueber die Entwickelung des sympathischen Nervensystems der Saugetiere. Archiv f. mikr. Anat., Bd. 70, pp. 266-317. Kohnstamm, O. Zur Theorie des Reflexes von hinterer Wurzel auf hintere Wur 1900 zel. Zentralbl.f. Physiol., Bd. 14, p. 457.

Kolliker, A. Ueber das Vorkommen von Nervenzellen in den vordern Wurzeln 1894 der Riickenmarksnerven der Katze. Verhdl. Gesell. Nalurf. u.

Aerzte. 66. Versamml., Theil 2, Halfte 2, p. 363. 1905 Die Entwickelung der Elemente des Nervensystems. Zeitschrift f. Wiss. Zool., Bd. 82, pp. 1-38. Ktjntz, A. A contribution to the histogenesis of the sympathetic nervous system. 1909 Anatomical Record, vol. 3, pp. 458-465.

1909 The role of the vagi in the development of the sympathetic nervous system. Anaiomischer Anzeiger, Bd. 35, pp. 381-390. Lillie, F. R. The development of the chick, pp. 256-261. New York, 1908


258 ALBERT KpNTZ

Marcus, H. Ueber den Sympathicus. Sitzungsberichi d. Gesell. f. Morph. a.

1909 Physiol, in Munchen.

Xf.al. H. V. The development of the ventral nerves in the Selachii. Mark Anni 1903 versary Volume, New York. pp. 291-313.

Xevmayer. L. Histogenese und Morphogenese des peripheren Nervensystema,

1900 der Spinalganglien und des Nervus Sympathicus. Handbuch der

vergl. u. experiment. Entvrickelungslehre der Wirbeltiere, Bd. 2, 3.

Theil. pp. 513-026.

Onodi, A. D. Ueber die Entwickelung des sympathischen Nervensystems.

1886 Zweiter Theil. Archivf. mikr. Anat., Bd. 26, pp. 291-313.

Paterson, A. M. Development of the sympathetic nervous system in mammals.

1890 Phil Trans, of the Royal Society, vol. 181, pp. 159-186.

Remak. On the independent alimentary nervous system. Berlin.

1S47 Schafer. A. E. Notes on the occurrence of ganglion cells in the anterior roots of 1881 the cat's spinal nerves. Proc. Roy. Soc. London, vol. 31, p. 3-48.

Schaper, A. Die frl'iesten Differenzierungsvorgange im Zentralnervensystem. 1897 Kritische Studie und Versuch einer Geschichte der Entwickelung

nervoser Substanz. Archivf. Entw.-Mech., Bd. 5. pp. 81-132. Schenck, S.. and Birdsall. W. R. Ueber die Lehre von der Entwickelung der ls7< Ganglien des Sympathicus. Mittcil. aus d. embryol. Instit. iii,

Wien. Van Wijhe, J. W. Ueber die Mesodermsegmente des Rumpfes und die Entwicke1889 lung des Exkretionssystems bei Selachiern. Archivf. mikr. Anat..

Bd. 33. Yk;xal, W. Memoire sur le developpement des tubes nerveux chez les embryons 1S83 de mammiferes. Archives de Physiol, norm, et pathol.. 3. serie,

T. I. Pari-.


The Control Of Phototactic Reactions In Hyalella By Chemicals

Hartley H. T. Jackson

From the Zoological Laboratory, University of Wisconsin

Loeb 1 in a brief preliminary paper has shown that specimens of Gammarus pulex, which are normally negatively phototactic, may be made positive if they are placed in certain chemicals of the proper degree of concentration. For example, positive phototaxis was produced when the amphipods were placed in solutions of hydrochloric, oxalic or acetic acid of about m/500. Loeb obtained similar results with a 2^m solution of ethyl alcohol, a m/6 solution of ether, and a m/25 solution of ammonium chloride, but a m/10 solution of boracic acid failed to produce such an effect. My own experiments were performed on another amphipod, Hyalella knickerbockeri, in the endeavor to ascertain if Loeb's results would hold true in this species, and to test the effect of various chemicals in different concentrations. The results obtained were similar to those of Loeb, but it was found thatHyalellas were made positive by boracic acid if they were dropped into a saturated solution. Tartaric acid produces no change in their reaction.

My results with salts were quite parallel to those of Loeb; I found some ammonium salts to make them decidedly positively phototactic; some potassium salts made them weakly positive; potassium bromide and potassium iodide made them strongly positive. Potassium chloride and potassium chlorate produced no marked change in their phototactic response, nor did any of the sodium salts, or magnesium sulphate. I tried several alkalies, but here, as with the salts, there seemed to be no relation be ^oeb, J. The Control of Heliotropic Reactions in Fresh-water Crustacean, by Chemicals, especially CO>. University of California Publications. Physiology vol. 2, p. 1. 1904.


260 HARTLEY H. T. JACKSON

tween the chemicals used and the reactions. Some potassium salts produced a change in the phototaxis of the animals, other potassium salts did not; some acids produced a change in the phototaxis, others did not. It is the same with the alkalies; ammonium hydroxide causes all the animals to become positive immediately when they are dropped into a solution of .0075 per cent, but when the animals are put into a solution of potassium hydroxide, or sodium hydroxide, or potassium carbonate of any concentration which will not kill them outright, there is no change in their phototactic response; they still remain negative. Loeb 2 claims that in all probability light produces chemical changes in the eye or skin of the animals, and that these changes are responsible for the phototactic reactions. If this be true, it might seem not improbable that some definite relation would be found between the classes of chemicals employed and the reaction, but such experiments as have thus far been made tend to prove that no such relation exists.

In endeavoring to test more thoroughly the effects of chemicals upon phototaxis, experiments were conducted in the dark room, the source of light being an electric tungsten bulb of 350 candlemeters intensity. Seven chemicals were used, namely, ethyl alcohol, ammonium hydroxide, and hydrochloric, nitric, acetic, picric, and chromic acids. I first determined the lowest per cent solution of each of these chemicals which would cause a reversal in the phototactic reaction of the Hyalellas; in other words, the least per cent solution which would make them positive when they were dropped directly into it. I found that ethyl alcohol of .074 per cent, ammonium hydroxide of .0075 per cent, hydrochloric acid of .0067 per cent, nitric acid of .0053 per cent, acetic acid of .01 per cent, picric acid of .0053 per cent, and chromic acid of .0046 per cent would produce this result. In each of these cases, all or nearly all the animals would be in the positive end of the dish making frantic efforts to get nearer the source of light. My next experiment was to place several of the animals in an oblong glass dish, four inches long and one and one-half inches

2 Loeb, J. Comparative Physiology of the Brain and Comparative Psychology. New York, 1900, and I.e.


PHOTOTACTIC REACTIONS IN HYALELLA 261

wide, containing twenty-five cubic centimeters of distilled water. Very slowly and gradually the concentration of the solution was increased by adding constant small amounts of a given chemical at intervals of five to fifteen minutes. A careful record was kept of the number of animals positive and the number negative at certain concentrations throughout the experiments. In every case the light was moved and the reaction of the animals tested from each end of the dish; that is, when the animals had oriented themselves in one end of the dish, the light was then transposed to that end and the reaction tested again. The results were surprising ; in each case the amphipods remained decidedly negatively phototactic, even though the concentration was carried far beyond the point at which they became positive when dropped directly into the solution. The concentration in each case was increased to the point where the majority of the animals died; nevertheless they were negative, and decidedly so, throughout the experiment until death occurred. Thus, when the concentration is gradually increased, the Hyalellas were negative in .64 per cent solution of ethyl alcohol, .05 per cent solution of ammonium hydroxide, .022 per cent hydrochloric acid, .05 per cent nitric acid, .43 per cent acetic acid, .029 per cent picric acid, and 0.22 per cent chromic acid. The results will appear more evident by a study of the accompanying table. In this table the first column of figures indicates the per cent solution; following each of these figures, to the right, is indicated the number of animals which were positive and the number negative at each reading of a given chemical at a concentration given in the first column. The line drawn across each column indicates the point at which the animals were positively phototactic when dropped directly into the solution. Where the sum of the positive and negative specimens is not equal to the sum of the positive and negative used at the beginning of the experiment, it indicates that a number of animals died from the effect of the chemical.

It is evident that it is not chemical change in the tissues which caused the reversal of reaction, for, if it were, it would be impossible to increase the concentration and still have the animals


262


HARTLEY H. T. JACKSON


TABLE OF PHOTOTACTIC REACTION IN HYALELLA.


PER CENT SOLUTION


MTRIc ACID


ACETIC ACID


PICRICACID


CHROMIC ACID


HYDROCHLORIC ACID


AMMONIUM HYDROXIDE


ALCOHOL


USED


No. +

No.

+

No.

+

No.

+

No.

+

No.

+

No.

+

0000


21



17



20



18


1 20


20


20


.0008


21



17



20



18


21


20


1 1!)


0016


21


1


16



20



18


21


20


20


' .0023


21



17



20


1


17


21


20


20


.0031


21



17



20



18


21


20


20


.0038


21



17



20



18


21


20


20


.0046


21 21




17 17



20


1

2


17 15


21 21


20 20


20


.0053



20


20


0060


21



17


1


19


1


13


21


20


20


.0067


21



17


4


16



14


21


20


20


0075


21



17


2


16



9


21


20


20


.0100


21



17


1


14



9


21


1 19


20


015


21



17


1


12



7


14


19


20


.022


21



17


1


11



4


3


1 15


20


.029


21


1


16


2


8





1 12


20


.036


19



17







12


20


043


13



17







7


20


050


4


1


16







5


20


.060




17








20


.067




17








20


074




17








1 19


100




17








1 19


150



1


16








20


.220




17








20


.290




17








1 18


.360




14








19


.430




(i •








18


.500











18


.570











18


.640











18


PHOTOTACT1C REACTIONS IN HYALELLA 263

remain negative ; even though the increase be gradual, the chemical change in the eye or skin of the animal would take place as readily when the necessary concentration was reached as when the animal was dropped directly into the concentrated chemical. Moreover, the duration of immerison in the chemical is greater when the concentration is gradually increased; this would tend to produce a more complete chemical change in the animal than when it is dropped momentarily into the chemical. It is probable, therefore, that these various changes of reaction are due, not to chemical changes in the eyes or skin of the animal, but to a sudden stimulation or shock to the nervous system.

I wish here to express my acknowledgments and thanks to Dr. S. J. Holmes for his kindly criticisms and aid in preparing this paper.


Accepted by the Wistar Institute of Anatomy and Biology, March 21, 1910. Printed July S, 1910.


THE DEVELOPMENT OF THE HYPOGLOSSAL GANGLIA OF PIG EMBRYOS

C. W. PRENTISS

Northwestern University Medical School

EIGHT FIGURES

The use of dissected pig embryos for classwork in embryology suggested itself to the writer some years ago. Upon trial it was found that very instructive preparations could be made and with much greater ease than might be expected. It is difficult for students to grasp the relations of developing organs as seen in sections and a dissected embryo showing the primitive organs in position is very helpful in remedying this evil. It is the intention of the author to present, in a future paper, the results of his work along these lines, with directions for dissection and figures of the more instructive preparations. The form and relations of the various organs may be seen as accurately as in reconstructions made from serial sections by experts. There is this disadvantage in dissection, that some of the finer details of structure may be lost. For research it commends itself as a check to errors which may occur in making reconstructions; for it enables one in a short time to study a considerable number of embryos.

The nervous system lends itself most easily to dissection. The mesenchyma crumbles away from the more tenuous nervous tissue of suitably prepared embryos, making it possible to lay bare the entire nervous system of pigs varying in length from 6 to 20 mm. The smaller embryos are more easily dissected, as no cartilage or bone is encountered.

My dissections brought out some points in connection with the cerebral nerves which have not hitherto been cleared up, and my

THE JOURNAL OF COMPARATIVE NEUROLOGY AND PSYCHOLOGY, VOL. 20, NO. 4.


266 C. W. PRENTISS

purpose in the present paper is to describe the rudimentary ganglia which occur between the first cervical nerve and the vagus, ganglia which I shall refer to as the hypoglossal ganglia because of their undoubted connection with this nerve. My descriptions will necessarily deal with the development of the last four cranial nerves, the glosso-pharyngeal, the vagus, the spinal accessory, and the hypoglossal.

LITERATURE

The occurrence of hypoglossal ganglia was first described by Froriep ('82) in the sheep and the ox. He traced the development of a single ganglion, anterior to the first cervical, which it resembled in form, though smaller in size. The single distal root of this ganglion joined the most caudad root of the hypoglossal nerve. Anterior to this hypoglossal ganglion the neural crest was undifferentiated. Froriep and Beck ('95) found a precervical ganglion present in the adult throughout those groups of mammals in which the first cervical ganglion was well developed.

Martin ('91), investigating cat embryos, found five ganglionic masses posterior to the jugular ganglion, and five roots of origin for the hypoglossal nerve. He concludes, therefore, that these ganglia are the dorsal ganglia of the hypoglossal, though he gives no figures in support of his view.

Lewis ('03) in his excellent paper on the anatomy of a 12 mm. pig found extending caudad from the jugular ganglion "a beaded commissure ending in a small knob. In the track of the commissure, but separated from it, is an irregular ganglionic mass. After another interval there appears a small fragment, then follows the first cervical ganglion." In one case he found a small fiber bundle connecting the irregular ganglionic mass (Froriep 's ganglion) with the hypoglossal nerve, but considers its "relation with the commissure" as "far more striking than its resemblance with a spinal ganglion." He finds the ganglion "connected with the commissure in pigs of 17 mm." In a dissected pig of the same length "the hypoglossal ganglion appeared as a detached part of


HYPOGLOSSAL GANGLIA OF PIG EMBRYOS


267


the ganglionic chain running forward to the vagus. This commissure could not be subdivided into definite ganglia ; it was characterized by irregular swellings and spurs."

Streeter ('04) in tracing the development of the peripheral nerves in human embryos finds a ganglionic crest extending from the first cervical to the superior ganglion of the glossopharyngeal and partly ensheathing the fibers of the spinal accessory nerve. In embryos 10 to 13 mm. long the neural crest becomes differentiated into four or five rather diffuse cell masses. Froriep's ganglion resembles the others, being irregular in form and without roots. The hypoglossal nerve originates as four or five parallel roots. There is no correspondence between these and the rudimentary ganglia, nor are the ganglia segmentally arranged. He considers the three or four anterior cell masses as cerebral ganglia and "not to be confused with the precervical ganglion of Froriep."

MATERIALS AND METHODS


The number and length of the embryos dissected are given in the following table:


STAGE


LENGTH


LENGTH OF EXAMPLE


NUMBER DISSECTED


I


5- 7 mm.


6 mm.


5


II


8-10 mm.


8.5-9 mm.


4


III


12-14 mm .


13-13.5 mm.


10


IV


17-20 mm.


17-18 mm.


10


V


28-30 mm.


28 mm.


4


VI


41-50 mm.



2



Total


35






All drawings were made with the aid of an Abbe camera lucida and a Zeiss a* objective. The embryos were fixed in Zenker's fluid and the dissections were first stained, cleared in creosote and drawn as transparent objects. It was thus possible to locate microscopic cell masses and trace the course of very small fiber bundles. The dissection was then transferred to alcohol and examined as


268


C. W. PRENTISS


an opaque object by reflected light to obtain the contour of the different structures. By making several dissections of the same stage I believe that the finer structures were more accurately and completely reproduced than could be done by serial reconstructions.

DESCRIPTION OF DISSECTIONS

Stage 1. 6 mm. In this embryo (fig. 1) the ganglia were connected from the glossopharyngeal (IX) back to the caudal region


crist


qanq Froriep



I*- gang sup. , qangjugul

it

- gang, pet roj I


Fig. 1. Dissection of a 6 mm. pig showing in the hypoglossal region the undifferentiated neural crest. See explanation of figures on page 282.


by continuous bands or loops of cells, undifferentiated portions of the neural crest. The ventral roots of the spinal nerves (d C 2 ) are large, but the dorsal ganglia show little differentiation into fibers, though short distal and proximal roots are present. The hypoglossal originates as five or six parallel roots, resembling those of the spinal series but uniting, in the later embryos of this stage, to form a common trunk (XII). The spinal accessory (XI), as an arched bundle of fibers, could be traced from the fourth cervical ganglion cephalad to the vagus. Dorsal to the accessory and partly ensheathing it is a flattened band of cells, the


HYPOGLOSSAL GANGLIA OF PIG EMBEYOS 269

neural crest (crist. new.), extending forward to the jugular ganglion of the vagus (gang, jugul.). Opposite the posterior root of the hypoglossal a marked ventral loop and thickening in the crest (gang. Froriep) shows the position of Froriep's ganglion. Anteriorly the crest of cells is broader and a Tew short proximal rootlets are present. A depression separates it more or less completely from the cells of the jugular ganglion which is flattened and diffuse with 8-10 short proximal roots. The glossopharyngeal is short and its superior ganglion is joined to the jugular ganglion by a small cord of cells.

The remarkable features at this stage are then: the early development of the spinal roots; the resemblance of the hypoglossal to a series of ventral spinal roots; the existence of a nearly undifferentiated neural crest between the jugular and the first cervical ganglion.

Stage 2. 8.5-9 mm. In embryos of this stage (figs. 2 and 3), the roots of the spinal nerves are longer and more fibers are developed. The first cervical ganglion is distinctly double in fig. 2. It is still connected with the second cervical ganglion by a loop of cells. The neural crest between the first cervical and the jugular ganglia shows the most marked change. The slight enlargement opposite the posterior root of the hypoglossal which we saw in the first stage has now grown to be a spindle-shaped mass of cells (gang. Froriep) with two proximal roots and a distal bundle of fibers which extends to the root of the hypoglossal. This ganglion (Froriep's) is still connected with cellular loops (nod.) of the neural crest, but in this respect it does not differ from the cervical and sacral ganglia of this stage. It strongly resembles one of the two cell masses composing the first cervical ganglion. Anterior to it is a smaller mass of cells (gang, hypogl.) from which a proximal root is developing. This is the "terminal knob" of the "commissure" figured by Lewis in the 12 mm. pig (1903, pi. I). Anteriorly the crest shows five diffuse irregular cell masses which become gradually larger toward the jugular ganglion with indications of proximal roots. The jugular ganglion is of more definite form and is pointed ventrally. Dorsal


$"*


hypoq

qanq Fror

nod. J J


Fig. 2. Dissection of an 8.5 mm. pig showing Froriep's ganglion with a distal root to the hypoglossal nerve, and a double cervical ganglion.



rrujelencephalon


gang.hypoql. ,._/

gang /

Froriep _/

nod _/.__



-m>~ qanq iup


qanqjuqul. TXT.


__ qanq petros

v __ qcingnodos




Fig. 3. Dissection of a 9 mm. pig showing an early stage in the differentiation of the neural crest to form the hypoglossal ganglia.


HYPOGLOSSAL GANGLIA OF PIG EMBRYOS


271


to the ganglion nodosum of the vagus {gang, nodos.) a small bundle of fibers is given off to form the peripheral portion of the spinal accessory {XI').

Two embryos of 8.5 mm. showed the structure of fig. 2. In two others of 9 mm. no distal root has developed from Froriep's


gang sap.



Fig. 4. Dissection of a 13 mm. pig to show a seiies of hypoglossal ganglia and two double cervical ganglia.

ganglion (fig. 3). The first cervical is not distinctly double and shows no connection with the second cervical ganglion. The proximal roots are longer, however, and the neural crest near the jugular ganglion is better differentiated. Two flat distal strands of mixed cells and fibers pass down parallel to the spinal accessory fibers and enter the vagus.

This stage brings out three important points: (1) the first cervical ganglion frequently originates as a double structure; (2)


272


C. W. PRENTISS


Froriep's ganglion sends fibers to the hypoglossal at an earlier stage than has been described by other investigators; (3) its resemblance to one of the divisions of the first cervical ganglion is marked.


/



qang. hypogl gang Froriep


nod.


rmjelencephalon


qanqsup qang.jugul "Z EC



Fig. 5. Dissection of a 13 mm. pig showing a scries of eight hypoglossal ganglia and the persistence of the neural crest from the superior ganglion to the first cervical.

Stage 3. 13 mm. This embryo, of about the same age as that studied by Lewis, is characterized by a further elongation of the distal and proximal roots and by a greater differentiation of the neural crest anterior to the first cervical nerve. Five of the ten embryos dissected belonged to the types shown in fig. 4 and fig. 5. Here we see a cord of cells passing cephalad from the first cervical ganglion. In fig. 5 it joins Froriep's ganglion. In fig. 4


HYPOGLOSSAL GANGLIA OF PIG EMBRYOS 273

it passes under its proximal roots. Froriep's ganglion possesses now one, now two proximal roots, with a distal bundle of fibers entering the posterior root of the hypoglossal nerve, and it is united to a more irregular ganglionic mass by a strand of cells which varies in size in different embryos. This second ganglion also shows one or two proximal roots, and a distal root is in evidence. A third cell mass more cephalad shows a proximal root, is pointed at its distal end and projects ventrad to the spinal accessory. Anterior to these three occurs a series of five cell masses, more diffuse, more closely united and elongate in the anteroposterior line. A small strand of cells unites the more cephalad of these to the jugular ganglion, and this in turn is connected with the superior ganglion of the glossopharyngeal. Each of these cell masses possesses from two to four proximal roots, and two pairs of the adjacent roots are joined by cellular loops. Distal roots either join the spinal accessory or form part of flattened bundles which pass to the trunk of the vagus.

In three of these five embryos the first cervical ganglion was distinctly double as in figs. 4 and 5. In the other two it gave evidence of a double origin. An interesting point was the difference in structure exhibited on the right and left sides of the same embryo. For example, in two other cases Froriep's ganglion was well developed with a distal hypoglossal root on the right side, small and without hypoglossal root on the left side. In two cases no hypoglossal root was found on either side, a condition similar to that figured by Lewis, who, however, found a small distal fiber bundle in a second embryo.

To sum up: The 13 mm. embryo shows (1) the neural crest anterior to the first cervical ganglion differentiated into about eight ganglionic masses; (2) the two posterior of these send roots to the hypoglossal in the majority of cases, and the condition figured by Lewis is apparently the exception, rather than the rule; (3) a persistence of cellular cords still unites the various links in this chain of ganglia with each other and with the first cervical and jugular ganglia; (4) the first cervical ganglion is often double in structure, and all the spinal ganglia are elongate and unlike the rounded nodules figured by Lewis; (5) in the ganglionic chain


274


C. W. PRENTISS


it is not possible to distinguish an anterior cranial series belonging to the vagus complex, and a pre-cervical group of spinal ganglia as maintained by Streeter ('04) ; (6) the connection between the chain of ganglia and the vagus is not as marked in most cases as Lewis figures, and when well developed represents a persistence


qanqFror



i.qanq.sup.


.qancj.jiiqul qancj.petros.


__qanq. nodos. XE


Fig. 6. Dissection of a 17 mm. pig showing three distal roots passing from the hypoglossal ganglia to the ventral roots of the hypoglossal nerve and a fourth incomplete root.

of the ganglionic crest — a persistence common also to the spinal ganglia; (7) the spinal accessory root could be traced back to the sixth cervical ganglion.

Stage 4- 17-18 mm. Further elongation of the nerve roots is accompanied by a differentiation of the hypoglossal ganglia greater than at any other stage (figs. 6 and 7) . The ganglia are relatively smaller but a continuous cord of cells may still be traced


HYPOGLOSSAL GANGLIA OF PIG EMBRYOS


275


from the first cervical to the jugular ganglion. In fig. 7, the first cervical ganglion is double. Froriep's ganglion is independent of the rudimentary crest and shows two proximal roots, one connecting with a small lateral spur. The distal root is large and


gang hypoql qonq Froriep


mijelencephalon



HE.

qanq nodos.


Fig. 7. Dissection of an IS mm. pig with three distal roots passing from the hypoglossal ganglia to the ventral roots of the hypoglossal nerve. The first cervical ganglion is double and the neural crest is persistent from the first cervical to the jugular ganglion.

its fibers join the caudal root of the hypoglossal some distance ventral to the spinal cord. Next anterior are three connected masses of ganglion cells which increase in size cephalad. Two proximal and two distal roots are seen. The distal roots are extremely small and it is doubtful whether their fibers enter the


276 C. W. PRENTISS

hypoglossal. Anterior to these three ganglionic masses is a large ganglion elongate in the line of the ganglionic crest. Five groups of proximal roots arise from the myelencephalon and distal roots join the spinal accessory. The anterior of these is the largest. Soon after it unites with the spinal accessory a short very slender cord of cells and fibers passes over to the jugular ganglion (fig. 7, x). This is the only connection between the hypoglossal ganglia and the jugular ganglion of the vagus.

The hypoglossal ganglia are best developed in the 17 mm. embryo shown in fig. 6. In this case three distal roots join the hypoglossal from as many ganglia, and a fourth distal spur is present. The first cervical ganglion is only partially divided. Comparing the hypoglossal ganglia with one division of the first cervical ganglion as seen in fig. 4, the resemblance is plain.

From ten dissections at this stage we would note the followingpoints: (1) The hypoglossal ganglia here reach their highest differentiation; (2) in every case Froriep's ganglion was present with a well developed hypoglossal root, — in three cases two such hypoglossal roots were present, in one case the ganglion was forked and in one case (fig. 6) there was evidence of four hypoglossal ganglia with distal roots ; (3) this stage proves that the connection between the jugular ganglion and the hypoglossal ganglion is of little importance other than showing that both are derived from a common neural crest ; (4) as observed in preceding stages, there is great variation in the hypoglossal ganglia of different individuals, and on the two sides of the same embryo; no two were exactly alike; (5) the root of the spinal accessory could be traced back to the eighth cervical ganglion.

Stage 5. 28-30 mm. In succeeding stages the hypoglossal ganglia show retrogressive changes as to structure and relative size. Fig. 8 shows the persistence of a single hypoglossal ganglion (Froriep's) posteriorly. Anteriorly three closely connected ganglia are seen, the more posterior sending a spur backward, which ends abruptly. This condition was found in two cases. In one case a double hypoglossal ganglion was present, and in one case a small spurred fragment occupied a position near the


HYPOGLOSSAL GANGLIA OF PIG EMBRYOS


277


middle of the hypoglossal chain, that is, about midway between Froriep's ganglion and the jugular ganglion.

Stage 6. 41-50 mm. Two dissections showed conditions similar to the preceding. A double Froriep's ganglion was found


qanq. hijpoq 1 qanq.Froriep


qanq. sup. qanq.juqul



_A qanq.petros.


Fig. 8. Dissection of a 28 mm. pig showing Froriep's ganglion and a distal root to the hypoglossal nerve. The middle portion of the series of hypoglossal ganglia, present in earlier stages, has disappeared.


in one case relatively smaller than in the embryo of 30 mm. Its arrested development was shown by its unchanged position and small size. It lies still partly dorsal to the spinal accessory, while the cervical ganglia have shifted ventrad owing to the elongation of their roots, and to their own growth.


278 C. W. PRENTISS

THEORETICAL DISCUSSION

It seems evident from the different dissections we have made that the hypoglossal nerve first develops as several ventral spinal roots (5 to 6 in number) which arise independently, lie parallel to each other and are in series with spinal nerves. Later these independent roots unite to form the trunk of the hypoglossal nerve. Comparing the earlier stages with the later, it would seem that the more anterior roots atrophy and this is in harmony with the observations of Bremer ('08).

Allowing that the hypoglossal is a composite of ventral spinal roots, then we should expect to find their ganglia, if present, between the first cervical and jugular ganglia. We do find a chain of ganglia occupying this very position, but they are rudimentary, appear late and soon show retrogressive changes. They arise from the same neural crest as do the spinal ganglia and root ganglia of the vagus and glossopharyngeal. They form a continuous series, but show variations in form characteristic of all rudimentary structures. As far as their early development is concerned they cannot be divided into a pre-cervical and a cerebral group, nor is there an overlapping of spinal and cerebral ganglia, according to the theory of Froriep.

Objections have been raised as to whether these ganglia were really homologous with spinal ganglia. The points made have been: (1) Difference in form, these rudiments rarely resembling spinal ganglia; (2) their frequent connection with the vagus rather than with the hypoglossal; (3) their lack of segmental arrangement; (4) their many variations and irregularities of form .

As to their form, in the sheep Froriep found it similar to that of the cervical ganglia. In the pig where they are best developed they are usually spindle-shaped, but broader forms, resembling spinal ganglia, have occasionally been observed. In many mammals too, including man, the first cervical ganglion loses its typical form and may become vestigial. In the pig the first cervical is smaller than the other spinal ganglia and develops


HYPOGLOSSAL GANGLIA OF PIG EMBRYOS 279

later. I have frequently found it double, consisting of two spindleshaped masses of cells. Other spinal ganglia show the same condition. The first cervical ganglion possesses always several proximal roots (4-5) and the distal root arises as two distinct bmidles. Froriep's ganglion never shows more than two proximal roots, generally only one, and never more than one distal root. As these ganglia are also spindle-shaped I would regard them as not homologous with a spinal ganglion, but as comparable to one of the spindle-shaped divisions of such a cervical ganglion as seen in fig. 4. Two of the rudimentary hypoglossal ganglia with their two distal roots would be exactly homologous with a single spinal ganglion. The separation of the two parts of the ganglion could be accounted for as due to their arrested development ; as they do not appear until late and the pre-cervical region grows more rapidly, the two masses of cells representing a ganglion would be separated to a greater or less extent. At any rate, we find that the first cervical ganglion is frequently divided in the same way, sometimes the second cervical, and the same thing may occur throughout the spinal series. The irregularities of structure and constant variation which we find in the hypoglossal ganglia is merely typical of all rudimentary structures. Lewis has objected that the connection of the hypoglossal ganglia with the vagus is more marked than their relation to the hypoglossal. He figures the hypoglossal ganglia (his " beaded commissure") as continuous with the jugular ganglion. I have shown that the direct connection with the jugular ganglion is only important as showing their common development from the neural crest. Occasionally this connection entirely disappears and it is to be compared to the loops of cells which may persist between the proximal roots of two adjacent spinal ganglia. Furthermore as many as three ganglia may be connected by distal roots with the roots of the hypoglossal. It is my opinion that the anterior ganglia of this series originally related to the hypoglossal, have become connected with the vagus complex, just as in man some fibers from the first cervical ganglion have joined the spinal accessorv.


280 C. W. PRENTISS

The present lack of segmental arrangement displayed by these ganglia does not preclude their metameric origin. Their development begins considerably later than that of the spinal nerves, and the rapid growth of the region they occupy, before they make their appearance, may cause them to shift their positions with relation to their myotomes. They certainly appear in regular series and their early development is similar to that of the spinal ganglia.

As to the number of dorsal ganglia represented in the hypoglossal series, no absolute statement can be made. The evidence of comparative anatomy goes to show that four or five spinal nerves have been added to the cranial series as a result of the union with the cranium of a corresponding number of vertebrae. Meeks ('09) finds in an Acanthias embryo three rudimentary spinal ganglia located between the vagus and the first spinal nerve, the ganglion of which would correspond to Froriep's ganglion in mammals. According to this evidence, four dorsal ganglia have become rudimentary structures in mammals and the corresponding ventral roots have united to form the hypoglossal trunk. Regarding each pair of the eight ganglionic nodules found in the 13 mm. embryo as homologous to a single spinal ganglion, then we would have the same number of ganglia, four, represented between the first cervical and the jugular. The more anterior roots of the hypoglossus, which are found in the early embryos but disappear in the later stages, represent ventral roots of the vagus and glossopharyngeal according to the observations of Bremer ('08).

CONCLUSIONS

1. The jugular and superior ganglia of the vagus and glossopharyngeal nerves, the hypoglossal ganglia and ganglia of the spinal nerves arise in the pig embryo from a continuous neural crest, as observed by Streeter in human embryos.

2. The hypoglossal ganglia are retarded in their development, but appear in embryos of 13 mm. as a series of eight connected cell masses of nearly equal size.


HYPOGLOSSAL GANGLIA OF PIG EMBRYOS 281

3. According to their development, the hypoglossal ganglia can be divided only artificially into a cephalic cerebral group and a caudal pre-cervical group.

4. The first cervical and other spinal ganglia are often of double origin, composed of two spindle-shaped masses, and generally possess two distal roots.

5. The spindle-shaped ganglion of Froriep with its single distal root would therefore represent but one half of a spinal ganglion.

6. The degree of development of the hypoglossal ganglia varies in different embryos ; in the same embryo the right side may be better developed than the left, and vice versa. This is good evidence of their rudimentary or vestigial character.

7. One, frequently two or three, and in one case four hypoglossal ganglia possessed single distal roots and the fibers of three of these joined the hypoglossal nerve.

8. The connection of the hypoglossal ganglia with each other and with the jugular ganglion represents a persistence of the neural crest. It is similar to the connections which were found persisting between the roots of the spinal ganglia.

9. If we consider a pair of hypoglossal ganglia as the equivalent of a single spinal ganglion, four such ganglia would be represented in pig embryos, between the jugular and the first cervical.

10. The hypoglossal trunk develops as five or six separate ventral roots, at first parallel and independent, later uniting to form a single nerve.

11. The hypoglossal ganglia reach their maximum development in embryos 17-20 mm. long, then retrograde, though ganglia at both ends of the series may persist in the adult.

12. The spinal accessory nerve develops very early, being well formed in the youngest embryos examined (5 mm. long). As development proceeds the fibers of the spinal accessory root may be recognized farther and farther caudad. In a pig of 17 mm. a few accessory fibers were traced to a point opposite the eighth cervical ganglion.


THE JOURNAL OF COMPARATIVE NEUROLOGY AND PSYCHOLOGY, VOL. 20, NO, 4.


282 C. W. PRENTISS

BIBLIOGRAPHY

Bremek, J. L. Aberrant roots and branches of the abducent and hypoglossal

1908. nerves. Jour. Comp. New. and Psych., vol. 18, pp. 619-639, 9 figs.

Froriep, A. Ueber ein Ganglion des Hypoglossus. Archiv f. Anat. u. Physiol., 1882. Anal. Abth., 1882, pp. 279-302.

Froriep, A. und Beck, W. Ueber das Vorkommen dorsaler Hypoglossuswurzeln 1895. mit Ganglien in Reihe der Saugethiere. Anat. Am., Bd. 10, pp.

688-696.

Lewis, F. T. The anatomy of a 12 mm. pig. Amer. Jour. Anal., vol. 2, pp. 211 1903. 225, 4 pis.

Martin, P. Die Entwickelung der neunten bis zwolften Kopfnerven bei der 1891. Katze. Anat. Am., Bd. 6.

Meek, A. The encephalomeres and cranial nerves of an embryo of Acanthias

1909. vulgaris. Anat. Am., Bd. 34, pp. 473-475.

Streeter, G. L. The development of the cranial and spinal nerves in the occipi 1904. tal region of the human embryo. Amer. Jour. Anat., vol. 4, pp. 83116, 4 pis.. 14 text figs.

Accepted by the Wlstar Institute of Anatomy and Biology, May 16, 1910. Printed September 9, 1910

EXPLANATION OF FIGURES

A.11 drawings were made with the aid of a camera lucida and have been reduced to a common magnification of about 25 diameters. The figures show in surface view the right side of the myelencephalon and spinal cord from a point anteiior to the origin of the glossopharyngeal nerve to a point just caudad to the first, second or third cervical ganglion. The following abbreviations have been employed :

CI, C2, first and second cervical ganglia; crist. neur., neural crest; gang. Froriep, Froriep's ganglion; gang, hypogl., hypoglossal ganglia; gang, jugul., jugular ganglion; gang, nodos., ganglion nodosum; gang, petros., ganglion petrosum; gang, sup., superior ganglion; nod., persisting cellular nodules of neural crest; sp. cord, spinal cord; IX, X, XI, XIII, glossopharyngeal, vagus, spinal accessory and hypoglossal nerves; XI', peripheral portion of spinal accessory nexve.


The Development Of The Sympathetic Nervous System In Birds

Albert Kuntz

From the Biological Laboratory, State University of Iowa

TEN FIGURES

CONTENTS

Introduction 283

Observations 284

1. Sympathetic trunks 284

a. Introductory 284

b. Primary sympathetic trunks 286

c. Secondary sympathetic trunks 287

d. Histogenesis 288

2. Prevertebral plexuses 293

3. Ganglion of Remak 295

4. Vagal sympathetic plexuses 296

a. Introductory 296

b. Myenteric and submucous plexuses 296

c. Pulmonary plexuses 299

d. Cardiac plexus 299

e. Histogenesis 300

Discussion of results, and conclusions 301

Summary 305

Bibliography 308

INTRODUCTION

The present investigation of the development of the sympathetic nervous system in birds has grown out of an investigation of the development of the sympathetic nervous system in mammals. It was undertaken in order to further exact knowledge concerning the development of the sympathetic nervous system, to extend

1 From the Laboratories of Animal Biology of the State University of Iowa. Prof. Gilbert L. Houser, Director.


284 ALBERT KUNTZ

the writer's observations on the histogenesis of the sympathetic nervous system in mammals, and to point out certain morphogenetic differences in the development of the sympathetic system in birds and in mammals, with a view to their phylogenetic significance.

Birds and mammals have become specialized along divergent lines. Their special habits of life have brought about modification in the course of ontogeny as well as in adult structure. The sympathetic nervous system, which is concerned primarily with the control of the purely vegetati\e functions, has not escaped the modifying influence of specialized habits. It is hoped, therefore, that a more exact knowledge of the development of the sympathetic nervous system in birds may throw some new light on the problems involving the structural and the functional relationships of the sympathetic system to the central nervous system.

Inasmuch as the literature bearing on the development of the sympathetic nervous system has been reviewed by the writer in a recent paper, 2 only such references will be made to the literature in this paper as seem to be necessary.

The observations set forth in the following pages are based on embryos of the chick. The embryos were fixed in chrom-acetoformaldehyde. The sections were cut to a thickness of 10 micra and stained by the iron-ha^matoxylin method. This method, as indicated in the earlier paper, was found best adapted for purposes of this research.

OBSERVATIONS

1. Sympathetic trunks

(a.) Introductory. — His, Jr. ('97) called attention to the fact that in the chick two pairs of sympathetic trunks arise in the course of ontogeny. These he has designated as the "primary" and the ' ' secondary ' ' sympathetic trunks. According to his observations, the primary sympathetic trunks arise about the close of the third

2 The development of the sympathetic nervous system in mammals. Journal of ( 'tmiparative Neurology and Psychology, vol. 20, no. 3.


SYMPATHETIC SYSTEM IN BIRDS


285


day of incubation, as a pair of cell-columns lying along the sides of the dorsal surface of the aorta. About the beginning of the sixth day, the anlagen of the secondary sympathetic trunks arise as cell-aggregates situated just median to the ventral roots of the spinal nerves. These cell-aggregates are at first independent of each other, but become united later by longitudinal commissures. Between the fourth and the eighth day of incubation, the primary sympathetic trunks disappear, except in the most anterior region, becoming resolved into the ganglia and nerves constituting the prevertebral and the peripheral sympathetic plexuses. According



Fig. 1. Diagramatic transverse section through the thoracic region of an embryo of the chick (130 hours incubation), ao., aorta; nc., notochord; p.sy., primary sympathetic trunks; sp.g., spinal ganghon; sp.n., spinal nerve; s.sy., secondary sympathetic trunks.

to His, Jr., the cells giving rise to both the primary and the secondary sympathetic trunks are derived exclusively from the spinal ganglia.

My observations on the development of the sympathetic trunks in the chick do not differ essentially from those of His, Jr., except in one particular. I find that the cells giving rise to the primary and the secondary sympathetic trunks in the chick, like the cells giving rise to the sympathetic trunks in mammals, are not derived exclusively from the spinal ganglia, as His, Jr., believes them to be, but that they have their origin, wholly or in part, in the neural tube.


286 ALBERT KUNTZ

Figure 1 has been introduced to show the relative positions of the primary and the secondary sympathetic trunks in an embryo of the chick in the 130- hour stage.

(b.) Primary sympathetic trunks.- — The primary sympathetic trunks arise about the beginning of the fourth day of incubation, as cell-aggregates lying along the sides of the aorta and along the dorsal surfaces of the carotid arteries. At the close of the fourth day (96- hour stage) , these cell-aggregates have assumed the appearance of loosely aggregated cell-columns (fig. 2, A, p. sy.). Well marked ganglionic enlargements do not occur, but the cell-columns are not of uniform diameter. In the posterior region, the anlagen of the primary sympathetic trunks arise a little later than in the anterior region, and remain less sharply limited. They are, at this stage, not directly connected with the spinal nerves. In the thoracic region where the spinal nerves are best developed, they extend peripherally a little beyond the level of the aorta. At a point a little above the level of the aorta, cells deviate from the course of the spinal nerves and wander through the mesenchyme, either singly or in small groups, toward the sides of the aorta (fig. 2, A and B, i. c. c. r.) where they become aggregated to give rise to the anlagen of the primary sympathetic trunks.

During the course of the fifth day of incubation, the primary sympathetic trunks become more conspicuous. They move dorsally and recede a short distance from the walls of the aorta until at the close of the fifth day (120-hour stage) they appear as conspicuous cell-columns lying along the dorso-lateral aspects of the aorta a short distance from its surface (fig. 2, C, p. sy.). The primary sympathetic trunks are now sharply defined in the anterior region and are connected with the spinal nerves by distinct cellular tracts. In the posterior region, the cell-aggregates are still loosely scattered along the sides of the aorta and the cellular tracts connecting them with the spinal nerves are less distinct. The primary sympathetic trunks have now reached their maximum development. During the course of the sixth day, they decrease materially in size until at the close of the sixth day (144hour-stage) they have almost disappeared. Their complete (lis


SYMPATHETIC SYSTEM IN BIRDS


287


appearance occurs first in the thoracic region, while the last remnants may be observed in the anterior cervical region.

(c.) Secondary sy?npathetic trunks. — The anlagen of the secon dary sympathetic trunks arise about the beginning of the sixth day (120-hour stage), as ganglionic enlargements on the median sides of the spinal nerves at the point of origin of the communicating


spn.


ic r m. sp.ii.


psy LC.c.r.

A


C




<s •* ssy sp.n.


Wp-W





\ ■« &.


D



i


«y


FSjVl.


Fig. 2. Transverse sections showing successive stages in the development of the primary and the secondary sympathetic trunks in the chick. A., primary sympathetic trunk (96 hours incubation), X 200. B., Primary sympathetic trunk (105 hours incubation), X 200. C, Primary and secondary sympathetic trunks (120 hours incubation), X 200. D., Secondary sympathetic trunk (144 hours incubation), X 100. ao., aorta; c.r., communicating ramus; i.c.c.r., cells migrating from spinal nerve to primary sympathetic trunk; i.c.n., indifferent cell undergoing mitosis; p.sy., primary sympathetic trunk; sp.n., spinal nerve; s.sy., secondary sympathetic trunk.


288 ALBERT KUNTZ

rami (fig. 2, C, s. sy.). These ganglionic enlargements are at first independent of each other, but become united later by longitudinal commissures. Like the anlagen of the primary sympathetic trunks, the anlagen of the secondary sympathetic trunks appear earliest in the thoracic region and latest in the sacral region. At the beginning of the sixth day there are as yet no traces of the anlagen of the secondary sympathetic trunks in the posterior half of the body. During the course of the sixth day, the secondary sympathetic trunks become larger and more conspicuous, while the primary sympathetic trunks become correspondingly smaller. The former, being located at the point of origin of the communicating rami, are connected with the latter by the cellular tracts which connect them with the spinal nerves (fig. 2, C).

As the communicating rami become fibrous, the anlagen of the secondary sympathetic trunks become removed a short distance from the spinal nerves. In the cervical and the thoracic region they are removed to the ends of the short communicating rami (fig. 2, D, s. sy.). In the posterior region of the body, the fibers of the communicating rami extend beyond the anlagen of the secondary sympathetic trunks. At the close of the sixth day, they may be traced through the cell-aggregates still remaining scattered along the sides of the aorta, into the anlagen of the prevertebral plexuses.

In the posterior region of the body, the distinction between the primary and the secondary sympathetic trunks is never well marked. Cells gradually become aggregated in the proximal part of the communicating rami to give rise to the secondary sympathetic trunks, while the cells constituting the primary sympathetic trunks migrate ventrally into the anlagen of the prevertebral plexuses. After the sixth day, the secondary sympathetic trunks become more distinct throughout their entire length, as the ganglionic enlargements become connected by the fibers of the longitudinal commissures.

(d.) Histogenesis. — As already indicated, the sympathetic trunks arise from cells which migrate peripherally from the cerebrospinal nervous system along the spinal nerves. As soon as fibers


SYMPATHETIC SYSTEM IN BIRDS


289


appear in the ventral roots of the spinal nerves (72-hour stage) cells may be traced from the motor niduli, across the marginal veil, into the proximal part of the ventral nerve-roots. Medullary cells become aggregated in the proximal part of the ventral


c.md.n.%



m.nrr


sp.n. cmvr


Fig. 3. Transverse section of the neural tube and the spinal ganglion of an embryo of the chick (105 hours incubation), X 190. c.m.d.n.r., cells migrating into dorsal nerve-root; c.m.v.r., cells migrating into ventral nerve-root; g.c, germinal cells of His; m.n.r., motor nerve-root; nc, notochord; sn.r., sensory nerve-root; sp.g., spinal ganglion; sp.n., spinal nerve.


nerve-roots and soon appear to migrate peripherally along the nerve-fibers. While the spinal ganglia are becoming differentiated from the neural crest, cells apparently having their origin in the neural tube wander out into the spinal ganglia. Evidence of this


290 ALBERT KTJNTZ

process may be observed as early as the 64-hour stage. During the fourth and the fifth day, after the spinal ganglia have become well differentiated, a few cells may be observed migrating from the dorsal part of the neural tube into the dorsal nerveroots (fig. 3, c. m. d. n. r.). It is probable that cells do not migrate from the dorsal part of the neural tube in any considerable numbers after the spinal ganglia have become differentiated. Cell migration into the dorsal nerve-roots is probably only a transient process which takes part in the development of the spinal ganglia.

As the cells in the ventral nerve-roots migrate peripherally, they mingle with similar cells which wander down from the spinal ganglia. As there is no recognizable difference between the cells which wander out from the spinal ganglia and those which migrate peripherally along the ventral nerve-roots, it is impossible to distinguish between the cells from these two sources after they have passed beyond the point of union of the sensory and the motor nerve-roots. As these cells migrate peripherally along the spinal nerve-trunks, some of them deviate from the course of the spinal nerves and migrate toward the sides of the aorta where they become aggregated to give rise to the primary sympathetic trunks. As migration proceeds, the cells which deviate from the course of the spinal nerves no longer migrate into the primary sympathetic trunks, but become aggregated at the median sides of the spinal nerves to form the ganglionic enlargements which constitute the anlagen of the secondary or permanent sympathetic trunks.

His, Jr., has expressed the opinion that the elements composing the primary sympathetic trunks are resolved into the ganglia and nerves of the prevertebral and the peripheral sjmipathetic plexuses. In view of the comparatively enormous development of the prevertebral plexuses and of the ganglion of Remak in birds, this is obviously the fate of the elements composing the primary sympathetic trunks in the posterior region of the body. There is no evidence, however, of the peripheral migration of cells from the primary sympathetic trunks in the anterior region. While there may be some migration posteriorly along the primary sympathetic trunks, it is more probable that most of the elements composing


SYMPATHETIC SYSTEM IN BIRDS 291

these trunks in the anterior region of the body are withdrawn into the anlagen of the secondary or permanent sympathetic trunks along the cellular tracts connecting the former with the latter. The last remnants of the primary sympathetic trunks in the anterior cervical region, as His, Jr., has suggested, probably atrophy.

The period of incubation being comparatively shorter in birds than in mammals, cell migration takes place much more rapidly. It is at its height in the chick during the fourth and the fifth day of incubation. During this time breaches occur frequently in the external limiting membrane of the neural tube just opposite



c




Fig. 4. Neuroblasts drawn with the aid of the camera lucida, X 825. a., in ventral nerve-root inside external limiting membrane (105 hours incubation); b., in ventral nerve-root outside external limiting membrane (105 hours incubation); c, in spinal nerve (105 hours incubation); t/., in communicating ramus (105 hours incubation); e., in ventral nerve-root (96 hours incubation): /., in spinal nerve (96 hours incubation).

the motor niduli, and medullary cells may be traced without difficulty from the motor niduli into the proximal part of the ventral nerve-roots (fig. 3, c. m. v. r.). Numerous accompanying cells are present in the spinal nerve-trunks as far as the latter may be traced. At the close of the sixth day, the number of cells present in the spinal nerves has materially decreased. While cells are still moving peripherally along the spinal nerves, it is probable that migration from the neural tube and the spinal ganglia has practically ceased.

The great majority of the cells migrating peripherally along the spinal nerves are characterized b} r very little cytoplasm, and


292 ALBERT KUNTZ

by large rounded or elongated nuclei usually having their chromatin aggregated into one or two dense masses. These are obviously the " indifferent " cells of Schaper. Among these are found a few cells which are characterized by large rounded or elongated nuclei showing one or two dense masses of chromatin, and a larger cytoplasmic body which is usually drawn out to a point at one side. Fig. 4 shows several of these cells drawn with the aid of the camera lucida. These are obviously the " neuroblasts" of Schaper. The majority of the cells present in the mantle layer in the neural tube answer to the descriptions given above for the two types of cells migrating peripherally along the spinal nerves. There can be no doubt, therefore, that the cells accompanying the fibers of the spinal nerves have the same histogenetic relationships as the cells which give rise to the neurones and the neuroglia cells in the central nervous system. They are all the descendants of the "germinal" cells (Keimzellen) of His.

These observations are in full accord with the writer's observations on mammalian embryos. There is a marked difference, however, in the chromatin structure of the embryonic medullary cells in birds and in mammals. In mammalian embryos, the migrant medullary cells are usually quite readily recognized by the chromatin structure of their nuclei. They also usually take a slightly deeper stain than the cells of the surrounding mesenchyme. In birds, the chromatin structure of the embryonic medullary cells differs very little from the chromatin structure of the typical mesenchyme cells. Nor are they as distinctly separated from the cells of the surrounding mesenchyme by differential stains as is the case in mammals. Although the difficulties in technique are greater in the chick than in mammalian embryos, there can be no doubt that the cells accompanying the fibers of the spinal nerves are migrant medullary cells. Such cells wander out of the neural tube into the ventral nerve-roots in considerable numbers. The number of cells present in the proximal part of the spinal nerves increases rapidly until the maximum rate of migration is reached, and then decreases rapidly until migration ceases, when only a comparative 1 }' small, but fairly constant, number of cells remain distributed along the nerve-fibers. Furthermore, a few of the


SYMPATHETIC SYSTEM IN BIRDS 293

cells present in the spinal nerves are obviously neuroblasts. Such cells have frequently been observed outside the neural tube and the spinal ganglia. Cajal ('08) described cells which he recognized as nerve cells in the bipolar phase, in the motor roots of the spinal nerves and in certain of the cranial nerves in the chick. These cells, he believes, correspond to the real motor cells in the neural tube. Mitotic figures occur occasionally all along the course of migration and in the sympathetic anlagen. We are not to suppose, therefore, that all the cells taking part in the development of the sympathetic trunks actually migrate as such from their sources in the cerebro-spinal nervous system. Doubtless, many arise by the mitotic division of " indifferent " cells along the course of migration.

2. Prevertebral plexuses

The prevertebral plexuses are derived directly from the primary sympathetic trunks. They arise about the middle of the fourth day (108-hour stage), as cell-aggregates lying along the ventro-lateral aspects of the aorta from the suprarenal bodies posteriorly. In this region the primary sympathetic trunks are not sharply limited ventrally. Sympathetic cells may be traced from the latter directly into the anlagen of the prevertebral plexuses. In the sacral region, the aorta is soon completely surrounded ventrally by a ring of loosely aggregated sympathetic cells (fig. 5, hyp.).

The cell-aggregates constituting the anlagen of the prevertebral plexuses increase very rapidly. At the close of the fourth day (120-hour stage), these plexuses have become well established. Distinct lines of cells may be traced from the primary sympathetic trunks directly into cell-aggregates of considerable size lying along the median sides of the suprarenals, and wandering sympathetic cells may be observed all along the sides of the aorta from the suprarenals posteriorly. The limits of the anlagen of the several prevertebral plexuses cannot be determined at this stage. Traces of one or the other of these plexuses are not wanting in any transverse section in this entire region.


294


ALBERT KUNTZ


As incubation proceeds, the prevertebral plexuses assume more definite proportions. The cells increase in number and become more closely aggregated. At the close of the sixth day (144-hour stage), nearly all the cells which were present in the primary sympathetic trunks in the posterior region of the body have wan


1 i • •• ' i AM/ , f


Dies.

Fig. 5. Transverse section through the sacral region of an embryo of the chick (105 hours incubation), X 60. ao., aorta; g.r., ganglion of Remak; hyp., hypogastric plexus; mes., mesentery; nc., notochord; r. rectum; sp.n., spinal nerve.


dered down into the prevertebral plexuses. Neither cells nor fibers can be traced ventrally from the prevertebral plexuses as yet except in the sacral region. Here numerous cells may be traced from the hypogastric plexus directly into the ganglion of Remak.


SYMPATHETIC SYSTEM IN BIRDS


295


3. Ganglion of Remak

The ganglion of Remak arises about the middle of the fourth day, .as an oval cell-column lying in the mesentery just dorsal to the rectum (fig. 5, g.r.). Its greatest diameter occurs in the pos


mes~

Fig. 6. Transverse section of an embryo through the sacral region of an embryo of the chick (130 hours incubation), X 110. ao., aorta; g.r., ganglion of Remak; hyp., hypogastric plexus; mes., mesentery; r., rectum; sp.n. spinal nerve.


terior region. It increases in size very rapidly until at the close of the fifth day it has become a large and conspicuous column of closely aggregated cells, with a maximum diameter of about 85 micra (fig. 6, g.r.). Its diameter decreases anteriorly until it


29G ALBERT KUNTZ

terminates in the region of the genital ridges, in a slender cellular cord which Remak has called the intestinal nerve (Darmnerv).

This ganglion was described by Onodi ('86) and by His, Jr. ('97), but, as far as I have been able to learn, no worker before me has traced the cells composing it to their source. My preparations show conclusively that the cells giving rise to the ganglion of Remak are derived directly from the anlagen of the hypogastric plexus. In transverse sections through the posterior sacral region, where the mesentery is broad and the rectum lies close to the anlagen of the hypogastric plexus, cells may be traced from the latter directly into the ganglion of Remak (figs. 5 and 6).

The ganglion of Remak has no counterpart in mammals. It may be of interest to note at this point that an examination of several embryos of the turtle (kindly placed at nry disposal by Dr. F. A. Stromsten, of these laboratories) has shown that while there is no well defined ganglion in this type, corresponding to the ganglion of Remak, there are numerous cell-aggregates associated with the rectum, which evidently constitute the prototype of Remak's ganglion. It is probable, therefore, that this ganglion, so enormously developed in birds, is correlated with oviparous habits.

4- Vagal sympathetic plexuses

(a.) Introductory. — In an earlier paper I have shown that in mammals the sympathetic plexuses related to the vagi; viz., the cardiac plexus and the sympathetic plexuses in the walls of the visceral organs, have their origin in cells which migrate from the vagus ganglia and the walls of the hind-brain along the fibers of the vagi. I have, therefore, designated these plexuses as the " vagal sympathetic" plexuses. My observations on embryos of the chick show clearly that in birds also these plexuses have their origin in cells which migrate from the hind-brain and the vagus ganglia.

(b.) Myenteric and submucous plexuses. — Inembryosof thechick in the 130-hour stage, the vagus trunks may be traced posteriorly along the walls of the oesophagus just a little below its ventral


SYMPATHETIC SYSTEM IN BIRDS 297

level. The ganglia of the trunk lie close to the walls of the oesophagus just distal to the origin of the trachea. The bifurcation of the trachea occurs farther anteriorly in birds than in mammals, and the bronchi are comparatively longer. Anterior to the bifurcation of the trachea, cells deviate from the course of the vagi along the fibers of their growing branches and wander into the walls of the oesophagus. These cells are so slightly differentiated at this stage that it is no longer possible to trace them after they have entered the denser tissues of the oesophageal walls. Beyond the bifurcation of the trachea, the vagus trunks bend laterally and ventrally round the bronchi and extend along the ventro-lateral aspects of the oesophagus, continually approaching each other posteriorly. At the point where the vagi begin to bend round the bronchi, each vagus trunk gives rise to a slender branch which extends posteriorly along the wall of the oesophagus between the latter and the bronchus. These branches may be traced posteriorly but for a short distance at this stage.

At the close of the sixth day, the vagi have become more conspicuous. In the anterior region, definite lines of cells may be traced from the vagus trunks into the walls of the oesophagus where they become aggregated into more or less distinct groups arranged in two broken rings (fig. 7, m.s.p.). Posterior to the bifurcation of the trachea, cells may be traced dorsally from the vagus trunks into the walls of the oesophagus (fig. 9, m.s.p.). The vagus branches lying between the walls of the oesophagus and the bronchi have become more conspicuous and may be traced posteriorly as far as the region of the lungs. Posterior to the region of the heart, the vagus trunks lie close together and apparently break up to form a plexus ventral to the oesophagus.

During the seventh and the eighth day of incubation, the sympathetic plexuses in the walls of the digestive tube become well established. Branches of the vagi may be traced into the walls of the oesophagus, and the cell-groups constituting the anlagen of the myenteric and the submucous plexuses assume a more definite arrangement.

The sources of the cells giving rise to the myenteric and the submucous plexuses in the walls of the small intestine could not

THE JOURNAL OF COMPARATIVE NEUROLOGY AND PSYCHOLOGY, VOL. 20, NO. 4.


298


ALBERT KUNTZ


be definitely determined. In the early stages, single sympathetic cells could not be traced in the dense tissues of the walls of the digestive tube. It is difficult, therefore, to determine whether or not such cells migrate posteriorly in the walls of the digestive tube, as is the case in mammalian embryos. It is probable, however, that such is the case. On the other hand, it is probable that some of the cells which take part in the development of the myen


m , s P



|. IP--^:/*.



Fig. 7. Transverse section through the oesophagus and the vagi of an embryo of the chick (144 hours incubation), X 80. ao., aorta; m.s.p., cells giving rise to myenteric and submucous plexuses; oe., oesophagus; t., trachea; vag.n., vagus trunks.

Fig. 9. Transverse section through the oesophagus and the anlagen of the cardiac plexus of an embryo of the chick (144 hours incubation), X 80. a.c, atrial cavity; a.s., atrial septum; car. p., anlagen of cardiac plexus; m.s.p., cells giving rise to myenteric and submucous plexuses; oe., oesophagus; vag.n., vagus trunks.

teric and the submucous plexuses in the posterior region of the intestine wander out from the ganglion of Remak. There is no evidence of cells entering the sympathetic plexuses in the walls of the digestive tube from the sympathetic trunks or from the prevertebral plexuses, except through the ganglion of Remak, until


SYMPATHETIC SYSTEM IN BIRDS


299


fibrous connections are established between the former and the latter. There can be little doubt, therefore, that most of the cells taking part in the development of the myenteric and the submucous plexuses in the walls of the small intestine migrate posteriorly from the anlagen of these plexuses in the anterior region of the digestive tube.

(c.) Pulmonary plexuses. — In transverse sections through the region of the lungs of embryos in the 144-hour stage, fibers may be traced laterally from the branches of the vagi lying between the oesophagus and the bronchi. Cells wander out along these



, @&


P&fll§#l



vag. b


Fig. 8. Transverse section through the region of the lungs of am embryo of the chick (seventh day of incubation), X 80. ao., aorta; oe., oesophagus; p.p., anlagen of pulmonary plexuses; vag. b., branches of the vagi; vag.n., vagus trunks.

fibers and become aggregated to give rise to the anlagen of the pulmonary plexuses, (fig. 8, p.p.).

(d.) Cardiac plexus. — In transverse sections through the region of the head in embryos in the 120-hour stage, cells may be traced ventrally from the vagi into the septum of the atria where they become aggregated into small groups which constitute the anlagen of the cardiac plexus. In later stages, these cell-groups become more conspicuous until at the close of the sixth day they appear as distinct cell-aggregates in the atrial septum (fig. 9, car. p.).


300


ALBERT KUNTZ


(e.) Histogenesis. — In sections through the head-region of embryos in the 96-hour stage, medullary cells may be traced from the walls of the hind-brain into the rootlets of the vagus and the spinal accessory nerves (fig. 10, c.m.vag.r.). That these cells migrate peripherally from the walls of the hind-brain in considerable numbers cannot be doubted. In many sections they may be observed pushing into the nerve-rootlets in cone-shaped heaps as the latter traverse the marginal veil. Occasionally medullary cells are observed half in and half out of the neural tube, and


c.m.vag.r.


vag$



gj§i <£> ® <& <B®%,¥




j^U® <2> <5) ® ®£^


elm.


-dm. g.c.


Fig. 10. Transverse section through the wall of the hind-brain of an embryo of the chick (96 hours incubation), X 550. c.m.vag.r., cells migrating into roots of the vagus; el.m,., external limiting membrane; g.c, germinal cells of His; i.l.m., internal limiting membrane; vag.r., roots of vagus nerve.


many are present in the nerve-rootlets just outside the external limiting membrane. With similar cells which wander out from the vagus ganglia, these cells migrate peripherally along the fibers of the vagi. As these cells migrate peripherally and the vagi give rise to fibrous branches, cells wander out from the vagus trunks and give rise to the vagal sympathetic plexuses. That such is the origin of the cardiac plexus and the sympathetic plexuses in the walls of the visceral organs in the chick cannot be doubted. The figures of cells migrating from the vagi into the anlagen of these plexuses are perfectly clear. Nor can cells be traced into these plexuses from any other


SYMPATHETIC SYSTEM IN BIRDS 301

source, except possibly in the posterior region of the intestine, until fibrous connections have been established between the latter and the sympathetic trunks. By this time the great majority of the cells taking part in the development of the vagal sympathetic plexuses are already present. The connections of these plexuses with the sympathetic trunks must, therefore, be looked upon as secondary.

The period of migration of cells from the hind-brain and from the vagus ganglia along the vagi is coextensive with the period of migration of cells from the neural tube and the spinal ganglia along the spinal nerves. The cells which migrate peripherally along the vagi are cells of the same character as those which migrate peripherally along the spinal nerves; viz., they are the " indifferent" cells and the "neuroblasts" of Schaper. The cells giving rise to the vagal sympathetic plexuses, therefore, have the same histogenetic relationships as those giving rise to the sympathetic trunks. Mitotic figures occur occasionally along the vagi and in the anlagen of the vagal sympathetic plexuses. We are not to suppose, therefore, that all the cells taking part in the development of these plexuses actually migrate as such from their sources in the hind-brain and the vagus ganglia. Doubtless, many arise by the mitotic division of indifferent cells along the course of migration.

DISCUSSION OF RESULTS, AND CONCLUSIONS

The observations set forth in the preceding pages have shown that in birds the sympathetic nervous system has its origin in cells which migrate peripherally from the neural tube and the cerebrospinal ganglia. The cells giving rise to the sympathetic trunks and the prevertebral plexuses, including the ganglion of Remak, migrate peripherally along the spinal nerves, while the cells giving rise to the vagal sympathetic plexuses migrate peripherally along the vagi. These observations agree essentially with the writer's observations on the histogenesis of the sympathetic nervous system in mammals. My observations on the histogenesis of the sympathetic system agree with the findings of Froriep ('07)


302 ALBERT KUNTZ

in embryos of Torpedo and of the rabbit only in regard to the sympathetic trunks and the prevertebral plexuses. Froriep succeeded in tracing medullary cells from the neural tube into the ventral roots of the spinal nerves. According to his observations these cells, with similar cells which wander out from the spinal ganglia, migrate peripherally along the spinal nerves. At the origin of the communicating rami, cells deviate from the courses of the spinal nerves and give rise to the sympathetic nervous system. Inasmuch as Froriep does not admit of the existence of sympathetic sensory neurones, he concludes that all the sympathetic neurones in the sympathetic trunks and the prevertebral and the peripheral sympathetic plexuses arise from cells which have their origin in the ventral half of the neural tube and migrate peripherally along the ventral roots of the spinal nerves. My observations have shown conclusively that the cardiac plexus and the sympathetic plexuses in the walls of the visceral organs do not arise from cells which migrate peripherally along the spinal nerves, but have their origin in cells which migrate from the hindbrain and the vagus ganglia along the vagi. As I have pointed out in an earlier paper, experimental evidence indicates the existence of sympathetic sensory neurones in some of these plexuses. It is probable, therefore, that the sympathetic excitatory neurones arise from cells which migrate from the neural tube along the fibers of the motor nerve-roots, while the sympathetic sensory neurones, wherever such neurones exist, arise from cells which migrate peripherally from the cerebro-spinal ganglia. According to this interpretation, the sympathetic neurones are homologous with the afferent and the efferent components of the other functional divisions of the peripheral nervous system.

Froriep further believes that the axones which constitute the fibers of the motor roots of the spinal nerves are the vehicles by means of which medullary cells are transported peripherally along the spinal nerves, and that cells are carried from the spinal nerves into the anlagen of the sympathetic trunks by the axones which constitute the motor fibers of the communicating rami. He is not convinced as to whether such peripheral transportation is


SYMPATHETIC SYSTEM IN BIRDS 303

accomplished by the peripheral growth of the axones alone or whether cells may also migrate peripherally, independently of the growth of the axones. My observations do not enable me to offer any adequate explanation of the process by which the cells giving rise to the sympathetic nervous system are carried peripherally from the cerebro-spinal system. The growing nerve-fibers, doubtless, constitute an important factor in the peripheral transportation of these elements. They are not sufficient, however, to account for the entire process alone. Nor is the presence of nerve-fibers absolutely necessary to the peripheral migration of sympathetic cells. In embryos of both birds and mammals, cells may be traced from the spinal nerves into the anlagert of the sympathetic trunks before fibers are present in the communicating rami. Likewise, cells migrate ventrally from the sympathetic trunks into the anlagen of the prevertebral plexuses before postganglionic fibers appear.

Held ('09) and Marcus ('09) have recently taken exception to Froriep's views concerning the origin of the cells giving rise to the sympathetic nervous system. Held has attempted to show, for the entire vertebrate series, that the cells present in the motor nerve-roots play no part in the development of the sympathetic nervous system. He still regards the sympathetic system as an offshoot from the spinal ganglia. In the light of the present investigation, such a position is untenable. My preparations show conclusively that medullary cells migrate into the motor nerveroots in considerable numbers. These cells migrate peripherally along the spinal nerves just as certainly as do the cells which wander down from the spinal ganglia. Inasmuch as the great majority of the cells migrating peripherally along the spinal nerves are cells of an indifferent character, there is no reason to suppose that the cells which wander down from the spinal ganglia give rise to sympathetic neurones, while those which migrate from the neural tube along the fibers of the motor nerve-roots do not.

Marcus has attempted to show that the cells which Froriep observed in the ventral roots of the spinal nerves do not wander out from the neural tube, but migrate thither from the neural


304 ALBERT KUNTZ

crest. In early stages of embryos of torpedo, he has observed cell-chains connecting the neural crest with cell-aggregates in the ventral nerve-roots. He concludes, therefore, that the neural crest represents the sole source of the cells giving rise to sympathetic neurones. I have found no evidence of cells migrating from the neural crest into the ventral roots of the spinal nerves in embryos of birds and mammals. Cell-chains connecting the neural crest with the cell-aggregates in the ventral nerve-roots, doubtless, do occur in embryos of the lower vertebrates. I have observed such cell-chains in embryos of Amblystoma. This does not, however, preclude the possibility of cells migrating from the neural tube directly into the ventral nerve-roots. In the same embryos in which these cell-chains were observed, I was able to trace medullary cells from the ventral part of the neural tube directly into the ventral nerve-roots.

As has already been pointed out, the cells migrating peripherally from the neural tube and the cerebro-spinal ganglia along the spinal nerves and along the vagi are the descendants of the " germinal" cells of His; viz., the "indifferent" cells and the "neuroblasts" of Schaper. They are, therefore, homologous with the cells giving rise to the neurones and the neuroglia cells in the central nervous system. Inasmuch as some of these cells give rise to the sympathetic nervous system, the latter bears a direct genetic relationship to the central nervous system, and the sympathetic neurones are homologous with the afferent and the efferent components of the other functional divisions of the nervous system. The histogenetic relationships of the sympathetic neurones were considered at some length in my paper on the development of the sympathetic nervous system in mammals. They will, therefore, not be considered further at this point.

A comparative study of the morphogenesis of the sympathetic nervous system in birds and in mammals reveals some striking points of difference which evidently have phylogenetic significance. Two pairs of sympathetic trunks arise in the course of ontogeny in birds, while in mammals a single pair of sympathetic trunks is developed. In the early stages in mammalian embryos,


SYMPATHETIC SYSTEM IN BIRDS 305

the prevertebral plexuses show their maximum development in the region of the suprarenals. In the early stages in the chick, these plexuses show their maximum development in the sacral region. This character in birds is obviously correlated with the enormous development of the ganglion of Remak which has no counterpart in mammals. Minor differences also occur in the development of the vagal sympathetic plexuses. These morphogenetic differences, doubtless, indicate that the sympathetic system has departed more widely from the ancestral type in birds than in mammals.

A study of the development of the sympathetic system in birds, as well as in mammals, warrants the conclusion that the nervous system is a unit of which the sympathetic system is a part homologous with the other functional divisions. It may be looked upon as one of the later accessions to the vertebrate nervous system which has arisen in response to the conditions of the vegetative life. The morphogenetic differences which have been pointed out in the development of the sympathetic system in birds and in mammals obviously indicate specializations in certain directions, which have arisen in response to peculiar vegetative functions.

SUMMARY

1. The primary sympathetic trunks in the chick arise about the beginning of the fourth day of incubation, as a pair of cellcolumns lying along the sides of the aorta and along the dorsal surfaces of the carotid arteries. The anlagen of the secondary sympathetic trunks arise about the beginning of the sixth day, as ganglionic enlargements on the median sides of the spinal nerves. These ganglionic enlargements are at first independent of each other, but become united later by longitudinal commissures. The primary sympathetic trunks reach their maximum development during the course of the sixth day, after which they decrease in size until they disappear. The observations just summarized agree essential^ with the results of His, Jr.


306 ALBERT KUNTZ

2. The author finds, however, that the cells giving rise to the sympathetic trunks are not derived exclusively from the spinal ganglia, as His, Jr., supposes, but that they are derived, wholly or in part, from the neural tube. Medullary cells migrate from the neural tube into the ventral roots of the spinal nerves. With similar cells which wander out from the spinal ganglia, these cells migrate peripherally along the spinal nerves. At a point a little above the level of the aorta, cells deviate from the course of the spinal nerves and, migrating toward the aorta, give rise to the primary sympathetic trunks. As migration proceeds, the cells which deviate from the course of the spinal nerves no longer wander into the primary sympathetic trunks, but become aggregated at the point of origin of the communicating rami and give rise to the anlagen of the secondary sympathetic trunks.

3. The prevertebral plexuses arise as cell-aggregates lying along the ventro-lateral aspects of the aorta from the suprarenals posteriorly. They are derived directly from the primary sympathetic trunks.

4. The ganglion of Remak arises as an oval cell-column lying in the mesentery just dorsal to the rectum. It arises from cells which the author finds to migrate ventrally from the hypogastric plexus.

5. The cardiac plexus and the sympathetic plexuses in the walls of the visceral organs, which the author has designated as the " vagal sympathetic" plexuses in an earlier paper, arise from cells which migrate from the hind-brain and the vagus ganglia along the fibers of the vagi. In the posterior region of the intestine, the myenteric and the submucous plexuses probably receive some cells from the ganglion of Remak.

6. The cells which migrate from the neural tube and from the cerebro-spinal ganglia along the spinal nerves and the vagi are the descendants of the "germinal" cells of His; viz., the "indiffer


SYMPATHETIC SYSTEM IN BIRDS 307

ent" cells and the " neuroblasts" of Schaper. They are, therefore, homologous with the cells which give rise to the neurones and the neuroglia cells in the central nervous system, and the sympathetic neurones are homologous with the afferent and the efferent com ponents of the other functional divisions of the peripheral nervous system. These observations agree with the author's observations on mammalian embryos.

7. Certain morphogenetic differences exist in the development of the sympathetic nervous system in birds and mammals, which the author interprets as indicating that the sympathetic system has departed more widely from the ancestral type in birds than in mammals. Such departure is no more than should have been expected in the specialized avian branch of the vertebrate series.


308 ALBERT KUNTZ

BIBLIOGRAPHY

Cajal, S. R. Nouvelles observations sur revolution des neuroblastes, avec:

1908. quelque remarques sur l'hypothese neurogenetique de HensenHeld. Anal. Am., vol. 32, pp. 1-25 and 65-78.

Froriep, A. Die Entwickelung und Bau des autonomen Nervensystems. Med. 1907. naturwiss. Archiv., vol. 1, pp. 301-321.

Held, H. Die Entwickelung des Nervensgewebes bei den Wirbeltieren. Leipzig. 1909.

His, Jr., W. Ueber die Entwickelung des Baucksympathicus beira Hiihnchen 1897. und Menschen. Archiv Anat. u. Entwg. Supplement.

Kuntz, A. A contribution to the histogenesis of the sympathetic nervous system.

1909. Anat. Record, vol. 3, pp. 158-165.

1909. The role of the vagi in the development of the sympathetic nervous

system. Anat. Anz., vol. 35, pp. 381-390.

1910. The development of the sympathetic nervous system in mammals.

Jour. Com. Neur. and Psy., vol. 20, no. 3, pp. 211-258

1910. A comparative study of the development of the sympathetic nervous system in birds and mammals. Abstract of paper read before the American Society of Zoologists, Central Branch. Science, n.s., vol. 31, p. 837.

Marcus, H. Ueber den Sympathicus. Sitzungsb. d. Gesell. f. Morph. u. Physiol. 1909. in Miinchen, pp. 1-13.

Onodi, A. D. Ueber die Entwickelung des sympathischen Nervensystems. Archiv 1886. /. mikr. Anat., vol. 26, pp. 555-580.

Accepted by the Wistar Institute of Anatomy and Biology, June 10, 1910. Printed, September 14, 1910


The Origin Of The Cranial Ganglia In Ameiurus

F. L. Landacre

Eighty-Eight Figures

CONTENTS

Introduction 309

Historical sketch 312

Material and method 322

The differentiation of the neural plate 325

The differentiation of the Gasserian ganglion 330

The origin of the lateralis Vllth ganglia 332

The auditory vesicle and auditory ganglion 334

The fate of the preauditory placode 339

The origin of the geniculate ganglion 342

The differentiation of the postauditory lateral mass 349

The origin of the lateralis Xth and the early stages of the postauditory placode 353 The fate of the postauditory placode and the appearance of the lateral line

organs of the body 355

The placodal ganglion of the IXth nerve 362

The origin of the lateralis IXth ganglion 369

The origin of the third and fourth epibranchial placodes . 373

The origin of the fifth and sixth epibranchial placodes 375

The later history of the lateral mass ganglia of the Xth 378

General summary 384

Bibliography 389

Abbreviations 392

Explanation of figures 394

INTRODUCTION

The origin and relationship of the cerebral nerves have been, since the appearance of Balfour's classical researches on the nerves of Elasmobranchs in 1876, among the most interesting and puzzling problems in vertebrate morphology. Notwith


310 F. L. LAND ACRE

standing the enormous amount of labor expended on this problem, it can hardly be said that there is any general agreement among workers as to the mode of origin or fundamental relationships of the various cerebral nerves to each other.

The lack of agreement may be traced to several causes aside from the difficulty inherent in the problem. Chief among these causes must be placed the subordination of the study of the cerebral nerves to more general problems, such as the metamerism of the vertebrate head and the relation of the head to the trunk. These fundamental problems have been a source of intense interest to both morphologists and embryologists and it was only natural that the cerebral nerves should be used as a means of shedding light on them, even when the origin, composition and distribution of the nerves were not accurately known.

Owing to this method of approach, however, students of the cerebral nerves were for a long time handicapped by preconceived notions as to the relations which ought to exist between the trunks of the cerebral nerves and the various head segments. The idea of the serial homology of the cerebral nerves has been in a sense a barrier to a clear understanding of their true relationships. The chief function of the nervous system as a coordinating mechanism was lost sight of in the attempt to reduce the cerebral nerves to some fundamental type that would bring them into harmony with each other and with the spinal nerves.

Some such fundamental relationship among the cerebral nerves as that involved in the idea of serial homology probably existed in a primitive vertebrate; but the cerebral nerves of the Ichthyopsida are not most easily understood on such a basis. The fundamental needs of the organism have so modified this primitive type that it is much easier and more in accord with the facts to take some functional unit, such as a sensory system, as a basis for the analysis of the cerebral nerves.

The fundamental difficulty in the attempt to homologize the trunks of the cerebral nerves comes from the fact that they are not units but composite in nature, so that the components of which they are made up furnish the logical basis for getting at their true relationships.


THE CRANIAL GANGLIA IN AMEIURUS 311

The realizacion of this fact and its use as a working basis in the attempt to determine the relationships of the cerebral nerves is the most striking characteristic of recent work on the nervous system of the lower vertebrates. The recognition of functional units rather than serial homologies as the keynote in this work has tended to free the study of the cerebral nerves from subordination to larger problems and to call attention to the need as well as to the possibility of a more thorough knowledge of the composition of the nerves before this knowledge can be applied to such problems as the metamerism of the head or to the relation of the head to the trunk. It has also served to bridge the gap that has existed between the structural and functional conceptions of the nerves by adopting a unit that is not only structural in character but functional as well.

Such an analysis of the cerebral nerves has been possible, owing to the fact that, among other things, there are structural differences between the various components that have made it possible to follow them throughout their whole course in favorable types. The analysis has been materially strengthened of course by a clearer understanding of the various types of end organs to which the fibers are distributed peripherally as well as of the central connections in the brain and cord.

The net result of this work seems to have changed the basis for homologizing the cerebral nerves from the nerve trunks to the nerve components. These are anatomical and physiological units characterized by similarity in function, peripheral distribution and central connections; but they are not uniformly distributed throughout the various cerebral nerves even in closely related types. Nerves with the same name and the same general topographical relations may vary totally in composition. This fact makes it evident that there could be no agreement as to the homology of the cerebral nerves as long as their exact composition was unknown.

A more thorough knowledge of the components of which the cerebral nerves are made up and of the composition and position of the ganglia from which they arise opens the question as to the origin of the discrete elements of these ganglia. Are


312 F. L. LANDACRE

they as distinct from each other in their mode of origin as they are in their structure and function in the adult?

Even a cursory reading of the literature on cerebral nerves shows a well defined movement away from the idea of concreteness of origin and toward that of discreteness of origin. Recently however, so far as I am aware, no attempt has been made to follow out carefully the development of the cerebral ganglia in some type in which the nerve components are known. The present paper is an attempt to fill this gap in our knowledge by tracing the various components back to their earliest recognizable stages, keeping in mind the fact that the origin of a definitive ganglionic mass of known composition and function in the adult is the end in view. The attempt to account for the fate of all embryonic cell masses, a part of which may go to form ganglia, presents a totally different problem and one with which we are unable to cope with present technical methods.

I wish to express my gratitude to Dr. CO. Whitman for his friendly criticism and guidance during the progress of this work.

HISTORICAL SKETCH

The following brief historical account aims to give only the important advances by which our knowledge of the discreteness in origin and structure of the cranial ganglia and nerves has grown.

According to Balfour ('75), before the appearance of his work on the spinal ganglia in Elasmobranchs, it was the prevailing opinion that the ganglia were derived from the mesoblast of the vertebrae. His had in 1868 ascribed their origin to the epiblast but his work seems not to have been confirmed up to the time Balfour published. Balfour, as is well known, traced the ganglia to the cord but believed that his work diverged from that of His no less than from that of his own predecessors. Goette and Semper, writing in 1875, however, had attributed the origin of the nerves to the epiblast and Goette had extended this observation to the head. In 1882 Van Wijhe and Hoffmann, working on teleosts, confirmed Goette's work and Van Wijhe showed that the neural


THE CRANIAL GANGLIA IN AMEIURUS 313

crest cells fuse with the lateral ectoderm in two places. This was confirmed for amphibia in 1886 by Misses Johnson and Sheldon, who like Van Wijhe found that the dorsal fusion is connected with the lateral line.

Tn 1885 Froriep described in mammals, what he designated as branchial sense organs on the facialis, glossopharyngeus and vagus nerves. The epidermal thickenings which he figures and designates as branchial sense organs are undoubtedly epibranchial placodes since they lie immediately over the gill slits and there are no lateral line nerves in the mammals. Froriep, however, thinks that the resemblance of these anlagen to those of the lateral line anlagen places them in that group of nerves. He further finds that the evidence for the addition of cells from the placode to the ganglia in these nerves is slight, since only in the earlier stages are the boundaries between the two structures indistinct.

Beard ('85, '87, '88) described primitive branchial sense organs in elasmobranchs, teleosts, amphibia, reptiles and birds and also stated that the epidermal thickening contributed cells to the ganglia derived from the neural crest. Beard rejected the term lateral line organ because the lateral lines originate in the head and substituted therefor the term branchial sense organ, and he seems always to have had in mind the lateral line organs of the head when he speaks of branchial sense organs; at least he confuses the dorso-lateral and the ventro-lateral placodes and, as von Kupffer ('91) intimates, it is impossible to tell which he is discussing at times. He figures and mentions, however, two points of contact of the neural crest ganglia with the skin.

Beard ('88, p. 882) refers to the fusion of the neural crest with the epiblast at the level of the notochord, and just above the gill cleft in the same paragraph, apparently not distinguishing between them, although in the same paper he criticizes Onodi for not having seen the second fusion of the ganglion anlage with the epiblast. It seems a safe conclusion from Beard's papers that he took all fusions of the neural crest to be concerned in the formation of the lateral nerves.

In discussing the origin of the neural crest portion of the ganglia Beard ('88) insists that the origin of the spinal ganglia from the

THE JOURNAL OF COMPARATIVE NEUROLOGY AND PSYCHOLOGY, VOL. 20, NO. 4.


314 F. L. LANDACRE

Zwischenstrang as described by His is not correct but that in all cases he finds the ganglion arising from a mass of cells derived from the epiblast at the point of the entering angle between epiblast and cord and that the Zwischenstrang takes no part in its formation and is present after the ganglion has detached itself from the epiblast and that the ganglion anlage can always be detected before the closing in of the medullary plates. Beard failed, however, as his predecessors had failed, to distinguish between the dorso-Iateral placodes and the early stages of the adult lateral line organs. Under the term branchial sense organs he may have been describing either lateral line organs in the head, or dorso-lateral placodes which may give rise to ganglion cells. Von Kupffer, working on Petromyzon, in 1891 made a sharp distinction on the one hand between dorso-lateral placodes or those placodes concerned in the origin of the lateral line and lying at the level of the notochord, and on the other hand the epibranchial placodes which arise just over the gill slits. He also states that the epibranchial ganglia are concerned in the origin of the branchial nerves, so that one may infer from his work that there were separate origins for what we now designate as lateralis and visceral sensory nerves.

Miss Piatt ('95) showed that in Necturus the lateral ectoderm gives rise not only to the lateral line ganglia and nerves which appear in three primitive longitudinal ridges, but also to mesectoderm, cells proliferated from the lateral ectoderm and from the neural crest and assuming the position and characteristics of mesoderm. Miss Piatt seems first to have made a distinction between the earliest thickening of the lateral epidermis, i.e., the dorso-lateral and epibranchial placodes, and the early stages of the lateral line organs. These seem to have been confused by previous workers. She says (p. 500) that "the large dorso-lateral and epibranchial ganglia are formed from cells that split off en masse leaving the ectoderm external to them for the time thin. A sensory ridge may appear later in the exact place where the ganglion arose, as happens in the supra-orbital line, or sense organs may form at either side of the ganglionic anlage, as in the vagus region." She calls attention (p. 497) to the fact that the growing point of


THE CRANIAL GANGLIA IN AMEIURUS 315

the dorso-lateral placode may simulate very closely a lateral line organ (fig. 5), but that it does not give rise directly to these organs since this growing point is found in segments where later there are no lateral line organs.

Miss Piatt's work marks in a way the culmination of the work begun by Balfour in 1876. A more thorough analysis of the cerebral nerves of the adult was needed before the full significance of the dorso-lateral and epibranchial placodes could be determined. While she seems to place the dorso-lateral and epibranchial placodes in the same category, the determination that not all cells proliferated from the neural crest and epidermis go to form ganglia and that the dorso-lateral placodes do not in some cases give rise directly to sense organs, but that these appear later, mark distinct advances in our knowledge of the origin of structures derived from the epidermis. Much confusion has arisen apparently from ignorance of these facts.

Locating a specific ganglion of known composition in the adult and then tracing it to its origin would have prevented errors that have arisen unavoidably from pursuing the opposite course and rinding a segmental ganglion and nerve for every area in which cells are being proliferated from the neural crest and ectoderm. As a general conclusion from her work, she thinks that the root of a sensory nerve is no index to the segmental value of that nerve, the position of the nerve root being in a great measure the expression of the coordinate relations which the central nervous system serves. Miss Piatt emphasizes another point of the greatest importance; it is in connection with the distinction between definitive ganglion cells and nerve fibres on the one hand, and the anlagen from which these come on the other. Just as not all cells derived from the ectoderm go to form ganglia and nerves, so not all cells grouped about ganglia and the growing points of nerves are concerned directly in the nervous functions of the nerves and ganglia.

The appearance of Strong's paper in 1895, and more particularly of Herrick's in 1899, mark as it appears to the writer, a turning point in the study of the cerebral nerves. Evidently no homology of the cerebral nerves can be established without an exact knowl


316 F. L. LANDACRE

edge of the central ending, ganglionic relations, and peripheral distribution of the various components of these nerves. Confusing acustico-lateralis ganglia with communis or visceral ganglia or failing to distinguish between perfectly discrete regions in the brain leads only to confusion. The fact that in many types a difference in size of fibres, in addition to the characteristics mentioned above is present, has made possible the determination of the exact composition of the trunks of the cranial nerves and has made unnecessary the almost interminable and confusing discussions of their homologies. The determination by Herrick ('99) that there are in the Vth, Vllth, VHIth, IXth, and Xth nerves of Ichthyopsida three, and apparently only three, chief groups of sensory components and the exact definition and description of these components and their ganglia marks in a general way the advance in our knowledge of cranial nerves since Strong's paper appeared.

The three components mentioned above are, first, the general cutaneous, characterized by ending in the brain in the spinal fifth tract, by having ganglia (in Menidia) in the Vth and Xth nerves situated intracranially, by having medium sized fibres and by being distributed peripherally to the skin as free nerve endings. Second, the acustico-lateralis, or special cutaneous, characte ^ized by ending in the brain in the tuberculum acusticum, by having ganglia in the Vllth, VHIth, and Xth nerves, by having large fibres and by being distributed to the ear and lateral line organs only. Third, the communis or visceral sensory system characterized by ending in the brain in the nucleus of the fasciculus solitarius or its equivalent in the vagal, glossopharyngeal, and facial lobes, by having ganglia in the Vllth, IXth and Xth nerves, by having small sized fibres and by ending peripherally in taste buds wherever found, and general mucous surfaces. While this paradigm holds for teleosts generally, the irregularity with which some of these components are distributed in the various cerebral trunks makes it clear why there was so much confusion and disagreement in the earlier attempts to determine the homologies of these trunks.


THE CRANIAL GANGLIA IN AMEIURUS


31:


A comparison of Koltzoff's excellent paper ('02) on the embryology of Petromyzon with the work of Johnston ('05 b) on the nerve components of the same type will serve to introduce the point of view from which the author has worked the early stages of the ganglia in Ameiurus. KoltzofT does not list a single nerve component paper in his bibliography and was in fact working primarily on the segmentation of the head, and only incidentally on the origin of the ganglia. The results of these two papers are summarized in the following diagram :


TABLE I


Summarizing the results of Koltzoff's work on the development of the ganglia in Petromyzon, and of Johnston's on the adult structure of the same form


PROF. OR TRIG. OR

TRIG. I. V TRIG. II. V


Neu. Crest.. Neu. crest Neu. crest Neu. crest Neu. crest Neu. crest Neu. crest

Segmental


sc .K

o <-> o ~~^


I { D. L. Plac. . D.L. Plac. D.L. Plac. D.L. Plac. D.L. Plac. D.L. Plac. D.L.Plac.


Ep. Plac Ep. Plac.


Ep. Plac.Ep. Plac. Segmental


§ a S S

co o


General Prof. N.

Cutaneous . and Gan..


Trig. N.Gen. Cut.


Gen. Cut. Gen. Cut. X.


and Gan.. VII IX.


Segmental


Lat. X.

Acustico- Lateralis Lateralis Lateralis Lateralis Not

Lateralis . . . Prof Trig VII Auditory. IX Segmental


Communis [ or Visceral Com. VII.


Com. X. Com. IX. Segmental


It will be seen at a glance that every nerve into whose ganglion cells derived from the epibranchial placode enter contains in the adult visceral fibres; every nerve to whose ganglion the dorso


318 F. L. LANDACRE

lateral placodes contribute cells contains acustico-lateralis fibres in the adult; and, finally, every nerve, except the eighth, to whose ganglion the neural crest contributes cells contains in the adult general cutaneous fibres.

If it were not for the single exception in the case of the VHIth nerve, a very natural conclusion from this comparison would be that the epibranchial placodes gave rise to visceral ganglia, the dorso-lateral placodes to acustico-lateralis ganglia and the neural crest to general cutaneous ganglia. The error in this conclusion arises from the assumption that the epibranchial and dorsolateral placodes give rise to all of the ganglia into which they enter. This may occur in some cases but certainly does not in all. There is a very general agreement among workers on the origin of the cerebral ganglia that the neural crest cells fuse with cells derived from either the dorso-lateral placodes or the epibranchial placodes or both, so that the assumption that these placodes give rise to all of the ganglia into which they enter is not only unwarranted but contrary to evidence presented in a large number of cases. If neural crest cells enter into the auditory ganglion of Petromyzon, which is the most specialized of all the acustico-lateralis ganglia, it would not be surprising if they should enter into the less specialized ganglia, such as those of the Vllth nerve. An examination of the literature on the composition of the auditory ganglion shows a great deal of diversity of opinion as to the presence of the neural crest in the region of the auditory vesicle. Johnston ('06) states that it is absent and Rabl ('92) and Hoffmann ('94) both take the same position. On the other hand, Beard ('88), Van Wijhe ('82), von Kupffer ('94), Piatt ('95) and Neal ('97) hold that it is present. There seems little doubt that it is present in some types in the earlier stages at least. It does not follow from this fact that neural crest cells enter into the auditory ganglion. Dohrn ('90) holds that the neural crest is at first continuous between the acusticofacialis and the glossopharyngeus, but later disappears, that it is sometimes present on both sides and sometimes only on one side and that the masses of cells which he sometimes finds are to be interpreted as remnants of the former connections of preauditory and postauditory neural crest. The fact that there are general


THE CRANIAL GANGLIA IN AMEIURUS 319

cutaneous fibres in the Vllth in Petromyzon, that the crest is at first continuous but apparently not so later, and that when the Vllth and IXth contain general cutaneous fibres they also contain neural crest cells seem to the writer to minimize the discrepancy between Kolzoff's and Johnston's work, since these cells may enter into the Vllth and IXth in Petromyzon. The matter is complicated by the fact that neural crest cells of the Vllth lie between the anterior portion of the auditory vesicle and the neural tube and come into contact with the vesicle at or near the place where the cells of the auditory ganglion are proliferated from the vesicle so that the relations are confused and it is difficult to be certain of the exact conditions. Leaving out of consideration the VHIth ganglion, we can infer that the general cutaneous ganglia come exclusively from the neural crest, while the acustico-lateralis and visceral ganglia come in part from the dorso-lateral and epibranchial placodes respectively but may contain neural crest cells in some cases.

Without anticipating the conclusions drawn from a study of Ameiurus, it may be well to call attention to two facts: first, the acustico-lateralis system of nerves and ganglia is usually treated as a special cutaneous system on account of the fact that the tuberculum acusticum which is the acustico-lateralis center in the medulla oblongata is a specialized portion of the general cutaneous center of the spinal cord extending into the medulla oblongata. The close relationship of the general cutaneous and acusticolateralis systems is thus based on good anatomical evidence. Second, the communis or visceral system is really double in character; it consists of a general visceral portion which supplies general mucous surfaces, and it also contains a special visceral or gustatory portion whose fibres end peripherally in taste buds, so that the presence of neural crest cells in the acustico-lateralis and visceral ganglia might be explained in the first case on the basis of the close relationship between the general cutaneous and the acusticolateralis systems of nerves and ganglia and in the second case on the basis of the double composition of the communis or visceral system.

The determination of the relation of neural crest cells to those


320 F. L. LANDACRE

cells derived from the dorso-lateral and epibranchial placodes was one of the chief objects in taking up the study of Ameiurus. Unexpected difficulties were met, owing to the fact that Ameiurus has no well defined neural crest in the head such as is found in other types, so that my conclusions are based on the assumption that there are in Ameiurus cells which correspond to the neural crest of other types. Aside from this peculiarity, it has proven to be an exceptionally favorable type. The epibranchial placodes are large and easily followed and the epibranchial ganglion of the IXth nerve seems to be a pure placodal ganglion, a fact which enables us to come to a definite conclusion in regard to the relation of the neural crest cells to those derived from the placode. It seems to be necessary to determine this point definitely before we can come to any satisfactory conclusion as to the relation of the spinal and cerebral ganglia to each other. The acustico-lateralis and gustatory nerves are peculiar to the head but the general visceral and the general cutaneous are not, being found in the trunk. The problem of the relation of spinal and cerebral nerves becomes largely a question as to how the general visceral and general cutaneous nerves are related in origin to the special visceral and special cutaneous nerves in the head. Aside from the fact that it was necessary to choose some type in which the nerve components are known, the chief reason for taking Ameiurus lies in the enormous size of the gustatory system of sense organs and nerves. The taste buds are distributed over practically the whole body surface as well as in the mouth* pharynx and oesophagus, and gustatory fibres a/e found in ten out of twelve of the chief nerves arising from the trigemino- facial ganglia. A type with such an enormously hypertrophied gustatory system seemed likely to furnish a good basis for determining the exact mode of origin of the special visceral ganglia and nerves.

A summary of the cranial ganglia of Ameiurus as given by Herrick ('01) is shown in the following diagram:


THE CRANIAL GANGLIA IN AMEIURUS


321


TABLE II

Showing the components of the Vth to Xth ganglia in Ameiurus; compiled from Herrick's ('01) description. The presence of any component is indicated by "pres," ' or by the name of the ganglion. The jugular and lateralis Xth are placed over the third and fourth epibranchial ganglia of the Xth not to indicate their segmental position but because this is their relative position.



V


VII


VIII


IX


X


X


X


X


Gen Cut.. . .









Jugular Pres.


Acus. Lat. . .



Lat. VII. Pres.


Pres.


Pies.





Communis . .



Pres.


Pres.


Pres.


Pres.


Pres.





This table is slightly different from Herrick's description. There is a lateralis ganglion in the region of the IXth nerve which he seems to attribute ('01, p. 208) morphologically to the Xth, although it is associated with the root of the IXth. I find it to be entirely distinct in origin from the Xth nerve and have placed it in the diagram with the IXth. The general cutaneous ganglia are found in the Vth and Xth nerves. The acustico-lateralis ganglia are found in the Vllth, where there are two divisions, the dorso-lateral and the ventro-mesial ; and in the Vlllth, IXth, and Xth. The visceral ganglia are found in the Vllth, IXth and in the four branchial ganglia of the Xth. The jugular or general cutaneous Xth and the lateralis Xth are placed in the diagram over the last epibranchial ganglion of the Xth, not to indicate their segmental position but because they form the posterior portion of the vagus ganglion. Attention will be called later in the body of the paper to the fact that this table makes no distinction between the general and special visceral ganglia, both being catalogued as the visceral sensory or communis system, although Herrick distinguishes the two groups functionally and further distinguishes the two types of fibres structurally.


322 F. L. LANDACRE

MATERIAL AND METHOD

In a type which develops so rapidly as Ameiurus, there being only five days between fertilization and hatching, the age seems to be a better way to designate a series than the body length. Series must be taken at close intervals and, while individuals of a given age may not vary by an appreciable amount in length, there are frequently found in the same series quite perceptible variations in the degree of differentiation of the ganglia and sense organs, so that while the age is not an absolutely accurate method of designating a series since the growth varies with temperature, I have used this on account of the difficulty of separating embryos by their length.

As to method, I have followed consistently the plan of locating definite ganglia in older series after they were well defined and tracing these back to the earliest recognizable stages. This plan seems to be absolutely necessary, since only in a few cases do the definitive ganglia use all of the material from which they are formed and in some cases, particularly the general cutaneous ganglia, only a very small portion of the mass from which the ganglion is formed finally enters into the composition of the ganglion. Some confusion seems to have arisen, especially in the earlier work on nerves and ganglia, from taking it for granted that all of the material from which a ganglion forms enters into a ganglion.

The material on which this work was done is in part, the same as that used in a former paper by the author (Landacre, '07) and consists of series of Ameiurus melas of which the absolute age is known since the process of oviposition was observed; and in part, particularly the early stages, the work was done on a large number of graded series of Ameiurus nebulosus of which the relative age was known but the absolute age was not known. The series of Ameiurus melas are indicated by their ages. Of the nebulosus material nine stages were used of which the last two series, VIII and IX, are quite similar to each other and correspond closely to the forty-nine hour series of A. melas. Series I seems to be about twenty-four hours old compared with a similar stage of A. melas. The remaining series II to VII inclusive are all younger


THE CRANIAL GANGLIA IN AMEIURUS 323

than the youngest of the A. melas series, which was forty-nine hours old. In the following brief characterization of the stages the optic cup, the optic lens, and the auditory vesicle are used chiefly as a means of separating them in addition to the more minute differences to which attention is called in the body of the paper.

Stage I. An early stage (24 hours?) of A. nebulosus in which the blastoderm is nearly flat, the neural keel being elevated slightly above the yolk at the anterior end only. The blastoderm has not extended sufficiently far posteriorly and ventrally to ' be cut on the ventral side of the yolk sac at the posterior end of the embryo. Fig. 1 was drawn from this stage.

Stage II. Optic vesicle present, but solid. Nuclei of cells in optic vesicle uniformly distributed with no indication of the peripheral arrangement of the nuclei which precedes the formation of a cavity. Nuclei of auditory vesicle irregularly arranged with no indication of a cavity. The auditory vesicle can be located only by comparison with an older series. Lateral mass not broken down into mesectoderm at any point. Figs. 2 and 3 are drawn from this stage.

Stage III. Optic vesicle solid but with nuclei of cells uniformly arranged around the periphery of the vesicle preparatory to the formation of the optic cup. Auditory vesicle with its cells elongated and the nuclei arranged around the periphery. The lateral walls of the cup not in contact, the future cavity of the cup filled with spherical cells. The solid lateral mass of the preceding stage changed into a loose mass of tissue in the region of the Gasserian ganglion, particularly just anterior and posterior to this point. The pre- and postauditory placodes are present. Figs. 4, 5, 6, 7, 8, 9 were drawn from this stage.

Stage IV. Optic vesicle with slight cavity. Epidermis not thickened preparatory to the formation of a lens. Auditory vesicle with a slight cavity and the lateral walls of the vesicle in contact. Preauditory placode just detached fom the auditory vesicle. Postauditory placode still in contact with the auditory vesicle. Gasserian and geniculate ganglia not sufficiently well defined to determine approximately their boundaries. Hyoid gill pocket


324 F. L. LANDACRE

not yet in contact with the epidermis. Figs. 14, 15, 16, 44, 45, 46 were drawn from this stage.

Stage V . Optic vesicle and optic stalk completely open. Lens indicated by slight thickening of the epidermis. Auditory vesicle with well defined cavity and the proliferation of cells from the vesicle to form the ganglion beginning. Gasserian and geniculate ganglia distinguishable from the loose lateral mass cells surrounding them, although their boundaries cannot be definitely determined. Hyoid gill pocket in contact with the epidermis. First stage in the proliferation of cells from the auditory vesicle to form the lateralis IX ganglion. Hyoid gill pocket in contact with the epidermis. Figs. 10, 11, 13, 17, 18 and 63 were drawn from this stage.

Stage VI. Lens thickening well defined but changing gradually at the borders into epidermis, i.e., not sufficiently thick to have its borders well defined. First epibranchial placode appears. Hyoid gill pocket still in contact with the epidermis. Figs. 19, 20, 21, 22, 23, 24 were drawn from this stage.

Stage VII. Lens constricted sharply at its borders preparatory to its detachment from the epidermis. Nuclei of cells in lens not yet located in the periphery of the cells. First epibranchial placode is proliferating cells into the mesoderm to form the first epibranchial ganglion but the ganglion is still in contact with the epidermis. The jugular ganglion of the X can be located. The lateralis X ganglion is in process of being proliferated from the postauditory placode. Hyoid gill pocket still in contact with the epidermis. Figs. 12, 25, 26, 27, 28, 29, 30, 40, 41, 42, 43, 47, 48, 49, 50, 51 were drawn from this series.

Stages VIII — IX correspond closely to a forty-nine hour embryo of Ameiurus melas being possibly slightly younger. There seems to be no well defined distinctions between stages VIII-IX or between these and A. melas forty-nine hours. The three figures drawn from these two series could have been taken from A. melas. Fig. 61 was drawn from Stage VIII and figs. 52, 62 from Stage IX.

All sections were cut 7n thick and it will be convenient to express the spatial relations of structures in sections.


THE CRANTAL GANGLIA IN AMEIURUS 325

THE DIFFERENTIATION OF THE NEURAL PLATE

Ameiurus presents a rather striking difference from the descriptions usually given of the formation of the neural cord in its relation to the origin of the neural crest and dorso-lateral placodes.

Balfour ('75) described the neural crest as growing out of the neural cord but seems not to have worked stages sufficiently early to determine its exact mode of origin. Marshall (77) gives substantially the same description of its origin. Beard ('88), however, describes it as arising in elasmobranches, teleosts, amphibians, reptiles and birds as a thickening of the epidermis, lying lateral to the neural plate and always distinguishable from that structure before the neural tube is formed. There is substantial agreement among all the earlier descriptions of the cerebral ganglia in attributing to the neural crest a definite structure related more or less closely to the dorsal portion of the neural plate as it folds off to form the neural cord (Harrison '01, textfigs. 1-5) . Few authors, so far as I am aware, have found any close relation between the neural crest and dorso-lateral placodes in their earliest stages. The neural crest ganglion is almost always described as growing down ventrally from its point of origin and coining into contact with the epidermis at the point of origin of the dorso-lateral placode on a level with the notochord. Wilson and Mattocks ('97, p. 659) do, however, describe the dorso-lateral placode of the salmon as arising not by a thickening of the epidermis but by the thinning out of the neural shield which leaves the placode isolated, lying in a lateral position. In the salmon the placode is at first located in the lateral portion of the neural shield just as in Ameiurus.

The early stages of Ameiurus differ from the usual descriptions in that the neural crest and dorso-lateral placodes are not differentiated from each other at first but appear as a large lateral thickening lying on either side of the neural plate. This I have designated as the lateral mass (fig. A.) It doubtless contains regions comparable to the neural crest and certainly to the dorso-lateral placodes of other authors; but since these are not recognizable in the early stages and in fact the neural crest is never recognizable


326


F. L. LANDACRE


as a structure distinct from the lateral mass, and since the lateral mass contains much material that does not go to form ganglia before undergoing differentiation and has a definite structure and position of its own, it seems better to characterize it as indicated above.

The lateral mass in an early stage in which the neural plate is still nearly flat has the appearance indicated in fig. 1. In Stage




dor./atma.ss vu. ■pre. au. p/aoode


D


Fig. A. A camera tracing of the blastoderm corresponding to fig. 1, showing the broad, flat neural plate.

Fig. B. A camera tracing corresponding to fig. 3. showing the position of the lateral mass and of the intermediate region.

Fig. C. A camera tracing four sections anterior to fig. 9, showing the relation of the lateral mass to the neural cord.

Fig. D. A camera tracing corresponding to fig. 9, showing the relation of the preauditory placode to the dorso-lateral mass. Posterior to this point of a few sections, the auditory vesicle occupies the position of the preauditory placode and dorso-lateral mass.


II (fig. B), it is separated from the neural keel by a thinner region which I shall designate as the intermediate region to distinguish it from the lateral mass. This thickening extends throughout the whole length of the head and into the cord region. The lateral mass does not become incorporated into the neural cord which forms from the thick central mass, or neural keel, lying between the intermediate regions on either side.


THE CRANIAL GANGLIA IN AMEIURUS 327

As the neural keel deepens and assumes the form of a cord (fig. C), and as the blastoderm rises on the yolk and assumes a rounded form, the lateral cell masses are brought gradually into a lateral position, still retaining their connection with the dorsal half of the cord by an intermediate slightly constricted area.

In an embryo in which the optic vesicle has reached a stage in which the future optic cup is slightly larger than the stalk (Stage II), this lateral thickening begins some five or six sections posterior to the stalk and extends from this point back beyond the region in which the Xth ganglion is later formed, more than one hundred sections. Posterior to this point where the keel is forming in the region of the spinal cord it gradually becomes reduced in size as the keel becomes shallower. Throughout this whole region the lateral cell mass has a structure quite uniform at first, varying only in shape, being somewhat thicker and more closely applied to the sides of the brain in the anterior region (fig. 2), and somewhat thinner and more dorsally attached to the cord in the posterior region, particularly posterior to the position in which the auditory vesicle develops.

Six sections posterior to the optic vesicle (fig. 2) , the lateral mass is applied to the dorsal half of the cord and is homogeneous in structure. This section lies in the region anterior to that in which the Gasserian ganglion later appears. In the region in which the Gasserian ganglion forms and posterior to it (fig. 3) , the lateral mass is broader and thinner and the attachment to the neural cord is less extensive. This condition of the lateral mass persists throughout the region in which the Gasserian ganglion forms and back of this until we come to the region just anterior (fig. 4) and just posterior (fig. 5) to the auditory vesicle, where, in a slightly older embryo (Stage III), there is soon noticeable a slight differentiation of the lateral mass into a thicker dorsal portion, the dorso-lateral mass (D. L. M., figs. 4 and 5), connected with the cord by the intermediate region, and a ventral mass (Pre. PL, fig. 4, Post. PL, 5) slightly separated from this dorsal mass on its mesial border by a constriction. This ventrally differentiated mass, or dorso-lateral placode (pre. au. placode, fig. D) is present throughout the whole auditory region (fig. 6) J


328 F. L. LANDACRE

and extends somewhat anterior and posterior to the auditory region, where it becomes merged completely with the dorso-lateral mass to form the lateral mass. Posterior to the auditory vesicle, in the region of the IXth nerve, it presents the appearance shown in fig. 5. This ventrally differentiated mass shown in figs. 4 {Pre. PL), 5 (Post. PL) and 6 (Au. Ves.) develops later into the auditory vesicle and the pre- and postauditory placodes (dorsolateral placodes).

The fate of the lateral mass, as a whole, varies in different regions of the head. In the regions between the optic stalk and the Gasserian ganglion it becomes converted entirely into mesectoderm. Fig. 7 is taken from an embryo slightly older (Stage III) than that from which figs. 2 and 3 were taken, and is identical in position with fig. 2, with which it should be compared. The lateral mass is here free from the cord on its mesial border nearly to the dorsal surface of the cord; while on its lateral border it is free from the epidermis up to about the same level. The ventral two-thirds of the mass is converted into a rather loose mass of mesectoderm in which the cell boundaries are indefinite and in which there are numerous intercellular spaces.

The later history of this mass shows that it is converted completely into a very loose mass of mesectoderm with large intercellular spaces and with faint cell boundaries, but with well defined nuclei. I have detected during this change of the lateral mass into mesectoderm no mitotic figures in any of my sections. Posterior to the region in which the Gasserian ganglion forms and between that ganglion and the lateralis Vllth the lateral mass is converted chiefly into mesectoderm, except its ventral border which represents the forward extension of the primordium of the auditory vesicle, or the preauditory placode. Fig. 8 from the same embryo is taken through the region in which the Gasserian ganglion forms. The lateral mass is here detached from the cord mesially, except at its dorsal border, but it is attached to the epidermis throughout its whole length. The dorsal and particularly the dorso-mesial portion of the solid lateral mass is beginning to be converted into a looser cell mass. Fig. 9 is taken just anterior to the auditory vesicle in the position in which the lateralis Vllth ganglion will


THE CRANIAL GANGLIA IN AMEIURUS 329

appear. The dorsal portion of the lateral mass which is still quite solid and has definite cell walls is later converted into the lateralis Vllth ganglion and possibly in part into the anterior portion of the auditory ganglion. The ventral portion of the lateral mass (fig. 9, Pre. PL) is slightly differentiated from the dorsal and represents the preauditory placode. As one reads back in the same series this preauditory placode becomes larger (fig. 4) and the dorsal portion of the lateral cell mass smaller until the auditory vesicle is reached (fig. 6.) The change in size of the preauditory placode is almost imperceptible. It extends farther and farther dorsally until the condition of fig. 4 is reached. The auditory vesicle here has not yet incorporated all the lateral cell mass; at least all the cells of the lateral mass have not yet assumed the radial form with distally arranged nuclei which is so characteristic of the auditory vesicle.

The future lateralis Vllth ganglion is at this time quite large in front where it lies dorsal to the preauditory placode, while posteriorly it becomes smaller and assumes a dorsal and mesial position with reference to the placode. It does not extend posteriorly beyond the anterior end of the vesicle. In the region of the vesicle the whole lateral cell mass is converted into the auditory vesicle.

From this lateral cell mass are differentiated directly, first the mesectoderm lying immediately anterior and posterior to the Gasserian ganglion and posterior to the auditory vesicle. Secondly it gives rise to the Gasserian ganglion and posterior to the ear gives rise to the jugular ganglion. It also gives rise to a large part of the geniculate ganglion, excepting of course the placodal portions, and to the greater portion of the visceral ganglion of the Xth; all of it, in fact, except those portions derived from the third, fourth, fifth and sixth epibranchial placodes. In contrast with these structures which are derived primarily from the lateral mass, we have the VHIth ganglion and the lateralis IXth which come largely, if not exclusively, from the auditory vesicle, and the lateralis Xth which comes exclusively from the postauditory placode. These may be considered as coming secondarily from the lateral mass, the auditory vesicle and postauditory placode representing the primary derivatives from this structure.

THE JO0RNAL OF COMPARATIVE NECROLOGY AND PSYCHOLOGY, VOL. 20, NO. 4.


si

. ■. -. : "

' tIt - - - : - - - - :hr

:--- : r :_:.--; -. : - - _ - . :\... .-- __:_.-:.- : -..

- r _ ■ " . - - - _ - : - ' - -' ■ .- - : - - - : ; r

~: " "

---:..- . ... _. — ' _. - ^Jer^liQw^va'.viiaitdieioese^odefinhafesbeccMiiesonie _ -

- - " :: . . ; :ir

-Tj_zr-i se^*i bii :i ilr _ . r _ - - ~-z ' - - :- - - :izi :" ~~-m_:-:

- -" -^ ~-.~ _r-_ ;:i '. ~J1 ~. .' - ~. :_- „i: - - -1"-:-. : t1. ~»^'~g -.i^t _ l__" i_ :t t-~— Vt -: _ -_._ ." --:'- : ~ — :'i- H-es aJbore tjl^- level of tfae iwiiliU of tihe n-rur-sJ carnaL TThnreig notiirag prwfjt "i""~r i r:-:: r.^z-. 7;_ ~- i - _- - -:_ ;-i: _._ _ .- - :_-;:_ -_;-;m eor_ ridi^niM ' n-riTJL-sr tjl^ r:>;T n.r "i- ~: '.._•:

"-- 2-j - -- :;.i_ -_- i _.__2 _: ---ru : : -i_-2_ ir= .JL-r. : r.v

-r;- :'_-n _7l- :j_ _ -;\ : :»;rd:n ::' --- -_-—;.__ . _. i

-:-i_rr :/_-r ..— v "-- - - :— :t_ v." - -;- > 2. -;.:7^i_:^_. ^-- c— : 7 ^~


-- ■ 7 . ptoplas ith j

nearly all the early ganglionic mas- _ieinrG5. The later i cleaner borders and is quite destine* from

" . . z

■ i - I -- -. . Z2JL2 melas fig. S - ~he eii

quite distil I receding t . l in ?:-r.:ao"

- - . _

-- - rain a? - !

■ ■ - . .. -- very _ 7

be no doubt that it does Dot us

_

honiolc g be Denial er Che origin of

- - :: :. 7 : - . - - s to be doe to the theses tares I

gesof 1 :he

• rial mass beeom _--..-.-■ . -- .

ei :r::

rial mass although in



332 F. L. LANDACRE

THE ORIGIN OF THE LATERALIS VII GANGLIA

The origin of the lateralis Vllth ganglion resembles closely the origin of the Gasserian with the exception that the lateral mass (figs. 4 and 9, Stage III) giving rise to the lateralis Vllth never in my series breaks down so completely into a loose mass of cells as does the Gasserian. There is no break in continuity between the lateral mass cells and the ganglion. Posterior to the hyoid gill cleft, for some distance, the greater portion of the lateral mass breaks down into mesectoderm, between Stages III and V, and it is difficult to assign any definite boundary to the anterior end of the lateralis Vllth ganglion at first. Its anterior end, as in the case of the Gasserian, is situated more ventrally than the posterior end and comes into contact with the mesoderm of the mandibular arch anterior to the middle region of the arch. The anterior portion of this lateral mass later forms part of the geniculate ganglion of the Vllth nerve, but at this time the anterior limit cannot be outlined, not, in fact, until the placodal portion of the Vllth is detached from the epidermis, which is some hours later (A. nebulosus, Stage VIII). The whole ganglionic mass has an upward and backward trend, finally coming into contact with the auditory vesicle on its anterior, mesial and ventral walls. The whole ganglionic mass of the lateralis Vllth is quite homogeneous in structure, and shows no evidence of separating into the two ganglionic masses, the dorso-lateral and ventro-mesial, of which it is later composed. Its anterior boundary when it can be determined is not overlapped by the Gasserian ganglion but is overlapped by its root. Its posterior end, just where it comes into contact with the vesicle, is more closely applied to the neural tube than to the epidermis. The fibrillated root appears later at the posterior end, where it is in contact with the middle region of the cord.

Fig. 11, from the same embryo as that from which fig. 10 was taken, lies four sections anterior to the auditory vesicle. The lateralis Vllth ganglion shows the same irregular boundaries that characterized the Gasserian ganglion and is surrounded on all sides, except that next to the epidermis, by loose mesectoderm


THE CRANIAL GANGLIA IN AMEIURUS 333

in which cell walls are indistinguishable. The cell boundaries are still recognizable in the ganglion, however, showing its continuity with the lateral cell mass (see fig. 9) . During all the earlier stages of this ganglionic mass it is impossible to locate definitely either its anterior or posterior limits. Its posterior end is in contact with the anterior end of the auditory vesicle at first and for some time also with the auditory ganglion, while its anterior end as mentioned above is in contact, if not continuous, with the posterior end of the geniculate ganglion. It is not until about the eighty-sixth hour that the boundaries of the two portions of the lateralis Vllth ganglion, the dorso-lateral and the ventro-mesial, can be determined; these boundaries are shown in figs. 34 to 37 and in fig. 83.

While these two divisions of the ganglion in the early stages are fused and their boundaries are difficult to determine, there is no difficulty in following the history of the lateral mass up to the time that the definitive ganglia appear. The principal variation which I have observed is in the length of the dorso-lateral portion which is always better defined than the ventro-mesial and sometimes extends well forward over the geniculate; in other cases it is quite short. One of the principal difficulties in determing definite boundaries in the lateralis Vllth ganglion comes from the fact that the root of the geniculate extends throughout the whole length of the lateralis Vllth and enters the brain at almost the same point as the root of that ganglion.

The general appearance and form of the Gasserian and lateralis Vllth ganglia resemble closely the conditions as described by Miss Beckwith ('07) in Amia, although she did not describe stages sufficiently early to determine whether the so-called neural crest arises along with the auditory vesicle as the lateral portion of the neural plate or whether the auditory vesicle arises as a thickening of the lateral epidermis and the neural crest as a more dorsally situated derivative of the epidermis. The tendency of the masses out of which the Gasserian and geniculate ganglia later form to break down first into loose tissue seems to be present in Amia also.

The origin of the Gasserian ganglion, which is a pure general cutaneous ganglion, and of the lateralis Vllth, which is a pure acustico-Iateralis ganglion, from the lateral mass in adjoining


'VA4 F. L. LANDACRE

segments receives its best explanation in the interpretation given to the acustico-lateralis system as a special cutaneous system. This system is a special cutaneous system in the sense that its center in the medulla, the tuberculum acusticum, is a specialized derivative of the dorsal horn, or general cutaneous column, of the cord (Johnston '056) . While the ear and the lateral line organs are unique in structure and function and the acustico-lateralis fibres can be distinguished from general cutaneous, the two systems stand in the relation indicated above on account of the relation of their central endings. This view is materially strengthened by the fact that the ganglia of both systems in the case of the Vth and lateralis Vllth have similar modes of origin.

Owing to the intimate relation of the auditory vesicle and placodes with the lateralis Vllth ganglia, structurally and in point of time, I shall leave that portion of the lateral mass anterior to the lateralis Vllth and posterior to the Gasserian ganglion, i.e., the communis Vllth or geniculate ganglion to be taken up in connection with the epibranchial ganglion of the Vllth nerve. From the time that the lateralis Vllth assumes definite form until the visceral Vllth appears this remnant can be located but not sharply defined. It lies just over the mesoderm of the hyoid gill bar and is somewhat denser than the more dorsal! y situated mesectoderm but does not differ sufficiently from either the mesectoderm or the mesoderm to enable one to determine its exact boundries. It is less dense than the more ventrally situated mesoderm.

THE AUDITORY VESICLE AND AUDITORY GANGLION

It is not necessary to descibe the auditory vesicle in detail but I have examined it carefully to determine if there is present any portion of the lateral mass in this region in addition to that portion forming the. auditory vesicle and also to determine the exact mode of origin of the auditory and lateralis IXth ganglia. I find that the whole of the lateral mass in the auditory region is converted into the vesicle and that the greater portion of the auditory ganglion comes from the anterior end of the vesicle, but that the ganglion lies in such close proximity to the lateralis Vllth that


THE CRANIAL GANGLIA IN AMEIURUS 335

it is not possible to determine definitely whether there are lateral mass cells in the VHIth or not, since it does not become distinct from the lateralis Vllth for some hours. Fig. 6 is characteristic of the median portion of the auditory vesicle before a definite cavity appears, at which stage it resembles the pre- and postauditory placodes (figs. 14, 45). The ventral portion of the vesicle is at this stage (fig. 6) well defined, while the dorsal portion is not yet fully differentiated from the lateral mass. The character of the lateral mass in the anterior third of the vesicle at this stage is quite like that of fig. 5 which is taken just posterior to the auditory vesicle. At the extreme posterior end where the differentiation into dorso-mesial and ventro-lateral or placodal regions has not yet appeared it resembles the condition shown in fig. 3. The ventral and mesial portions of the vesicle (fig. 6) are characterized by having elongated radially arranged cells with their nuclei situated at the periphery. Ventro-laterally the elongaled cells pass gradually into the epidermis. Dorsally, however, the elongated radially arranged cells pass into a mass of irregular cuboidal cells which on its ventral border is faintly delimited from the neural tube but dorsally becomes continuous with the tube.

The next figure (12, Stage VII), taken from a slightly older embryo, shows that this mass is practically all incorporated into the vesicle, there being a few scattered cells that possibly may be converted into mesectoderm. As to the fate of the cells connecting the vesicle with the medulla there is another possible way in which they may be disposed of, that is, they may be incorporated into the medulla. There is no way of determining whether this is done, however, and the appearance indicates that they become part of the vesicle. The important fact here is that there are no lateral mass cells left in the region of the vesicle that could be homologous with the neural crest of other authors. Fig. 12 is typical for the whole length of the vesicle at this stage.

In comparing Koltzoff's ('02) and Johnston's ('056) work on Petromyzon attention was called to the fact that the only point of disagreement was in regard to the presence of a neural crest which Koltzoff finds in the auditory region, and that an examina


336 F. L. LANDACRE

tion of the literature throws little light on the question. Dohrn ('90), however, taking this matter up specifically states that the neural crest is present at first in the auditory region but that it disappears. It was hoped that Ameiurus would throw light on the question, but the fact that there is no specific neural crest in Ameiurus renders my conclusions somewhat unsatisfactory in this matter. Since, however, there is no remnant of the lateral mass left in the auditory region after the vesicle forms, we may conclude that whatever the relation of the neural crest of other types to the lateral mass of Ameiurus the position taken by Dohrn is strongly supported by the evidence from this type and I find myself in agreement with Johnston ('06) when he states that the neural crest is absent in the auditory region. This statement will not hold for the region immediately anterior to the auditory vesicle, as was shown in describing the differentiation of the preauditory lateral mass. The lateral mass there furnishes the two lateralis ganglia of the Vllth nerve.

The exact mode of origin of the VHIth ganglion is complicated by the fact that, while it seems to come almost entirely from the anterior end of the auditory vesicle, it develops in such close proximity to the preauditory lateral mass which gives rise to the lateralis Vllth ganglia and is so closely fused with the ventromesial ganglion of these two ganglia that it is impossible to be certain of its exact composition. In the stage in which the auditory vesicle is first recognizable by the radial arrangement of its cells there is no mass of cells occupying the position of the future auditory ganglion; this, when it appears as a definitive mass, is situated as a cap of cells adhering closely to the ventral and ventromesial portion of the anterior portion of the vesicle and extending slightly beyond the anterior end. Here it comes into contact with the irregular mass of cells which differentiates later into the lateralis ganglia of the Vllth nerve. At this stage before the appearance of the auditory ganglion and up to the time the preauditory placode separates from the vesicle the anterior walls of the vesicle are well defined and there is no evidence that cells are being proliferated from it. However, shortly after the separation of the preauditory placode from the vesicle the walls of the anterior


THE CRANIAL GANGLIA IN AMEIURUS 337

end of the vesicle lose their characteristic regular outline, while throughout the remainder of the vesicle its walls are still characterized by their clean cut boundary and by the presence of a single row of distally located nuclei. At the anterior end, however, the ventral wall becomes indistinguishable and its nuclei become several layers deep and there is no perceptible boundary between the vesicle wall and the forming VHIth ganglion.

Fig. 13, Stage V, taken from the same embryo as figs. 10 and 11, lies four sections back of the anterior end of the auditory vesicle. The mingling of the vesicle and ganglionic cells and the numerous mitotic figures in this location indicate without doubt that the vesicle contributes cells to the ganglion. If it were not for the fact that the vesicle is constantly in contact with the ganglionic mass derived from the lateral mass, the lateralis Vllth ganglion, and seems to grow forward into this mass on its dorsal and mesial wall, one would be inclined to think that the whole auditory ganglion came from the vesicle. The conditions here are identical with those at the posterior end of the vesicle where the lateralis IXth ganglion is formed. That ganglion is proliferated from the wall of the vesicle after the vesicle is formed and can be followed from its first appearance until the vesicle ceases to contribute cells to it. The conditions at the posterior end of the vesicle are not complicated to the same extent by the presence of any contiguous ganglionic mass and one can be much more certain that the whole ganglion comes from the vesicle. Since, however, at the anterior end of the vesicle the cells derived from the vesicle are in contact with those derived from the lateral mass and there is no definite division into an Vlllth ganglion and two lateralis Vllth ganglia for some time after this, it is impossible to say positively that the whole of the Vlllth is derived from the vesicle. This mode of derivation is indicated by the conditions, however, and is further strengthened by the positions of the lines of cleavage separating the Vlllth from the lateralis Vllth which appear later, as well as by the manner in which the Vlllth ganglion adheres to the vesicle for a long time (figs. 35 to 39 and fig. 83).

I shall defer a discussion of the various ways in which the acustico-lateralis system of ganglia arises until after the description


338 F. L. LANDACRE

of the origin of the lateralis Xth. The fact that there may be lateral mass cells in the VHIth, however, does not affect the homogeneity of this system of ganglia, since the acustico-lateralis system is based on anatomical and physiological characters. We have lateralis ganglia, as in the case of the Vllth, derived solely from the lateral mass, probably from the neural crest in other types, and on the other hand lateralis ganglia, as in the case of the lateralis Xth to be described later, derived solely from the postauditory placode which is the posterior extension of the auditory vesicle. The VIII may be intermediate between these two extremes, since it possibly derives cells from both sources. The diversity in mode of origin shown by acustico-lateralis ganglia emphasizes the fact that the ultimate basis for the establishment of this system of ganglia and nerves rests on anatomical characters and on the central connections of this system in the brain and not on embryological evidence, since some of its ganglia come directly from the lateral mass like the general cutaneous ganglia, while others arise secondarily, coming from the auditory vesicle or placodes, and show a more specialized mode of origin.

There is a great deal of variation in the extent to which these three ganglionic masses (VHIth and two lateralis Vllth) become separated from each other at any given stage. Sometimes the dorso-lateral Vllth is free at one or both ends, while in other embryos of the same age one or both ends may be incompletel}* separated from the adjoining ganglia. There is also much variation in the relative lengths of the two divisions of the lateralis Vllth, the dorso-lateralis Vllth sometimes extending far foward on the lateral surface of the Gasserian. Up to the 86-hour stage which I have plotted and in which the three divisions are isolated (fig. 83) the variations strike one as representing different degrees of isolation simply, and in this stage are in such a condition as Herrick ('99) describes for the acustico-facial complex in Menidia. After the 86-hour stage the ganglia seem to be in various stages of assembling into the adult condition of Ameiurus in which they are much more closely fused than in Menidia. In fact, the 86-hour stage of Ameiurus seems to be in about the same condition as the adult ganglia of Menidia.


THE CRANIAL GANGLIA IN AMEIURUS 339

THE FATE OF THE PREAUDITORY PLACODE

Under the term preauditory placode I include the anterior extensions of the auditory vesicle from the point where the vesicle narrows at its future anterior boundary to the extreme anterior limit of this extension. At the time when the anterior and posterior boundaries of the vesicle can first be determined by the radial arrangement of its cells and by the size of the vesicle, the placode is represented by a slightly differentiated region in the ventral portion of the lateral mass. Fig. 4, Stage III, is taken four sections anterior to the auditory vesicle; at this stage the preauditory placode occupies the whole of the ventral portion of the lateral mass. Its shortest diameter lies in the dorso-ventral plane and the walls of its cells are quite distinct. The mesial surface of the placode is in contact with mesectoderm cells which seem to have been proliferated from its surface in this early stage. Farther forward (fig. 9, Stage III) the placode is reduced in size and its separation from the lateral mass is less distinct. Slightly anterior to this point it can no longer be detected, having merged completely with the remainder of the lateral mass.

The later history of this placode shows that for a time it becomes more definite in appearance, simulating closely a lateral line organ and on a small scale the changes in the auditory vesicle, and then later disappears entirely, the greater portion of it being converted into mesectoderm. It does not give rise to either ganglia or lateral line organs. The lateralis ganglia anterior to the vesicle can be located before the preauditory placode disappears and the preauditory lateral line organs do not appear for some hours after the disappearance of the last remnant of the placode. The first trace of lateral line organs in A. melas appears in an embryo of 75 hours. There is no trace of the preauditory placode left in an embryo of 49 hours. In A. nebulosus the last trace of the preauditory placode disappears some hours before this (Stage V), so that there is a period intervening between the disappearance of the last trace of the preauditory placode and the appearance of the first primordium of the preauditory lateral line organs of more than 26 hours. .


340 F. L. LANDACRE

The process of disappearance is as follows: When the preauditory placode is at its maximum size it extends from the anterior end of the auditory vesicle as far forward as the point where the hyoid gill pocket comes into contact with the ectoderm. Its posterior end is at first continuous with the auditory vesicle and its anterior end gradually thins out into ordinary epidermis. Fig. 14, Stage IV, is taken just anterior to the auditory vesicle at a time when the placode is at its maximum size. The radial arrangement of the cells of the placode and the partial formation of a cavity corresponding to the cavity of the auditory vesicle are evident. This figure should be compared with fig. 6, Stage III. In fact, in some series there is a small cavity in the placode corresponding to that of the auditory vesicle. The first change of a retrogressive nature that I can detect is illustrated in fig. 15, which is taken from the opposite side of the same embryo. Here there is an area corresponding to the thickness of one section intervening between the posterior end of the placode and the anterior end of the auditory vesicle. Here the placode seems to have broken down into mesectoderm, since we have only mesectoderm with a trace of the ventral portion of the placode left, where on the opposite side the placode and vesicle are continuous. The irregular outlines of the posterior end of the placode and of the anterior end of the vesicle also indicate that the placode has been converted into mesectoderm and that the two structures have not simply been carried apart by the growth of the embryo. It is not likely that the placode has moved forward bodily, since the anterior end remains constant in position.

In the next stage sketched from an embryo of the same age but slightly more developed, there is an area of six sections intervening between the posterior end of the placode and the anterior end of the vesicle where the placode has been converted into mesectoderm. The placode here (fig. 16, Stage IV) differs somewhat in appearance, being longer in its dorso-ventral axis and resembling less the auditory vesicle. While the nuclei are still situated in the inner extremities of the cells, the placode has the appearance of columnar epithelium. In a somewhat older series (fig. 17,


THE CRANIAL GANGLIA IN AMEIURUS 341

Stage V) the placode at this point has broken down almost completely into mesectoderm. The former position of the placode (Ep) is indicated by the slight elongation of the remaining cells and the presence of the small cavity which frequently exists on the outer surface between the elongated cells and the flattened layer of the epidermis. This section is taken nineteen sections anterior to the auditory vesicle and there is an area of 18 sections between the anterior end of the vesicle and the posterior end of the placode in which the placode has been converted into mesectoderm. The fate of the anterior remnant of the preauditory placode is somewhat peculiar. It never extends beyond the hyoid gill pocket in any stage. The region just posterior to the gill pocket is where the proliferation of cells takes place to form the epibranchial ganglion of the Vllth nerve. This epibranchial placode, as will be shown later, begins as a thickening of the epidermis differing somewhat in appearance from the disappearing preauditory placode; but it is difficult to determine the exact relation of the last trace of the preauditory placode to the first trace of the epibranchial placode arising in the same area. A valid reason for considering the disappearing preauditory placode as distinct from the early stages of the epibranchial placode of the Vllth nerve lies in the difference in histological character of the two structures. The last stage of the preauditory placode in which it can still be recognized as such has elongated cells with nuclei placed on the inner border and shows no mitotic figures, while the early stages of the epibranchial placode has its cells irregularly arranged and mitotic figures in all stages are very frequent. Fig. 18 from the same embryo as that from which fig. 17 is taken lies six sections anterior to fig. 17 and two sections from the posterior end of the point of contact of the hyoid pocket with the epidermis, the contact of the pocket with the epidermis extending over eleven sections. The resemblance to the placode is still noticeable, the cells being elongated and, except at the central portion being only one cell deep. There are no mitotic figures present at this stage. This figure represents the last recognizeable trace of the preauditory placode.


342 F. L. LANDACRE

THE ORIGIN OF THE GENICULATE GANGLION

Owing to the fact that both in point of time and in position there is such a close relation between the epibranchial ganglion of the Vllth nerve and the disappearing preauditory placode and that the cells derived from the epibranchial placode combine with cells from the lateral mass to form the definitive communis or geniculate ganglion of the Vllth nerve, it will be more convenient to describe the origin of the geniculate ganglion here than to follow the natural order and take up the differentiation of the postauditory lateral mass. The series of figures (19 to 24, Stage VI) is taken from an embryo slightly older than the preceding, and in it the last trace of the preauditory placode has disappeared and the thickening which later gives rise to the epibranchial Vllth is just forming. Fig. 19 is taken at the point of contact of the hyoid endodermal pocket, with the epidermis and corresponds in position to fig. 18. It differs in tw T o important respects; the epidermis dorsal and ventral to the contact with the endoderm shows no resemblance to the preauditory placode but is composed of cells whose outlines are very indistinct and whose nuclei are irregularly arranged. In addition to the disappearance of the placode in the region just mentioned, at the point of contact with the endoderm, the epidermis is closely fused with it and presents the appearance of being about to disappear entirely. This gill slit does not open completely in any of my series but the two layers of endoderm do separate in one of them, presenting the exact appearance of an open gill pocket. Fig. 20 is taken four sections posterior to fig. 19 at the point where the contact of the endoderm with the epidermis ceases and it is characterized by the irregular arrangement of its epidermal nuclei, some of which appear to have been proliferated mesially but are still in contact by cytoplasmic strands with the epidermis. Figs. 20, 21, 22, 23 are consecutive sections and lie just behind the point of contact of the gill pocket with the epidermis. In figs. 21 and 22 a proliferated mass of cells, quite small, is isolated slightly from the epidermis. The earliest stage of the epibranchial placode of the Vllth nerve is indicated by the thickened epidermis shown in these figures.


THE CRANIAL GANGLIA IN AMEIURUS 343

Tracing the placode back in a number of series leaves no doubt that this is the primordium of the placode. The manner in which this thickening of the epidermis increases in size and finally becomes detached, forming a portion of the geniculate ganglion, is quite easy to follow. Of the fate of the small mass of cells (x) shown in figs. 21 and 22, I am not certain. They may enter into the ganglion, but it is more probable that they become converted into mesectoderm. They lie a little anterior to the point of origin of the ganglion, and there is in some of my later series a small mass of cells anterior to the point of contact of the gill pocket with the epidermis, which resembles these, but their origin and fate I cannot determine definitely. . If the cells shown enter into the ganglion, it is by attaching themselves later to the main ganglionic mass derived from the placode. Of this, however, I have no evidence.

Figs. 22 and 23 show a typical appearance of the early stage of any epibranchial placode and resemble closely the first stages of the placodes of the IXth and Xth epibranchial ganglia. The nuclei are several layers deep and quite irregular in arrangement and mitotic figures are beginning to be quite numerous. Fig. 24 is taken four sections posterior to fig. 23 and shows the transition of the placode into ordinary epidermis. Here, however, mitotic figures are still rather numerous but two sections posterior to this the epidermis is unmodified.

A comparison of the embryo from which figs. 19 to 24 (A. nebulosus, Stage VI) were taken with the one from which figs. 17 and and 18 (Stage V) were taken leaves little doubt that, while the epibranchial placode appears practically in the place where the preauditory placode disappeared, there is no direct relation between the two histologically. All of the preauditory placode posterior to the hyoid gill pocket is converted into mesectoderm and, while there is no evidence that the placode at the point of contact is ever converted into mesectoderm beyond the presence of the small mass of cells mentioned above, the fact that a contact with endoderm is formed would obscure the conditions here somewhat. While it is impossible to say that none of the cells that were once a part of the preauditory placode or their direct descendants enter


344 F. L. LANDACRE

into the primordium of the epibranchial placode, the histological differences of the two structures and the fact that the one differentiates after the other has lost its characteristic structure show that the relation of the two placodes is more apparent than real. Added to this we have the evidence to be shown later that there is absolutely no relation between the postauditory placode and the epibranchial placodes of the IXth and Xth nerves.

The condition of the epibranchial placode in a somewhat older embryo in which the placode is still in contact with the epidermis is shown in figs. 25 to 30 (A. nebulosus, Stage VII). Fig. 25 is taken just posterior to the point of most intimate contact of the hyoid gill pocket with the epidermis. In this series the contact of the hyoid pocket with the epidermis, similar to that shown in fig. 19, (Stage VI) occupies only one section, while in the series from which fig. 19 was taken the contact is four sections in length. The hyoid pocket after coming into intimate relation with the epidermis gradually withdraws, and in the next stage following that from which fig. 25 was taken no longer reaches the epidermis at all.

Fig. 25 is taken at the extreme anterior end of the placode and the placode here differs from the epidermis anterior to it only in being somewhat thicker and in having its nuclei irregularly arranged in more than one row.

Fig. 26 is taken from the next section posterior to that from which fig. 25 was taken and there is here a decided thickening of the epidermis with numerous mitotic figures in various stages. In the next section (fig. 27) the thickening projects mesially as a well defined mass of cells and there are no less than twelve cells in various stages of mitosis in the placodal region of the epidermis. In the succeeding section (fig. 28) the proliferated mass of cells, the anterior rJortion of the future epibranchial ganglion, is larger and contains mitotic figures in addition to those in the epidermis. The ganglion is still attached to the epidermis in all these areas by a pedicle fully as thick as the ganglion itself. Two sections posterior to this point (fig. 29) the ganglionic proliferation changes somewhat in appearance, being longer and its attachment involving more of the epidermis ventral to it, so that


THE CRANIAL GANGLIA IN AMEIURUS 345

it passes gradually into the epidermis here, while on its dorsal surface it passes suddenly into rather thin epidermis. The appearance in this section indicates that the cells contributed to the ganglion before its detachment come mainly from the epidermis situated ventrally to its point of attachment. This conclusion is verified from a study of the epibranchial placodes of the IXth and Xth ganglia. The whole of the ganglion shown in fig. 29 is not derived from the placode. The portion situated mesially and dorsally (L. M . G. VII, fig. 29) is derived from the remains of the lateral mass lying anterior to the lateralis Vllth ganglion. The separation between these two constituents of the ganglion is faintly indicated in the figure. The presence of these lateral mass cells can be first detected in the section preceding the one from which fig. 29 is drawn and the separation is there somewhat more distinct than in the section sketched. In the section following the one from which fig. 29 was taken (fig. 30) the division between the two constituents is quite apparent and the portion of the ganglion derived from the lateral mass is slightly larger than the placodal constituent. A few sections posterior to this point the placodal portion of the ganglion ceases to be present, while the lateral mass portion reaches its maximum size in this embryo.

The posterior end of the lateral mass portion of the ganglion cannot be definitely determined, since at this stage it simply becomes looser in texture and finally before reaching the region of the lateralis Vllth can no longer be distinguished from that ganglion and from the ventrally situated mesoderm. The fact that this ganglionic mass is at its posterior end rather closely applied to the ventrally situated mesoderm makes it impossible to determine the exact boundary posteriorly before the ganglion assumes definite shape, as it does a little later after the detachment of the placodal portion; in addition to this fact the fibrillated root of the whole ganglionic mass appears at the posterior end of the ganglion and the point where this appears is in most of the cranial ganglia preceded by a more or less ill-defined mass of cells which renders the determination of exact boundaries difficult.

THE JOURNAL OP COMPARATIVE NEUROLOGY AND PSYCHOLOGY, VOL. 20, NO. 4.


346 F. L. LANDACRE

At this time the future ganglion consists of two constituents, first a placodal constituent proliferated from the thickened epidermis or epibranchial placode which is largest in its middle region and becomes smaller both posteriorly and anteriorly and is everywhere still attached to the epidermis although it later becomes completely detached. Secondly there is a lateral mass constituent lying on the dorso-mesial portion of the placodal constituent and ending posteriorly in loose lateral mass cells so that at this stage its posterior boundary cannot be definitely determined.

As the geniculate ganglion becomes larger the difference in size between the body of the ganglion and its root enables one to determine approximately the posterior extremity of the ganglion. This can be done with certainty only when the fibrillated root appears. At this time the trigemino-facial complex has the arrangement indicated in fig. 83 when the Gasserian and geniculate ganglia overlap for about one-half the length of the Gasserian. This overlapping is due apparently mainly to the forward growth of the geniculate on the mesial side of the Gasserian, the root of the Gasserian remaining constant in position and the body of the ganglion extending only slightly posterior to its original position. The forward extension of the geniculate is probably entirely due to growth in size of the ganglion, since I can find no evidence of the further proliferation of cells anterior to the point occupied by the placode.

Both the Gasserian and geniculate ganglia assume a definite form with well defined boundaries rather slowly, so that for some time they consist of condensed masses of cells with irregular outline. Since the extent of fusion of the acustico-facial complex is very great in the adult in Ameiurus, I have given a series of figures (31 to 39) to show the condition of the ganglia at a time when all the components are still separate and the roots and chief trunks of the nerves derived from these ganglia are still distinct. These figures are taken from the same embryo (A. melas 86 hours) as that from which fig. 83 was constructed. This plot should be compared with fig. 2 of Herrick's paper ('01). The principal point of interest here lies in the fact that the roots and chief


THE CRANIAL GANGLIA IN AMEIURUS 347

trunks derived from these ganglia are found to contain only general cutaneous or communis or lateralis fibres.

Fig. 31 is taken through the origin of the trunk of the Gasserian ganglion, and fig. 32 just anterior to the point where the trunk of the geniculate ganglion leaves the ganglion, and consequently lies between the trunks of the Gasserian and geniculate ganglia. These two trunks are entirely separate not only at their point of origin but throughout their whole course at this time. They form the supero-lateral and infero-mesial strands respectively of Wright ('84 b) . These combine and later separate to give rise to the maxillary and mandibular trunks which are mixed, containing general cutaneous, visceral, and lateralis components. Herrick ('01, p. 183) could not determine positively that these two strands were pure but there can be no doubt that they are pure at this stage and that the supero-lateral strand contains fibres from the Gasserian ganglion only and that the infero-mesial strand contains fibres from the geniculate ganglion only. Further, the roots are not yet in contact at any point nor do they contain lateralis fibres as yet.

Figs. 33 and 34 are taken through the root of the Gasserian ganglion and show that the root is no less distinct than the trunk. The dorso-lateral lateralis Vllth also appears in these sections. The roots of the geniculate and Gasserian ganglia are separated by 15 sections, while the trunks at this stage are six sections apart. The overlapping of the ganglia finally brings the trunks into the same plane but even then there can be no doubt that they are derived separately from their respective ganglia.

Fig. 35 is taken through the origin of the hyomandibular nerve (T. V. L. VII) which is at this time a pure lateralis nerve derived from the ventral lateralis Vllth ganglion (V. L. VII). This nerve later contains in the adult small strands of general cutaneous and visceral fibres (Herrick, '01) but is at first a pure lateralis nerve. The ramus oph. sup. VII is in the adult a pure lateralis nerve. In this stage it can be detected as a forward extension of the anterior end of the dorsal lateralis Vllth ganglion but it is only faintly fibrillated. The ramus oph. sup. V, I cannot detect at this stage but it is probably mixed at an early stage, since the


348 F. L. LANDACRE

Gasserian and geniculate ganglia from which it derives its two components lie side by side and their anterior ends from which the nerve arises are intimately in contact when it can first be detected. The arrangement of the roots of the geniculate, lateralis Vllth and auditory ganglia are shown in figs. 36 to 39. They are all quite distinct and can be followed to the point of contact with the medulla.

As mentioned above, from this time on the components of the Vth and Vllth ganglionic complex become more intimately fused so that it is not surprising that nerve trunks which are at first undoubtedly pure trunks, i.e., contain only one component, later become mixed and contain two or more components. The contiguity of the ganglionic masses furnishes a basis for the mixing. Those components which are not found in the nerve at first must grow into it later. They are unquestionably there in the adult and not there in the embryo. The geniculate ganglion wedged in between the Gasserian and lateralis Vllth ganglia is in a particularly favorable position to send its fibres into nerves derived primarily from other ganglia. This has gone so far in Ameiurus that 12 out of 14 of the chief nerves coming from the acusticofacialis complex contain visceral fibers, as compared with five in Gadus and Menidia.

There are, at least, three methods by which nerves become mixed. The first is illustrated by the maxillary and mandibular trunks which arise some distance from the ganglion, after the fusion of two pure strands, the supero-lateral which is pure general cutaneous and the infero-mesial which is pure visceral. These pure strands after coming into contact break up into two mixed nerves, the maxillary and the mandibular. A second method by which nerve trunks become mixed is illustrated by the hyomandibular which is at first a pure lateralis nerve but later becomes mixed by having general cutaneous and visceral fibres grow into it from their respective ganglia. A third method is illustrated by the rami oph. sup. V and VII which, owing to the contiguity of its ganglia, the Gasserian and geniculate, seems to be mixed from the first, the roots growing out from the ganglia in contact with each other.


THE CRANIAL GANGLIA IN AMEIURUS 349

DIFFERENTIATION OF THE POSTAUDITORY LATERALIS MASS

The changes which take place in the lateral mass posterior to the ear resemble in a general way those which take place anterior to the ear. The postauditory lateral mass gives rise to a post auditory placode situated ventro-laterally, continuous with the auditory vesicle, while the dorso-mesial region breaks down into a loose mass of cells which is converted partly into mesectoderm and partly into the general cutaneous and general visceral ganglia of the Xth nerve. As mentioned above, the lateral mass is at first entirely homogeneous. Fig. 3 represents the appearance throughout its whole extent at an early stage, except that the lateral thickening is slightly less marked as it passes into the region of the spinal cord.

The first evidence of differentiation posterior to the ear is shown in fig. 5 (A. nebulosus, Stage III) which is taken four sections posterior to the posterior end of the auditory vesicle. There is noticeable here a slight distinction between the ventro-laterally situated postauditory placode and the dorso-mesial region. The separation between these regions is indicated by the mode of contact of the cell walls on either side of the dividing line and by the appearance of a vacuole. This sketch is taken from the same embryo as that from which figs. 4, 6, 7, 8 and 9 were taken. The differentiation of the preauditory lateral mass is only slightly in advance of that of the postauditory mass. The changes in the dorso-mesial portion of the postauditory mass by which it becomes converted into mesectoderm resemble closely those that take place in the region of the Gasserian ganglion. On either side of the Xth ganglion the dorso-mesial portion of the lateral mass becomes converted into a very loose mass of mesectoderm, except at its ventral border; heie where it comes into contact with the mesoderm it remains slightly denser.

In the region where the Xth ganglion forms and possibly where the IXth forms the process of conversion into mesectoderm does not go so far and the lateral mass in these regions is in some embryos slightly denser than in the surrounding regions. These two regions however are onty slightly denser and, in a stage fol


350 F. L. LANDACRE

lowing the breaking down of the dorso-mesial portion of the lateral mass present the appearance of not having broken down so completely as surrounding areas. A comparison of figs. 40 to 43, (A. nebulosus, Stage VII) all from the same embryo, brings out these relations in an embryo in which the ganglionic regions are unusually well marked. Fig. 40 is taken one section posterior the auditory vesicle. The whole region between the cord and the endoderm is occupied by the derivatives of the lateral mass. Just dorsal to the endoderm (fig. 40, Y) the tissue is somewhat dense and there is also a slight condensation (fig. 40, G) at the level of the upper third of the medulla; this condensation extends into the next section (fig. 41 , G) but disappears in the section following that from which fig. 41 was taken. From this point back to the region where the Xth ganglion appears the lateral mass presents the appearance shown in fig. 42. The mesectoderm presents the appearance ofmesenchyme having well defined nuclei but with illdefined cytoplasmic branches forming a loose network. Just over the mesoderm the derivatives of the lateral mass in all three figures are somewhat denser. It will be recalled (pp. 330-331, 345) that it is this portion of the lateral mass that forms the Gasserian and geniculate ganglia in the preauditory region.

Figure 43 is taken through the region of the Xth ganglion just anterior to the anterior end of the lateralis Xth. This section is taken just back of the anterior end of the ganglionic mass (J. G. X, fig. 43) . The anterior end of the ganglion lies somewhat nearer the medulla than in the section figured, while posterior to this point the mass is found more ventral and lateral in position, finally coming into contact with the denser portion indicated in figs. 41 and 42, (F), lying just over the mesoderm. The loose mass of cells shown in fig. 43 (J. G. X) gives rise to the jugular ganglion to be described later.

Like the early stages of the Gasserian ganglion, the whole mass is extremely ill-defined with irregular borders passing gradually into the surrounding mesectoderm. The ganglionic mass does not reach the dorsal portion of the medulla and there is no indication of a migration of cells from that structure.


THE CRANIAL GANGLIA IN AMEIURUS 351

At this stage in the formation of the ganglion there is nothing to indicate its presence except the slightly denser massing of the cells and the somewhat denser character of the inter-nuclear material. There is one rather striking difference between the Xth and preauditory ganglia. Anterior to the ear all the ganglia trend down and forward from their points of attachment to the brain where the fibrillated roots will appear. The Xth ganglion derived from the lateral mass in the same manner trends down and back from the point where its fibrillated root will appear, reversing the relations to the medulla. These figures are taken from an embryo in which the condensations in the region of the IXth and Xth ganglia are usually well marked. From this stage up to the seventy-fifth hour in A. melas there is no well defined ganglion even in the region of the Xth, and in several series I can find no evidence of either ganglion. Between the 49-hour and the 56-hour embryo, A. melas, there is nothing that can be identified positively as a Xth ganglion, although one can be quite sure of the region in which it ought to be located. In embryos of 61 to 65 hours, A. melas, there are slight condensations in the region of the Xth, but they would hardly be detected if one did not read back from older series. The ventral derivative of the lateral mass (Y, figs. 41, 42) is present in the region of the Xth during this time, however, and it is this region which furnishes the greater portion of the visceral Xth. During the whole time under discussion the lateralis Xth ganglion is present and furnishes a good landmark in the attempt to locate the early stages of the visceral Xth.

By the seventy-fifth hour (A. melas), the ventral portion of the Xth ganglion derived from the lateral mass has assumed a definite form and can be followed from this time on easily. It occupies at this stage a ventro-mesial and mesial position with reference to the lateralis Xth which is an elongated cylindrical mass of cells with quite definite boundaries. The visceral Xth extends from near the anterior boundary of the lateralis Xth to a point beyond its middle portion and the motor root of the Xth extends along the mesial side of the lateralis Xth running forward and upward to the medulla. At this stage the lateral mass ganglion of the Xth is


352 F. L. LAND ACRE

situated low in the body on a level with the dorsal portion of the gill slit and is connected with the medulla by the narrow strand of loose cells described above. The outline of the ganglion is still indefinite and its ventral portion is closely applied to the mesoderm underlying it. All of the visceral portion of the Xth except that derived from the placodes seems to come from the ventral portion of the lateral mass, while the jugular or general cutaneous ganglion comes from the dorsal portion. The jugular ganglion cannot be detected for some time after the visceral Xth is formed. Owing to the intimate relation of the lateral mass portion of the Xth ganglion to the epibranchial placodes, it will be better to describe the lateralis Xth first and return to the visceral Xth later (p. 79). The fate of the lateral mass cells in the region of the IX ganglion will be taken up with the lateralis IXth. The relation of the visceral Xth ganglion to the jugular or general cutaneous Xth which appears later and is situated near the medulla intracranial^, is extremely interesting. It will be recalled that the lateral mass anterior to the ear gives rise to the Gasserian, a pure cutaneous ganglion, and to a part of the geniculate, a visceral ganglion which fuses with the ganglion from the epibranchial placode. Both these ganglia become closely fused in the adult and are situated intra-cranially. In the Xth, however, we have a complete separation of the general cutaneous and visceral portions; the former being small and situated intra-cranially, the latter large and situated extra-cranially and fusing with the placodal ganglion just as the geniculate does. From evidence to be brought out later it appears that the lateral mass contingent of the geniculate and visceral Xth is really a general visceral component, the special visceral or gustatory ganglia being derived from the placodes. In types having a well defined neural crest these same relations would seem to hold, i.e., the general cutaneous and general visceral ganglia of the head have a common derivation from the neural crest, which is homologous with the neural crest of the spinal cord. The chief difference between the brain and the cord in this respect lies in the greater size of the cranial ganglia and the extent to which the general cutaneous component and general visceral component, which are both represented in the spinal ganglia,


THE CRANIAL GANGLIA IN AMEIURUS 353

are separated in the head. In the more generalized Xth the general cutaneous is intra-cranial and the general visceral extracranial, while in the more highly specialized region of the Vth and Vllth both are intra-cranial.


THE ORIGIN OF THE LATERALIS Xth AND THE EARLY STAGES OF THE POSTAUDITORY PLACODE

The lateralis Xth is derived not from the lateral mass but from the postauditory placode as it moves away from the auditory vesicle. The first evidence I find of a separation between the postauditory placode and the auditory vesicle is shown in fig. 44, Stage IV. The dorsal portion of the auditory pit was present in the preceding section. In fig. 44 only the ventral portion remains as the placode. The portion which disappears is that most nearly in contact with the medulla. In the following section (fig. 45) the placode has lost all connection with the dorso-mesial portion of the lateral mass which is beginning to be converted into mesectoderm. The length of the placode at this time is only four sections. Fig. 46, Stage IV, shows the appearance of the placode in a series of the same age, but less developed, in which the auditory vesicle is continued into the placode with no break in continuity. The placode is here seven sections long.

The placode now moves back apparently after it is detached from the auditory vesicle. My series is incomplete at this period and I am consequently unable to describe the earliest appearance of the anterior end of the lateralis Xth or the rate at which the placode moves away from the auditory vesicle at first. At the time, however, when the anterior end of the placode has reached a distance of twenty-one sections from the vesicle, Stage VII, the placode has not changed in appearance but the lateralis Xth ganglion is present and has a length of ten sections and overlaps the placode for five sections. Anterior to the region of overlap the lateralis ganglion lies just under the epidermis between it and the mesectoderm as a rather irregular mass of cells (fig. 47) . Throughout the whole region of the overlap the placode is contributing


354 F. L. LANDACRE

cells to the ganglion. As one reads from the anterior towards the posterior end of the placode, the placode gradually becomes larger until we reach the middle region (fig. 50) ; it then gradually diminishes in size until it assumes the appearance of the ordinary epidermis. Coincident with the increase in size of the placode in its anterior portion there is a decrease in the number of cells in the overlapping ganglion (figs. 48, 49, 50) until just posterior to the middle (fig. 51 ) there are no ganglion cells present. This condition is duplicated in all my series in which the ganglion and placode overlap.

In an embryo of 56J hours (A. melas) the placode has moved beyond the posterior end of the ganglion and there is a space of three sections between the placode and the ganglion.

From this time on the placode moves steadily back from the region of the ganglion and as it moves away gradually loses its resemblance to the earlier condition. Its cells cease to have a radial arrangement and it is recognizable simply as a thickening of the epidermis. The appearance of the ganglion anterior to the region of the overlap is shown in fig. 52 (Stage IX) . The question as to whether the postauditory placode moves bodily away from the auditory vesicle or whether the placode at successive distances from the vesicle represents a localized differentiation at those points has been variously answered.' The anterior end of the placode seems to be partly converted into ganglion cells and partly to revert to ordinary epidermis as far as its appearance is concerned. The posterior end of the placode seems to arise by the conversion of epidermal cells into placodal cells. There is no recognizable train of cells left in the epidermis in the route of the moving placode. The lateral line organs which appear long after the p lacode has passed a given point appear approximately along the route traveled by the placode, but are not derived from the placode.

The reason for thinking that the placode moves by successive differentiations and does not migrate bodily rests on the fact that in the region of the placode, part of the placode becomes detached as the lateralis Xth ganglion and the remainder not so used is detached or passes gradually into ordinary epidermis, since there is no abrupt transition from placode to epidermis. After


THE CRANIAL GANGLIA IN AMEIURUS 355

the placode ceases to contribute cells to the ganglion the ganglion ends bluntly posteriorly and there is no strand of cells connecting the ganglion and placode. The trunk of the lateral line nerve of the Xth grows out as a new structure from the ganglion.

THE FATE OF THE POSTAUDITORY PLACODE AND THE APPEARANCE OF THE LATERAL LINE ORGANS OF THE BODY

After the postauditory placode is no longer in contact with the lateralis Xth ganglion it moves back some distance from the ganglion and remains approximately stationary from the fifty-sixth hour up to the one hundred and thirteenth hour (A. melas), during which time it gradually becomes smaller and less distinct. During the greater portion of this time it can be identified positively both by its appearance and by its position in the body and by the fact that there is nothing else with which to confuse it . In the mean time there have appeared in the region traversed by this placode in its backward movement and in the region from which the auditory vesicle was detached from the ectoderm four lateral line organs. These are the first four described by Herrick ('01, plate xiv, fig. 1). The first according to his nomenclature, which is innervated by the ramus oticus from the Vllth nerve, arises lateral to the anterior end of the auditory vesicle and appears first in my series in an embryo of 105 hours. The auditory vesicle at this time is beginning to form the semicircular canals. The second organ, innervated by the ramus supratemporalis IXth, develops between the posterior end of the auditory vesicle and the anterior end of the lateralis Xth ganglion. It appears first in the embryo of 105 hours. The third and fourth organs innervated by twigs from the lateralis Xth ganglion develop in close proximity to the posterior end of that ganglion from a common primordium which elongates and gives rise to two organs. The anterior portion of the common primordium gives rise to the third organ and the posterior to the fourth organ. This common primordium of the third and fourth organs appears first in an embryo of 86 hours and about nineteen hours before the appearance of the primordia of the first and second organs.


356 F. L. LANDACRE

At the time of the appearance of the common primordium of the third and fourth organs the postauditory placode lies about five sections posterior to it and can be distinguished easily from the primordium by the fact that the placode still retains its characteristic appearance, having elongated radially arranged cells, while the primordium bears no resemblance to a lateral line organ, being merely a thickening of the epidermis. Here, as in all other lateral line organs, one must first locate the organ after it is fully formed and then read back in series and locate the primordium. This can be done first by counting sections and second by using convenient land-marks in the body. This is necessary because none of the lateral line organs at first have any resemblance to the adult organs.

In an embryo of 99 hours (A. melas) while the postauditory placode is still recognizable as a thickening of the epidermis there have appeared at least two lateral line organs posterior to it. These two organs can be traced continuously in my older sections and lie posterior to the large branch of the Xth nerve, while the placode lies anterior to it. Owing to the fact that structures frequently do not develop uniformily even when a series of graded ages is used, it is not possible to make definite statements in regard to the last stage of the postauditory placode. Of its moving back from the vesicle and of its gradual reduction I am quite sure. In an embryo of ninety-nine hours (A. melas) while the third and fourth lateral line organs are still contained in the common primordium the placode is recognizable and back of this are two primordia of lateral line organs. In an embryo of one hundred and five hours the placode is present but no organs lie posterior to it. In an embryo of one hundred and thirteen hours the two lateral line organs previously mentioned as lying posterior to the placode are present in practically the same position as in the embryo of ninety-nine hours, having moved slightly away from the vesicle while the sections occupied by the placode in the preceding series are vacant and remain vacant in later series. These two organs persist in my later series and I infer that the postauditory placode disappears posterior to the position of the


THE CRANIAL GANGLIA IN AMEIURUS 357

fourth lateral line organ and gives rise to no organs. If any do arise from it it would be the fifth of the body lateral line.

As to the mode of origin of the organs, there seems to be little variation. Each one appears as a slight thickening of the epidermis scarcely perceptible at first except by its location at the seat of the future organ. This thickening becomes more pronounced by the elongation of the deeper cells and by their assuming a radial arrangement with frequently a small circular cavity just beneath the outer flattened layer of epidermis and at the apex of the radially arranged cells. This process of the elongation of the radially arranged cells continues until the organ assumes its permanent form. The variation mentioned above consists in the different rates at which this process is carried on and also the different stages of development at which the organs sink into canals. The facts brought out above differ materially from the usual description of the relation of the dorso-lateral placode to the lateral line organs and lateral line nerve but are in close agreement with the work of Miss Piatt on Necturus ('95). The fact that anterior to the ear the preauditory dorso-lateral placode disappears more than 26 hours before lateral line organs can be detected and that posterior to the ear at least four lateral line organs appear anterior to the placode, while it can still be recognized as such, leaves no doubt in the writer's mind that there is no evidence of a genetic relation between the dorso-lateral placodes and lateral line organs in Ameiurus. The lateral line organs are definite differentiations of the epidermis, just as are the taste buds, and the nerves which supply them grow from specific ganglia just as in the case of the gustatory nerves. There seems to be no doubt that the auditory vesicle and pre- and postauditory placodes are homologous, or rather the same structure, but the relationship of the lateral line organs and vesicle is on a somewhat different footing.

If the auditory vesicle phylogenetically is to be looked upon as containing sensory areas homologous to lateral line organs, in the ontogeny we have the troublesome fact that lateral line organs arise in the epidermis in the same area as that from which the vesicle arose; further than this the preauditory lateral line organs


358 F. L. LAND ACRE

are innervated from ganglia derived from the lateral mass or neural crest of other types while those posterior to the ear are innervated partly by ganglia derived from the auditory vesicle and partly by ganglia derived from the posterior extension of the vesicle or the postauditory placode.

The foregoing description of the relations existing between the pre- and postauditory placodes and the lateral line organs, differs so much from that of Wilson ('91, '97) in the sea bass and salmon that it merits a fuller discussion. The accounts differ as to the mode of origin of the lateral line organs, Wilson tracing them to the sensory lines derived from the postauditory placode and the preauditory placode (branchial sense organ of Wilson and Beard) while the present account traces them to differentiations of the epidermis not derived from the placodes. Wilson's statement of the case as given in his short paper on the salmon ('97) is as follows and will stand for the conception of those authors who agree with him. He states that in Serranus

The organs of the lateral line, the auditory sac, and the superficial sense organs of the head (presumably all) were derived from a common foundation. This common foundation has the shape of a long furrow (ectoderrnic) on the side of the head region. The furrow splits into three parts, the posterior part giving rise by division to the organs of the lateral line, the middle part becoming the auditor}- sac, the anterior part becoming a histologically developed branchial sense organ, situated in front of the single gill slit of the embryo, from which a (sensory) cord of cells is prolonged forwards.

Wilson finds practically the same condition in the salmon except that the pre- and postauditory furrows are only thickenings in this form, and cites Mitrophanow f 93) and Locy ('95) as agreeing substantially with him.

The paper on the salmon is brief and does not describe the exact mode of appearance of the definitive lateral line organs. An examination of Wilson's paper on the sea bass ('91) shows that the relation of the preauditory lateral line organs to the preauditory placode is much less definite than in the paper on the salmon. He


THE CRANIAL GANGLIA IN AMEIURUS 359

says (p. 246) in discussing the fate of the preauditory placode (branchial sense organ) that "the further development of the organ consists in the loss of its cavity, in histological differentiation, and in the transformation of its ill-defined anterior extremity into two cellular cords which doubtless serve as source for the production of new organs." On the following page in discussing the anterior sensory tract he states thac "during larval life one or two sense organs are found in this region and it is extremely probable that they arise from the dorsolateral tract. " In the following paragraph is the statement that "the anterior sensory tract is at the time of hatching very short and just what becomes of it I do not know." This anterior sensory tract is the tract which Wilson derives from the anterior end of the branchial sense organ and which gives rise by its forward extension to two tracts.

It may be well to call attention to the fact that where Wilson makes unequivocal statements in regard to the relation of the auditorj r vesicle to the preauditory placode, and to the post auditory placode in part, we are in essential agreement, i.e., in the origin of the auditory vesicle and the pre- and postauditory placodes from a common primordium, the detachment of these placodes from the vesicle and in the disappearance of the preauditory placode. Wilson finds that the preauditory placode disappears, while the postauditory continues to grow back giving rise to the lateral line, but I find that both the pre- and postauditory placodes finally disappear. With this exception up to this point the two accounts are quite similar. The presence of sensory cells in Wilson's branchial sense organ and their absence in Ameiurus is not a radical difference, especially since the preauditory placode in Ameiurus has such a close resemblance to the auditory vesicle.

As to the point of difference, a careful reading of Wilson's paper fails to show a definite relation between the branchial sense organ and the anterior sensory ridges, and the specific lateral line organs. The disappearance of the branchial sense organ and the anterior sensory tract, and the appearance of specific


360 F. L. LANDACRE

lateral line organs in the same region is not positive evidence of genetic relationship between the two structures. The divergence in our accounts begins in the description of the method by which the specific lateral organs arise, more particularly as to whether the sensory ridge on which Wilson finds the lateral line organs differentiating is an extension from the pre- and postauditory placodes. In Ameiurus I find no common sensory ridge from which the organs differentiate, and can trace the gradual disappearance of the pre- and postauditory placodes. Miss Clapp ('99, pp. 239 251) states that these sensory ridges originate in the auditory region but nowhere I think traces them to the auditory vesicle or its derivatives, while Miss Beckwith ('07) states definitely in tracing the genesis of the lines that they do not come from the auditory vesicle but arise as local differentiations of the epidermis on which the lateral line organs later appear (p. 28).

There is a bare possibility that the relation of the anterior sensory ridge to the branchial sense organ as described by Wilson in Serranus is one of contiguity only and not of genetic relationship. If this be admitted, and it is the point about which Wilson is least definite, then there are no insuperable difficulties in harmonizing the two views, for the lines seen by Allis ('89) and other workers may be interpreted in two different ways. They might be continuous ridges which break up into individual sense organs, or on the other hand they might be formed by a series of individual sense organ primordia whose extremities become confluent. There is evidence for both views. Wilson is very postitive concerning the mode of formation of the postauditory lateral line organs from a common primordium derived from the postauditory extension of the vesicle in the sea bass, while I have evidence which I do not think can be doubted that even the postauditory lateral line organs arise separately and have nothing at all to do with the postauditory placode. This conclusion was reached by Hoffmann ('94) and by Miss Piatt ('95, '96) in part of the organs in Necturus. The same conclusion is reached by Miss Beckwith in Amia. Allis's work ('89) does not really confirm Wilson, since his earliest stage was about one day old (after hatching) and this is too late to


THE CRANIAL GANGLIA IN AMEIURUS 361

trace the genesis of the sensory lines. Miss Beckwith shows that the lines from which the lateral line organs appear are entirely distinct from the auditory vesicle and that the supra-orbital, sub-orbital and mandibular sensory ridges arise separately. Locy states that he agrees with Wilson but gives no description of the origin of the specific lateral line organs. It will serve to clear the ground for future work if it is remembered that there are two distinct problems here; first, do the pre- and postauditory lateral line primordia grow out from the auditory vesicle or its extensions the pre- and postauditory placodes, and, second, do the lateral line organs appear individually or do they differentiate from a common primordium?

Admitting that the lateral line organs do differentiate in some types from common sensory ridges and in other types as individual organs, the question arises as to which is the more primitive method. If the appearance of individual organs as in Ameiurus, each of which is homologous with sensory areas of the auditory vesicle, is the primitive method, it is conceivable that the elongation and precocious appearance of the primordia from which these organs appear (and these primordia are always longer than the organs in Ameiurus) might result in a more or less definite line such as Wilson describes in Serranus and Miss Clapp in Batrachus. The growth of the postauditory sensory line would mean nothing more than the appearance of successive organs whose primordia have become continuous. If the differentiation of organs on lines which are extensions of the auditory vesicle is the primitive mode, it is more difficult to understand how the method by which they arise in Ameiurus could be derived from it. It is hard to see how organs that primitively come from a common primordium could later in phylogeny arise singly. The difficulty is increased when we remember that the lateralis Vllth ganglion does not come from the vesicle or any of its derivatives but from the lateral mass (neural crest). Perhaps the greatest obstacle to this view on theoretical grounds arises from the confusion which is introduced into current ideas of the primitive relation of the lateral line organs to the sensory areas of the audi THE JOURNAL OF COMPARATIVE NEUROLOGY AND PSYCHOLOGY, VOL. 20, NO. 4.


362 F. L. LAND ACRE

tory vesicle. If these organs arise as differentiations of sensory ridges which grow out from the anterior and posterior extension of the auditor} vesicle, they cannot be strictly homologous with the sensory areas of the auditory vesicle, since they are derived from it, unless we consider the ear as precociously developed on this ridge. If on the other hand the lateral line organs arose primitively as single organs, the sensory areas of the auditory vesicle and lateral line organs can be considered as homologous and the conditions in Serranus, Salmo, Batrachus, and Amia can be explained as modifications of this primitive method.

THE PLACODAL GANGLION OF THE IXth NERVE

The visceral ganglion of the IXth nerve stands in a rather unique position since it has no recognizable constituent derived from the lateral mass, and is therefore a pure placodal ganglion ; otherwise it presents many points in common with the placodal ganglion of the Vllth and those of the Xth to be described later. It begins as a thickening of the epidermis on a level with the dorsal limit of the first true gill slit, and extends posteriorly from this point. The thickening is characterized by the irregular arrangement of its nuclei, and by the number of mitotic figures. The stage of thickening is followed by a process of proliferating cells mesially and posteriorly, and this group of proliferated cells is finally detached en masse and acquires a connection with the more dorsally situated lateralis ganglion. Fig. 53 from a 562-hour embryo (A. melas) is taken just posterior to the point of contact of the visceral pocket with the epidermis. The placode is recognizable at this stage as a simple thickening of the epidermis extending from the level of the dorsal border of the future gill slit down to the level of the branchial artery and longitudinally occupying three or four sections. The epidermis is several cells thick and usually, especially during later stages, shows many mitotic figures. The epidermis is nearly always slightly indented at the point where the placode appears, but whether this is to be associated with the placode or with the gill slit is not clear. The placode differs structurally from the epidermis adjoining it in the


THE CRANIAL GANGLIA IN AMEIURUS 363

characters mentioned, and in addition it takes a slightly darker stain and its inner border is usually less ragged than that of the surrounding epidermis. The epidermis anterior and dorsal to the placode rarely has a well defined border on its inner surface, but presents the appearance of proliferating cells into the mesectoderm. The placode passes almost imperceptibly at first on its anterior and posterior borders into the epidermis. On the dorsal surface, however, the transition from placode to epidermis is quite sudden. In the region of the placode, however a well defined mesial border can be seen and, while it could not be said that no cells slip individually from this placode into the mesectoderm, there is no evidence of it.

This process of thickening of the ectoderm is followed a few hours later by a movement of the cells en masse into the region of the mesectoderm in a posterior and dorsal direction. Fig. 54, taken at the middle of the placode, shows its appearance in an embryo of 69 hours. It occupies about the same relative position as in fig. 53, but the auditory vesicle which was absent in fig. 53 appears in this section (not shown in the figure, however) , owing apparently to its extension backward. The placode has a much denser appearance here than the surrounding mesectoderm and its boundaries can usually be located with no difficulty except at the posterior end which is ill defined, except in the earliest stages before the placode begins to project mesially. The placode is shorter dorso-ventrally than in the preceding stage, and on the side from which the sketch was made not more than four sections long, but on the opposite side as much as seven sections long. On its ventral, anterior, and posterior borders it passes over into ordinary epithelium, as in the preceding stage described. Its mesial projection at the thickest part from which fig. 54 is taken is about three times the thickness of the adjoining epidermis. Cell multiplication is taking place as evidenced by the presence of mitotic figures; in fact, there is evidence of both cell migration from the epidermis after the placode begins to project mesially, and of the development of cells in situ in the projecting portion of the placode.


364


F. L. LANDACRE


In the next series, 75 hours, which I have not figured, little change has taken place, although the placode extends somewhat farther mesially; but in an embryo of 81 hours (fig. 55) a change can be noticed. Figs. 55, 56, 57 are consecutive figures taken from the same embryo. Fig. 55 is taken from the anterior portion of the placode and shows a decided constriction between the ganglionic mass and the epidermis with which it is still attached. The epidermis immediately dorsal to the attachment of the placodal ganglion has resumed its normal appearance, having a rough



edu/la

ud i/esick ud qanal/on.


e pi 'branch la t p/a code branch- art 'endoderm


Fig. E. A camera tracing corresponding to fig. 56, showing the position of the ppibranchial placode. NO. notochord; b.v. blood vessel.


inner border composed of cells which seem to be in process of becoming mesectoderm. Ventral to the placode the epidermis is still thick and the appearance here, as elsewhere in the formation of epibranchial placodes, indicates that the epidermis ventral to the ganglion is the chief source of whatever cells migrate into it.

The following section (fig. 56) presents approximately the same appearance except that the nuclei of the epidermis are separated from those of the ganglion which is attached to the epidermis by a cytoplasmic mass only. In fig. 57, however, we find the ganglionic mass detached from the ectoderm and lying free in the mesectoderm, while the epidermis resembles that of adjoining


THE CRANIAL GANGLIA IN AMEIURUS


365


regions. This portion of the ganglion was never attached to the epidermis at this point, but has grown backward and upward from the points indicated in the two preceding figures. The ganglion has now become longer than its point of attachment to the epidermis. Fig. 60, 93 hours, illustrates the first stage in which the epibranchial ganglion of the IXth is first completely detached from the epidermis. This section is taken through the point where the ganglion lies nearest the epidermis, i.e., near the anteror border of the ganglion. Careful measurements of the length of the ganglion and of its point of attachment in a series of stages make it evident that the ganglion grows as indicated.


TABLE III


Showing the relative length of the epibranchial ganglion of the IXth as compared with the length of attachment (A. melas).


Age in hours

Length of attachment to placode in

sections

Length of epibranchial ganglion in

sections


69


75


81


86


93


99


105


3


2


2


2



2


1


3


7


12


12


19


25


13


113


15


In the first series (69 hours) the mesially projecting mass is no longer than the thickening, but in the following series the ganglion occupies seven sections and the attachment two. In the later stages the ganglion elongates rapidly up to 99 hours, where it is twenty-five sections in length, and after that, owing to its becoming placed in a dorso-ventral position, is not so long. In an embryo of 93 hours the ganglion is not in contact with the skin, but is found in contact in two later series, 99 and 105 hours, and after 113 hours it is permanently detached. The placode does not present the appearance of adding cells to the ganglion after the 93-hour series. The anterior end of the ganglion seems to be simply lying in contact with the epidermis which no longer resembles a placode. The epidermis has not yet resumed its normal appearance, but shortly after this stage there is nothing in the character of the epidermis to indicate the point at which the ganglion arose.


366 F. L. LANDACRE

The introduction of the term ' branchial sense organ' by Froriep and Beard seems unfortunate. Froriep ('85) applied it to the epibranchial placodes of mammals, and Beard ('85-'86) to both epibranchial and dorso-lateral placodes apparently, and the term is used by Wilson ('91) in his paper on the sea bass. In all cases the term seems an unfortunate one. The epibranchial placodes of Ameiurus have no resemblance whatever to sense organs, particularly to gustatory organs, and the dorso-lateral placodes while resembling, in Ameiurus, lateral line organs do not give rise to lateral line organs. This is true in Necturus, as shown by Miss Piatt ('95). The branchial sense organs of Wilson are apparently the dorso-lateral placodes. Whatever may have been the phylogenetic history of these placodes, it is much better to designate them in such a way as not to commit oneself to a theory as to their phylogenetic origin until it can be shown what that origin was. I have, therefore, followed von Kupffer ('94) and have used the term 'placode' exclusively. The distinction brought out (p. 56-57) in the discussion of the relation of the pre- and postauditory placodes to the specific lateral line organs finds a striking parallel in the relations of the epibranchial placodes to their specific sense organs, the taste buds.

The lateral line organs arise in Ameiurus as definitely localized differentiations of the epidermis, and are not derived from the dorso-lateral placodes (pre-andpostauditory placodes of Ameiurus) , although they receive their innervation from ganglia derived from these placodes in the case of the X, IX and the VIII nerves.

The taste buds bear the same relation to the epibranchial placodes, as shown by the writer (Landacre '07). The taste buds appear simultaneously at the extreme anterior end of the oral cavity (ectoderm) and on the endoderm of the first three gill arches and spread posteriorally from the gills into the pharynx and oesophagus and from the anterior end of the oral cavity back into the mouth and externally over the lips, barbules and outer surface of the head and finally over the whole body.

The taste buds appear in well defined groups determined largely by the distribution of the rami of the nerves carrying gustatory fibres. These groups are isolated at first and later become con


THE CRANIAL GANGLIA IN AMEIURUS 367

fluent and always appear in order from anterior to posterior. The taste buds show one other peculiarity in their order of appearance. They appear, generally, at the peripheral distribution of the nerves innervating them anterior to the ear, and in the reverse order posterior to the ear. It is evident from these facts that they are in no way, in Ameiurus, closely related in place of origin to the epibranchial placodes. The epibranchial placodes, so far as any evidence secured from their ontogeny indicates, are to be looked upon as ganglion forming structures. The epibranchial placodes bear no resemblance to either single taste buds or groups of taste buds. These buds spread in the case of the oral and superficial head buds from the point of origin toward, and not away from the point of origin of the placodes.

An additional fact of significance is the time intervening between the detachment of a given placode and the appearance of the first taste buds innervated by fibres derived from the ganglion formed by the placode. In the IXth nerve, the placode which forms t lit' special visceral ganglion of that nerve becomes detached in a 93-hour embryo (A. melas), but the first taste buds on the pharynx, and, in fact, on any part of embryo, appear in a 113hour embryo ; a length of time sufficient to suggest tht the appearance of the taste buds is associated with the appearance of the nerve trunk, rather than with the proximity of the taste buds to the placode, since the IXth nerve has a fibrillated root and trunk which extends into the first gill arch in an embryo on 113 hours.

The process of growth by which the various components of the peripheral nerves find their appropriate sense organs furnishes one of the most puzzling problems in embryology. This is especially true of the gustatory nerves and taste buds, and in a measure is true of the lateralis system of nerves and sense organs. The solution does not seem to lie in deriving both nerves and sense organs from a common primordium, as can be done in the case of the olfactory nerve and optic nerve. The distance between the point of origin of taste buds supplied by the Vllth nerve and the point of origin of the ganglion of the Vllth nerve precludes such an explanation. The suggestion offered above that there is some coordination in the growth of the peripheral fibres and the appear


368 F. L. LANDACRE

ance of the taste buds seems much more in accordance with the facts.

A good deal of interest attaches to the dorsal extension of the epibranchial ganglion. There is nothing in its composition or mode of development to indicate that it is formed otherwise than by the extension dorsally of the epibranchial placode. If, however, there are lateral mass cells in this ganglion they must enter into its composition at this point, and for this I find no evidence.

Herrick ('01) states that he finds no fibres derived from this ganglion except gustatory or special visceral, and since there is no evidence that cells other than those derived from the epibranchial placode enter into its composition we are driven to the conclusion that the epibranchial placode furnishes those cells which give rise to gustatory fibres and that those cells which give rise to general visceral fibres in the IXth ganglion of other species must come from the lateral mass or from the neural crest.

The evidence for the presence of general visceral fibres in the IXth nerve is exactly like the evidence for the presence of lateral mass cells in the ganglion. Neither has been detected. There is a possibility of both having been overlooked, of course, but if they should be found to be present it would not in my estimation affect the conclusion drawn. It would simply fall short of a demonstration. The fact that the Vllth, IXth and various portions of the Xth ganglia possess gustatory fibres approximately in proportion to the extent to which the epibranchial placodes contribute cells to these ganglia materially strengthens the conclusion. The geniculate ganglion of the Vllth has a well defined lateral mass contingent, as was shown in describing this ganglion. That portion of the visceral Xth from which arise the large visceral rami is almost exclusively derived from the lateral mass, containing only small contingents from the reduced fifth and sixth epibranchial placodes. The number of gustatory fibres derived from this ganglion is correspondingly small.

The third and fourth epibranchial ganglia possibly contain small constituents from the lateral mass. These ganglia, however,


THE CRANIAL GANGLIA IN AMEIURUS 369

are in such close proximity to the large lateral mass ganglion of the Xth that general visceral fibres might easily grow into the nerves, so that we cannot infer the presence of lateral mass cells simply because general visceral fibres may be in the peripheral nerves of these ganglia. The IXth is unique in that neither general visceral fibres nor lateral mass cells have been found in its composition.

Aside from the explanation given above, there seems to be no reason for the presence of the epibranchial placodes in the head, and for the part they play in contributing cells to those ganglia only which supply fibres to gustatory organs. The gustatory organs are all innervated from the cerebral ganglia, and only those cerebral ganglia give rise to gustatory fibres which contain cells derived from the epibranchiaJ placodes, and the ganglion which seems to give raise to gustatory fibres only seems also to come only from the epibranchial placode.

My series at this stage are taken four hours apart, and the epibranchial ganglion acquires a connection with the proximal portion of the lateralis IXth (figs. 58 and 59) before losing its connection with the epidermis. However, the growing point of the epibranchial root is extremely thin. It passes up internal to the cardinal vein, which is in striking contrast with the growing parts of the epibranchial ganglia of the second and third true gills which pass up toward the brain external to the same vein (figs. 65, 66. 67, 68, and 69) . The growing point of the IXth between the thicker portion derived from the epibranchial placode and the lateralis IXth, is shown in fig. 58. In the preceding series (75 hours) this connection is only one cell thick and in an embryo of 69 hours it is entirely absent.

THE ORIGIN OF THE LATERALIS IXth GANGLION

After reaching the level of the auditory vesicle in its dorsal extension, the growing point of the root of this epibranchial ganglion comes into contact with a mass of cells derived from the posterior and ventral portion of the auditory vesicle, which later forms the lateralis ganglion of the IXth. It is not derived from


370 F. L. LANDACEE

nor directly related to the Xth, nor is it at first in contact with the auditory ganglion proper, but arises separately and can be followed continuously until it assumes the relations ascribed to it in the adult by Herrick ('01) . At one stage in the enlargement of the auditory vesicle its relations are somewhat confused on account of the fact that the vesicle grows back and around the ganglion, but it can be followed, as mentioned, continuously. There is a later stage also where the lateralis IXth comes into contact with the motor root of the IXth and before the motor root has grown ventrally as far as the root of the epibranchial ganglion, when it is impossible to distinguish between the lateralis cells and the cells of the growing motor root. This condition is illustrated in fig. 59. In this stage we have the lateralis ganglion and the motor root closely combined. The constituents which can be positively identified as entering into the IXth ganglion are: first, the cells derived from the epibranchial placode situated in the distal portion of the ganglionic mass in the early stage, and extracranially in the latter stages: and secondly, the proximally situated lateralis ganglion, derived from the posterior portion of the auditory vesicle and closely associated with the motor root in its early stages, and in the later stages situated intracranially.

In discussing the differentiations of the postauditory lateral mass (p. 349-50) attention was called to the appearance in one series of a slight condensation of the lateral mass derivatives, in the region of the IXth ganglion just posterior to the auditory vesicle. This condensation appears in some series and not in others, so that one cannot be positive that no cells derived from the lateral mass enter into the proximal or lateralis portion of the IXth ganglion in its later stages. The early stages of the lateralis IXth are quite definite in origin and distinct in outline, and if lateral mass cells do enter into its composition they would be homologous to the preauditory lateral mass, and the ganglion would have a double composition and would resemble the VHIth, if this ganglion contains lateral mass cells or the lateralis Vllth, rather than the lateralis Xth. The relations are difficult to unravel here, partly because of the presence of the motor root and partly because the cells


THE CRANIAL GANGLIA IN AMEIURFS 371

derived from the lateral mass are hard to distinguish from the surrounding mesectoderm. One can go no farther than to say that there is no well defined group of lateral mass cells entering into the lateralis IXth ganglion. The condition found in most of my series during the early proliferation of cells from the auditory vesicle before the definitive lateralis ganglion is formed, is illustrated in fig. 63. The lateralis IXth after it assumes a definite form is carried posteriorly by the backward growth of the auditory vesicle until it comes into the region of the slightly denser mass of mesectoderm, which at the time the lateralis IXth can first be detected is about seven sections posterior to the vesicle. This loose mass of cells, as mentioned above, is present in one of my series (A. nebulosus, Stage VII, figs. 40, 41, G), but in only one is it at all definite or does it present the appearance of an early stage of a ganglion. I take it to be the mass of cells which other authors have described as neural crest cells that enter into the IXth ganglion. ,

The early appearance of the definitive lateralis IXth ganglion is somewhat variable. It may not appear on both sides of the same embryo at the same time, and may vary in position, lying either on the ventro-mesial (fig. 61) or on the ventro-lateral side (fig. 62) of the posterior portion of the auditory vesicle. The definitive ganglionic mass may appear as a few loosely joined cells, some of which are still not entirely detached from the wall of the vesicle.

As in the case of the auditory ganglion, one can be quite sure that the bulk of the lateralis IXth ganglion is derived from the auditory vesicle. In the early stage of the vesicle there is no ganglionic mass in the position later occupied by the lateralis IXth, and the walls of the posterior end of the vesicle are clean cut and regular. Somewhat later the posterior wall becomes several cells thick, mitotic figures are numerous and a mass of cells is found attached to the ventro-mesial or ventro-lateral portion of the vesicle. No definite boundary can be determined for the forming ganglion, and cells are found in all positions between the wall of the vesicle and the body of the ganglion (fig. 63).


372 F. L. LANDACRE

There is a well defined space, however, between the anterior end of the lateralis IXth and the posterior end of the auditory ganglion. The backward growth of the vesicle carries the ganglionic mass of the lateralis IX into the region occupied in some series by the condensation of the lateral mass, so that it is impossible to tell whether these lateral mass cells enter into the lateralis IXth or not. If they do, the lateralis IXth resembles the auditory rather than the lateralis Xth, which is derived solely from the postauditory placode.

Of the four acustico-lateralis ganglia in Ameiurus, the double ganglion belonging to the Vllth is derived solely from the lateral mass and shows its affinities with the general cutaneous Gasserian ganglion most closely, not only in the source from which it comes but in its mode of origin.

The lateralis Xth, derived from the postauditory placode, is decidedly unlike the lateralis Vllth, both in general appearance and mode of origin. It is from the early stages a well defined rod-like mass of cells whose long axis coincides with the longaxis of the body, and is derived exclusively from the postauditory placode. Intermediate between these two extremes lie the auditory and lateralis IXth, which come largely if not exclusively from the auditory vesicle, but may have lateral mass cells in their composition.

The lateralis Xth is the most highly specialized of the four, and the reversal of the order of specialization, which usually proceeds from anterior to posterior, is at first sight striking. The whole acustico-lateralis system is, however, strictly a cranial specialization, the ganglia and nerves appearing first in the head and extending from there to the body, and having its center in the cranial region. The change from the generalized lateralis ganglion of the Vllth to the specialized lateralis ganglion of the Xth is in line with the order of appearance of the lateral line organs, which arise first in the cranial region, and later appear on the body. While usually a highly specialized structure lies anterior to a less specialized one, in the case of the acustico-lateralis ganglia the structure having a specialized mode of origin lies posterior to the one with a more generalized mode of origin.


THE CRANIAL GANGLIA IN AMEIURUS 373

THE ORIGIN OF THIRD AND FOURTH EPIBRANCHIAL PLACODES

The history of the first placode entering into the composition of the Xth ganglion, the third epibranchial, resembles very closely that of the epibranchial placode of the IXth ganglion, except that it forms its attachment to the brain in conjunction with the remaining roots of the Xth, by extending posteriorly until it joins the roots of the fourth placode. It can be detected first in my series (fig. 64) in a 69-hour embryo, at which stage it resembles very closely the early stage of the placode of the IXth nerve. In the next series (75 hours) the placode has proliferated quite a mass of cells mesially, but is not free from the epidermis at any point in its length, although three sections from its anterior end there is an evident constriction between the ganglionic mass and the epidermis.

In an 81-hour embryo the ganglion is not in contact with the epidermis at its posterior end, and has acquired a connection with the mass of cells lying over the site of the fourth and fifth epibranchial placodes of the Xth (fig. 79). In an 86-hour embryo the ganglion is still attached to the epidermis, but is somewhat larger and its root somewhat more evident. In a 93-hour embryo i^figs. 80 and 65) the ganglion is still attached at its anterior end. In fig. 66, 93 hours, taken from a section following 65, the ganglion is attached by a very narrow neck of cells and throughout the remainder o f its length is entirely free (figs. 67, 68, 69, 70) extending posteriorly into a narrow neck of cells which connects it with the last division of the Xth nerve. The root of this nerve, as mentioned above, lies lateral to the cardinal vein, as do all the other placodal rootsof theXth. The appearance of the epidermis near the placode resembles closely that in the region of the IXth. Anterior, dorsal and posterior to the placode the epidermis is usually one or two cells thick and quite ragged, presenting the appearance of proliferating cells into the mesectoderm. The thickening of the placode extends ventrally in the epidermis and, as in the case of the IXth, this region presents the appearance of furnishing most of the cells that move into the ganglion. Mitotic figures are occasional both in the epidermis and the placode.


374 F. L. LANDACRE

There seems to be no question that here, just as in the case of the second placode, we have first a thickening of the epidermis above and behind the second true gill slit, followed by a proliferation of cells dorso-mesially to form the body and root of the ganglion, and later the complete detachment of the proliferated mass from the epidermis. The attachment disappears last at the extreme anterior end in an embryo of 99 hours. The dorsal extension of this placodal ganglion resembles closely the dorsal extension of the placodal ganglion of the IXth nerve; but the fact that it grows back parallel to the dorsal extension of the fourth epibranchial ganglion, and that both soon come into contact with the lateral mass ganglion of the Xth whose boundaries are at this stage not well defined and which is surrounded by mesectoderm make it more difficult than in the case of the IXth to determine whether there are lateral mass cells in its composition. General visceral fibres, if present in this ganglion, might be traceable either to lateral mass cells incorporated into the ganglion, or they might easily grow into it from the large lateral mass ganglion of the tenth, situated over the fourth and fifth gills. The conditions are by no means as favorable for drawing a definite conclusion as to the composition of the ganglion as in the case of the epibranchial ganglion of the IXth.

The development of the fourth epibranchial placode which gives rise to the fourth epibranchial ganglion resembles closely that of the second and third. In an embryo of 81 hours the posterior extension of the placodal ganglion has come into contact with the lateral mass ganglion situated just over the fourth gill slit at a point just ventral to that at which the root of the second epibranchial ganglion joins the same mass (fig. 79) .

I have been unable to detect the placode in a 69-hour embryo. My 75-hour embryo is defective at this point. In the 81-hour embryo the ganglion is in contact with the epidermis throughout about half its length. The conditions are so similar here to those of the IXth, and first placode of the Xth, that I am sure that the method of appearance is the same. In fig. 68 (93 hours) is shown the appearance of the placode at the anterior portion of its attachment to the epidermis. Fig. 69 is taken two sections posterior


THE CRANIAL GANGLIA IN AMEIURUS 375

to 68 and shows the complete detachment of the posterior end of the ganglion from the epidermis. Fig. 70 shows the appearance of this ganglion seven sections posterior to the section from which fig. 69 was drawn, and just anterior to its union with the anterior portion of the lateral mass ganglion over the fourth gill slit. Fig. 71 shows the point of union of this ganglion with the anterior end of this lateral mass ganglion. The only difference between this placode and the third epibranchial lies in the thickness of its posterior portion, that part corresponding to the root of the third epibranchial ganglion. It lies much nearer its destination and apparently does not become so attenuated in reaching it. The later history of this ganglionic mass (figs. 81 and 82, Ep. G. IV) shows that it fuses much more closely with the ganglionic mass lying posterior to it than does the third, and may contain lateral mass cells, although I cannot be sure that they enter into its composition. It becomes completely detached from the epidermis in an embryo of 105 hours; while both the third and fourth epibranchial ganglia can be recognized in my oldest series, the fourth fuses much more closely with the remainder of the Xth than does the third.

THE ORIGIN OF THE FIFTH AND SIXTH EPIBRANCHIAL PLACODES

In the case of the fifth and sixth epibranchial placodes conditions are quite different on account of the fact that the lateral mass enters so prominently into the composition of this portion of the Xth. The relations are somewhat confused here on account of the enormous size of the lateral mass at the time the fifth and sixth placodes appear. The earliest trace of a placode I have been able to find for the fourth gill slit is represented in fig. 70 (93 hours) . This placode has the appearance that all the other placodes present at the time they begin to proliferate cells mesially. The appearance usually presented in my preparations is shown in figs. 72 and 73. The large lateral mass ganglion comes into contact with the epidermis and remains in contact about the length of time the other placodes retain their connection with the epidermis. Except for the condition shown in fig. 70 it would not be possible to as


376 F. L. LANDACRE

sert that we had a true placode here. However, I believe that the contact of the lateral mass ganglion with the epidermis is purely a secondary matter and that while it cannot be proven, in all probability the fifth epibranchial placode is contributing cells to the lateral mass ganglion. I can see no other interpretation for the presence of the early stage of the placode before the contact, or the persistance of the contact during the time usually occupied in the proliferation of cells from the placode. As mentioned above, however, I have been unable to separate this ganglionic mass into lateral mass and placodal portions.

Fig. 72 is taken through the middle of the fifth epibranchial placode of a 99-hour embryo. The attachment of the lateral mass to the placode is here slightly anterior to the anterior end of the lateralis Xth, which does not appear in the figures. Fig. 73 is from an embryo of 105 hours. The lateralis Xth and lateral mass are cut through the anterior portions. The placodal thickening is seen to extend ventrally in the epidermis as do those of the second, third and fourth placodes. The attachment to the epidermis does not continue up to the 113-hour stage, and judging by the distance of the ganglion from the epidermis in this stage may disappear some time before this, perhaps between the 105 and 113hour stages.

The sixth epibranchial placode is less prominent than the fifth. The lateral mass ganglion retains its contact a shorter time with the epidermis, which is slightly thickened, and I have not been able with certainty to identify the placode previous to the time the contact is formed. The series from which I describe it are taken from 6 to 8 hours apart, and this is not sufficiently close to be sure that the placode is not visible before the time of contact of the lateral mass with the ectoderm.

In an embryo of 113 hours there is present a second contact between the lateral mass cells and the epidermis. This contact occurs at the posterior end of the lateral mass portion of the Xth ganglion, which is quite large at this time and extends back of the fifth gill slit. The contact is present in an embryo of 113 hours from which fig. 74 was drawn, and must have been formed earlier but does not continue as long as the 120-hour stage. I have been


THE CRANIAL GANGLIA IN AMEIURUS


377


unable to determine to what extent cells move from the placode into the ganglion, since there seems to be no noticeable distinction between placodal cells and lateral mass cells in any of my embryos. The same reasoning applies here, however, that was used in the case of the fifth epibranchial placode. The location, time of appearance, and manner of thickening of the epidermis resemble the third and fourth placodes in their early stages when they are undoubtedly contributing cells to these ganglia. Its more transient character is in keeping with the reduction which has occurred in the fifth epibranchial ganglion as compared with the fourth, but seems more marked. It indicates a reduction of the placodes from posterior to anterior, and is to be associated with the reduction in the number of gills in the bony fishes as compared with cyclostomes and elasmobranchs.

The time at which the contact occurs and the length of the attachment of the IXth and the four epibranchial ganglia of the Xth is shown in the following table :


TABLE IV


Showing the time at which the epibranchial placodes of the ninth and tenth ganglia appear, the time at which they become detached, and the length of time of attachment.


EPIBRANCHIAL PLACODES


TIME OF FIRST APPEARANCE


TIME OF DETACHMENT


LENGTH OF TIME OF ATTACHMENT


Second


hours

56*.

69

75

93 105+


hours

93

99

105

-113

113+


hours 37£


Third


30


Fourth


30


Fifth


-20


Sixth


H — 8




This table shows that the fifth and sixth epibranchial placodes appear in serial order and become detached from the epidermis in the same order. They differ principally in having a shorter total time of attachment. There seems to be no reason for supposing that conditions are different here other than in the reduction of the placodes and in the presence of a well defined lateral mass ganglion which fuses with the placode.


THE JOURNAL OF COMPARATIVE NEUROLOGY AND PSYCHOLOGT, VOL. 20, NO. 4.


378 F. L. LANDACRE

The condition here seems to throw some light on the interpretation given these placodes by former workers. They have so often been described as arising at the time the neural crest ganglion fuses with the epidermis that I am unable to reconcile the conditions existing, under which the placodal ganglia of the IXth and first two divisions of the Xth arise in Ameiurus, with these descriptions, unless there are neural crest cells in the IXth and first two divisions of the Xth, which fuse with the epidermis in an early stage in these types. Until the nerve components have been thoroughly worked out in these types in which the neural crest ganglia have been described as entering into the IXth and first two anterior divisions of the Xth, so that we can be sure as to whether there are general visceral fibres in these nerves, it is useless to speculate as to either their mode of origin or composition. The presence of a neural crest in the region of the IXth and Xth nerves, however, does not prove the presence of neural crest cells in these ganglia, unless they can be definitely traced into them. The absence of a lateralis ganglion and fibres in the IXth nerve in Menidia (Herrick, '99) and apparently in Gadus (Herrick, '00, p. 296) and its presence in Ameiurus (Herrick, '01) indicate how useless it is to speculate as to the composition of these ganglia in one type because we know them in another which is closely related.

LATER HISTORY OF THE LATERAL MASS GANGLIA OF THE Xth

There are two of these ganglia in Ameiurus: (1) the small general cutaneous ganglion, the jugular, situated intra-cranially (Herrick, '01, p. 210) . This ganglion is present in Gadus (Herrick, '00, p. 297), where it is also intra-cranial. In Menidia, however, the ganglion is small and wedged in between the visceral ganglia and is extra-cranial.

(2) Beside the j ugular ganglion there is the large visceral ganglion situated extra-cranially over the fourth and fifth gill slits, the earlier history of which was taken up with the postauditory lateral mass. Some idea of its size at the time it comes into contact with the placodes of the fourth and fifth gill arches can beobtained from


THE CRANIAL GANGLIA IX AMEITJR1 - 379

fig. 73 which, however, is not taken through the largest portion of the ganglion. Herrick ('01) does not give a description of the communis ganglion of the Xth nerve in Ameiurus further than to call attention to the fact that it is typical, and in describing Gadus ('00) calls attention to the fact that it is similar to Menidia. In Menidia (Herrick, '99; the fifth branchial nerve is larger than the other four and gives rise not only to the fibres for the fifth gill arch but also to the fibres for the great visceral and oesophageal rami of the vagus. He also calls attention to the fact that in Alenidia the ganglia of the glossopharyngeus and first branchial ganglia of the vagus are composed of very large cells with medium and small cells intermingled among them, and that as we go toward the caudal end of the ganglionic complex, while there are still found cells of various sizes, the smaller ones become increasingly numerous, and suggests the hypothesis that the larger cells are related to taste buds and the smaller ones to visceral fibres.

The division of the extra-cranial Xth ganglion into four parts can hardly be so distinct in Ameiurus as Herrick describes it in Menidia, since in my oldest series the last two divisions are not easy to distinguish and the fusion is probably much more marked in the adult. At the time the fifth and sixth placodes of the Xth nerve are present the lateral mass ganglion appears as a dense mass of cells lying over the fourth and fifth gill slits. The first two divisions of the Xth which are placodal in origin can be distinguished easily, and the first of these in any series is distinct up to a late stage.

We have here in the Xth nerve, then, a general cutaneous or jugular ganglion derived from the lateral mass and lying intracranially, and an extra-cranial visceral ganglion in which it is not possible to separate the placodal cells from the lateral mass cells which greatly predominate over the cells derived from the fifth and sixth epibranchial placodes. This whole ganglionic mass is described by Herrick as a communis ganglion. We must add to this large extra-cranial mass the two anterior placodal ganglia which however can be distinguished easily from the combined lateral mass and placodal portion. These relations are rendered clear by figs. 79. 80, 81, and 82.


380 F. L. LANDACRfe

The embryological evidence based on the mode of origin of the Xth ganglion strongly supports Herrick's suggestion, based on a study of the adult condition in Menidia, since the general and special visceral ganglionic cells come from two different sources although combined into one ganglionic mass.

The intra-cranial or jugular ganglion is the last of the cranial ganglia to assume definite form, and can be recognized first as a definitive ganglionic mass in an embryo of 113 hours. At this time the backward growth of the ear has carried the lateralis IXth quite near to the future jugular ganglion, but they are separated here as in all preceding stages by a blood vessel. Between the 69-hour stage and the 113-hour stage the jugular ganglion is not definitely formed and appears gradually as a mass of cells surrounding the root of the Xth nerve and extending from near the medulla down around the root of the nerve to the extracranial portion. Preceding the stage of 69 hours, and particularly before the formation of a fibrillated root, it is not possible to distinguish a definitive jugular ganglion from those cells which will form the extra-cranial portion. The history of the jugular ganglion seems to be briefly as follows : First there is a loose mass of cells extending from the brain ventrally to a point where the epibranchial placodes will be formed (fig. 75). This is followed by a stage in which the visceral ganglion develops and the fibrillated root is present, and in which the jugular ganglion surrounds the root, sometimes thicker on one side, sometimes on the other, and extends down over the root to the enlarged extra-cranial portion. This is followed by a stage in which there is a definitive ganglionic mass present, which owing to the development of the cartilaginous skeleton can be designated as intra-cranial. This is connected with the extra-cranial portion by a mass of spindle shaped cells only, lying on and among the fibres of the root of the Xth nerve. These spindle shaped cells I interpret as sheath cells, so that the jugular ganglion is now quite distinct and can be easily differentiated from the extra-cranial communis ganglion after 113 hours in Ameiurus.

There is little change in the appearance of this ganglion until the stage of 81 hours, when the fibrous root of the Xth nerve is


THE CRANIAL GANGLIA IN AMEIURUS 381

present. The ganglion cells then surround the proximal part of the root as indicated in fig. 76, which is taken through the middle of the root and is not exactly transverse to the long axis of the embryo, so that nearly the whole length of the fibrous root shows. The cells occupying the proximal portion of the root where the definitive jugular ganglion appears surround the root, lying anterior and posterior to it.

There is little difference in size between the cells which occupy the position in which the ganglion is later found and those cells surrounding the future root. This condition persists for some time, but in an embryo of 113 hours there is a decided increase in the number and a change in the appearance of the cells as shown in fig. 77, which is taken through the middle of the root. Anterior and posterior to this point the cells are usually grouped on either side of the fibrous portion of the root.

A much later stage (fig. 78) from an embryo of 138 hours shows this ganglion after it has become a prominent portion of the vagus complex. The manner in which the ganglion cells are distributed through the root and their small size explains the difficulty one finds in separating the cells of the future ganglion from the spindle shaped cells of the root. The extra-cranial portion of the lateral mass I shall not describe further in detail. It is so large and increases in size so rapidly that it is by far the most conspicuous structure in this region of the body. It is too large, in fact, to draw under a camera at the same magnification I have used for the other portions of the ganglia. The figures (72 and 73) show it in the early stages and figs. 79, 80, 81 and 82 show its relative, size and relations. There is never any difficulty in locating the bulk of this ganglion after the 81st hour. There is the difficulty of determining its dorsal boundary, mentioned above, since it extends as a thin mass of cells dorsally to join the jugular. Its separation from the jugular is in all but these early stages, quite distinct.

The fact that this lateral mass ganglion consists of two parts, an intra-cranial which is general cutaneous, and an extra-cranial which is visceral supplying visceral surfaces, has an interesting bearing on the relation of spinal and cranial nerves.


382 F. L. LANDACRE

The spinal ganglia derived from the neural crest furnish both general visceral and general cutaneous fibres, but these two components are combined in the same ganglion, although probably having separate terminations in the cord. In the head, however, we have the lateral mass ganglion of the Xth, which probably corresponds to the neural crest ganglion of other types, differentiated into two ganglia, a general cutaneous, the jugular, situated intracranially, and an extra-cranial portion that is in all probability the general vesceral ganglion.

From the preceding description it will be seen that there are two quite distinct sources of origin for the cerebral ganglia in Ameiurus, the lateral mass and the epibranchial placodes. Unlike most types, there is no well defined neural crest. The lateral mass early gives rise to the dorso-lateral placodes represented by the auditory vesicle and the pre- and postauditory placodes, while the remainder of the lateral mass gives rise to the primordia of ganglia and to mesectoderm. The lateral mass doubtless contains beside the dorso-lateral placodes the homologue of the neural crest of other types, but the greater portion goes to form mesectoderm.

Of the components found in the adult ganglia, the special visceral or gustatory come from the epibranchial placodes, while the general visceral come from the ventral portion of the lateral mass. The general cutaneous component comes from the lateral mass also, but typically from the dorsal portion, as in the case of the jugular ganglion. The acustico-lateralis component shows its kinship to the general cutaneous component in that the lateralis ganglia of the Vllth come from the lateral mass. The auditory and lateralis IXth come chiefly from the auditory vesicle but may have lateral mass cells in their composition. They show a somewhat more specialized mode of origin, while the lateralis Xth comes entirely from the postauditory placode and is the most highly specialized in mode of origin of the acustico-lateralis ganglia. The geniculate ganglion was seen to be composed of constituents from both the ventral portion of the lateral mass and from the epibranchial placodes.


THE CRANIAL GANGLIA IN AMEIURUS


383


The visceral ganglion of the IXth nerve seems to be a pure placodal ganglion. The first two epibranchial ganglia of the Xth resemble the IXth, but may possibly contain lateral mass cells. The last two epibranchial ganglia of the Xth are quite small and combine with the large lateral mass portion so that the posterior portion of the visceral Xth is largely lateral mass in origin. These relations are shown in table V :


TABLE V


Showing the source of the various components of the cranial ganglia of Ameiurus. The presence of any given component is indicated by the word present.


DERIVATION


V


VII


VIII


IX X


X


X X


COMPONENT


X CO


Dorsal portion


Present

<


Present


?


•>



Pres- Present ent


General cutaneous

Acustico 1 at era! is


S


Dorsolateral placodes




Present


Present



Pies- Pres- Acusticoent ent lateralis



Ventral portion



Present



9


f


Pres- Present ent


General visceral


Epibranchial placodes



Present



Pres- j Present ent


Pres ent


Pres- Present ent


Special visceral


In this table the general cutaneous, general visceral and acustico-lateralis portions of the Xth are placed over the last two epibranchial ganglia not so much to indicate their segmental position as because they occupy this relative position. The lateralis Xth extends however much posterior to the visceral portion.

A comparison of this table with table II compiled from Herrick's work on Ameiurus (p. 321) shows that the ganglia are as discrete in their mode of origin as are the components of the adult ganglia


384 F. L. LANDACRE

and nerves, with the single exception that we have portions of the acustico-lateralis ganglia, as shown in other types, arising from the same source as the general cutaneous. This is to be explained on the basis of the relationship of the two components. In sharp contrast with this, however, we have distinct sources of origin for the special and the general visceral ganglia which are combined in the preceding table (II) and which in the adult are closely fused, particularly in the geniculate and in the posterior portion of the tenth.

The differences between the two tables may be summarized briefly as follows: The visceral ganglia of the adult in table II (p. 321) are broken down in the table V into those portions derived from the epibranchial placodes, i.e., the special visceral or gustatory ganglia, and the portions derived from the lateral mass, i.e., the general visceral.

The acoustico-lateralis ganglia of the adult in table II are broken down in table V into those portions derived from the lateral mass, i.e., the lateralis Vllth and possibly portions of the vTIIth ganglion and of the lateralis IXth, and those portions derived secondarily from the auditory vesicle and placodes, i.e., all of the lateralis Xth and most if not all of the auditory and lateralis IXth ganglia.

GENERAL SUMMARY

1 . The neural plate in Ameiurus differentiates longitudinally into three regions: a median region, the neural keel, which later becomes the neural tube, and two lateral regions, the lateral masses, separated from the neural keel by constricted areas.

2. After the body has assumed a rounded form, the lateral masses come to lie on the sides of the body still retaining their connection with the neural cord by constricted areas. Part of the lateral mass on either side differentiates into the auditory vesicle and the pre- and postauditory placodes, which are extensions of the vesicle and resemble it in structure. These represent the dorso-lateral placodes of other authors. The remainder of the lateral mass breaks down more or less completely into loose tissue in


THE CRANIAL GANGLIA IN AMEIURUS 385

which the general cutaneous, general visceral and some of the acustico-lateralis ganglia form, the remainder not thus used going to form mesectoderm. In the auditory region all of the lateral mass is converted into the auditory vesicle.

3. The Gasserian ganglion arises near the anterior end of the lateral mass, over the mandibular bar, and is first recognizable as a slightly denser area on either side of which the lateral mass breaks down into mesectoderm. Just posterior to this region the ventral portion of the lateral mass, over the hyoid bar, gives rise to part of the geniculate ganglion which later combines with the portion derived from the epibranchial placode, the two constituents not being separable shortly after fusion.

4. Just anterior to the auditory vesicle a portion of the lateral mass gives rise to the lateralis Vllth ganglia which does not differentiate into the dorso-lateral and ventro-mesial ganglia until later. The posterior end of this ganglionic mass is in contact with the auditory vesicle. Both the Gasserian and lateralis vTIth, and that portion of the geniculate derived from the lateral mass are, at first, small and ill defined, with irregular borders which pass almost imperceptibly into mesectoderm. Between the regions in which these ganglia appear the lateral mass breaks down into mesectoderm.

5. The auditory ganglion arises chiefly, if not exclusively, by the proliferation of cells from the anterior end of the auditory vesicle, but in such close contact with the preauditory lateral mass that one cannot be certain that there are no lateral mass cells in it.

6. The lateralis ganglion of the IXth nerve arises also chiefly, if not entirely, by proliferation of cells from the posterior end of the auditory vesicle, but it is carried by the backward growth of the vesicle into the region of the root of the IXth nerve, where a slight condensation of lateral mass cells is sometimes present, and it may possibly contain lateral mass cells.

7. Between the IXth and Xth ganglia and for some distance posterior to the anterior end of the Xth, the lateral mass breaks down completely into mesectoderm. In the region of the Xth nerve the dorsal portion of the lateral mass breaks down into


386 F. L. LANDACRE

loose tissue, in which the general cutaneous jugular ganglion later appears, while the ventral region which forms the general visceral portion of the Xth retains its continuity to a greater extent. The ventral region of the lateral mass which enters into the Xth ganglion resembles in appearance and corresponds in position to that portion of the preauditory lateral mass which enters into the geniculate ganglion.

8. The preauditory placode which is the anterior continuation of the auditory vesicle extends as far forward as the hyoid gill slit, but between the ear and the hyoid gill slit it breaks down completely into mesectoderm. Its anterior extension disappears at the exact point where the epibranchial placode of the hyoid arch appears but seems to have no direct relation to it other than in its position. The preauditory placode does not give rise to the preauditory lateral line organs, there being a period of some hours between the disappearance of the placode and the appearance of the first lateral line organs.

9. The postauditory placode, which is the posterior extension of the auditory vesicle, becomes detached from the vesicle and moves back by successive differentiations of the epidermis to a point back of the fourth lateral line organ, where it gradually disappears. Before its disappearance there have appeared four lateral line organs anterior to it and at least two posterior to it. It disappears in the region of the fifth, but probably does not give rise even to that. The lateralis Xth ganglion arises by the proliferation of cells from the postauditory placode after it has moved some distance back of the auditory vesicle. After the placode ceases to contribute cells to the ganglion it moves beyond the posterior limit of the ganglion, losing all connection with it, and does not give rise to the lateral line nerve.

10. Both the pre- and postauditory lateral line organs are formed by gradual differentiations of the deeper layer of the epidermis, sometimes singly, sometimes two from a common primordium and are entirely distinct in origin from the placodes.

11. The epibranchial ganglia all have a common mode of origin. The epibranchial placode of the hyoid arch appears first as a thickening of the epidermis dorsal and slightly posterior


THE CRANIAL GANGLIA IN AMEIURI S oS,

to the point of contact of the endodermal pocket of the hyoid gill slit with the epidermis. This thickening is characterized by the irregular arrangement of its nuclei and by the large number of mitotic figures. The thickening of the epidermis is followed by an active proliferation of cells mesially, which come into contact with the ventral portion of the lateral mass in this region. The proliferated mass later becomes detached and after some hours the geniculate ganglion, which is thus composed of a lateral mass contingent and a placodal contingent, assumes definite form and comes into intimate relation with the Gasserian ganglion.

12. The epibranchial placode of the first true gill slit arises in a similar manner, appearing first as a slight thickening lyingdorsal and posterior to the first gill slit. The thickening is accompanied by active mitosis, proliferation of cells, and finally by complete detachment, en masse, of the proliferated cells. The proliferated epibranchial ganglion is in this case, however, apparently a pure placodal ganglion, since no lateral mass cells could be detected entering into its composition. Its dorsal extension comes into contact with the remainder of the IXth before the ganglion is completely detached from the epidermis however.

13. The epibranchial ganglia of the second and third gill arches have exactly similar modes of origin, but their dorsal extensions soon come into contact with the lateral mass portion of the Xth. While the second and third epibranchial ganglia are definite in outline and mode of origin, their proximity to the Xth makes it difficult to be sure that there may not be lateral mass cells in their composition.

14. The epibranchial ganglion of the fourth and fifth true gills, owing to the fact that they are in the region of the large lateral mass ganglion of the Xth, present a somewhat different history. The fourth can be detected before the lateral mass of the Xth comes into contact with it, and while its early stages resemble those anterior to it, it does not become detached before the fusion occurs between it and the lateral mass portion. In the case of the fifth, the epibranchial ganglion cannot be detected in my series before the fusion, so that while there is every reason for thinking that the fifth and sixth epibranchial placodes contribute cells


388 F. L. LANDACRE

to the Xth, the relative amount cannot be determined and the two components cannot be separated. The visceral ganglia of the second and third gill arches are apparently like the IXth, pure placodal ganglia with possibly a small contingent of lateral mass cells, while the remainder of the tenth is composed of a few placodal cells derived from the small fifth and sixth epibranchial placodes united to a large lateral mass contingent.

15. The ganglia described in this paper as general cutaneous, acustico-lateralis, and visceral, have been followed to a late stage and shown to be the ganglia described by the authors under these names. The evidence from Ameiurus is little short of a demonstration that there are separate origins for the general visceral and for the special visceral systems, the former coming from the ventral portion of the lateral mass, or, in other types, from the neural crest, and the latter from the epibranchial placodes. For the acustico-lateralis system, there are two sources of origin. The lateralis Vllth comes entirely from the lateral mass and in other types apparently from the neural crest. The auditory and the lateralis IXth come chiefly from the auditory vesicle but may contain lateral mass cells, while the lateralis Xth comes entirely from the postauditory placode. The description of the acusticolateralis system as a special cutaneous system finds a strong support in embryological evidence. The lateralis Vllth ganglia represent an intermediate stage between the general cutaneous ganglia, and the acustico-lateralis ganglia of the IXth and Xth nerves, resembling the former in mode of origin and the latter in structure and function. The latter are derived from the dorso-lateral placodes represented by the auditory vesicle and postauditory placodes. The general cutaneous ganglia of the Vth and Xth come from the dorsal portion of the lateral mass also, or neural crest of other types.


THE CRANIAL. GANGLIA IN AMEIURUS 389

BIBLIOGRAPHY

Allis, E. P., Jr. The cranial muscles and cranial and first spinal nerves in Amia 1897. calva. Jour. Morph., vol. 12, no. 3.

1889. The anatomy and development of the lateral line system in Amia

calva. Jour. Morph., vol. 2, no. 3.

1900. The lateral sensory canals of Polypterus bichir. Anat. Anz., vol.

17, no. 23.

Bailey, F. R. and Miller, A. M. Text-book of embryology. Chapters 17 and 1909. 18, by Oliver S. Strong. New York.

Balfour, F. M. On the development of the spinal nerves in elasmobranch fishes. 1875. Trans. Royal Soc. London, vol. 166.

1878. A monograph on the development of the elasmobranch fishes. London.

1885. Comparative embryology. London.

Barbieri von Ciro. Ricerche sullo svillupo dei nervi cranici nei teleostei. 1907. Morphologisches Jahrbuch, vol. 37.

Beard, J. On the segmental sense organs of the lateral line and the morphology 1884. of the vertebrate auditory organ. Zool. Anz., Jahrg. 7.

1885-86. The system of branchial sense organs and their associated ganglia in Ichthyopsida. A contribution to the ancestral history of vertebrates. Quart. Jour. Micr. Sc, (n.s.), vol. 26, pp. 95-156.

1887. The ciliary or motor-oculi ganglion of the ophthalmicus profundus in sharks. Anat. Anz., Jahrg. 2, no. 18 and 19, pp. 565-575.

«. 1888. Morphological studies II. The development of the peripheral

nervous system of vertebrates Parti. Elasmobranchii and Aves. Quar. Jour. Micr. Sc, vol. 29 (n.s.) no. 114.

1888 a. A contribution to the morphology and development of the nervous system of vertebrates. Anat. Anz., Bd. 3, p. 874, 899.

1896. The history of a transient nervous apparatus in certain Ichthyopsida. Part I. Raia batis. Zool. Jahrb., Abtheil. fur Anat. u. Ontog., Bd. 9.

Beckwith, Cora J. The early development of the lateral line system of Amia 1907. calva. Biological Bulletin, vol. 14, Dec.

Bunker, F. S. On the structure of the sense organs of the lateral line of Ameiurus

1897. nebulosus. Anat. Anz., Bd. 13, nos. 8-9.

Clapp, Cornelia M. The lateral line system of Bactrachus tau. Jour. Morph., 1899. vol. 15. no. 2.

Cole, Frank J. On the cranial nerves of Chimaera monstrosa Linn, with a dis 1896. cussion of the lateral line system and the morphology of the chorda

tympani. Trans. Royal Soc. Edinburgh, vol. 38, Part III, no. 19.


390 F. L. LANDACiRE

Cole, Frank J. Observations on the structure and morphology of the cranial 1898. nerves and lateral sense organs of fishes with special reference to

the genus Gadus. Trans. Linn. Soc. London, series 2, Zool. 7.

Dohrn, A. Bemerkungen iiber den neuesten Versuch einer Losung des Wirbel1890. thierkopf-Problems. Anat. Anz., Jahrg. 5, Heft 2 u. 3.

Fkoriep, A. Ueber Anlagen von Sinnesorgans am Facialis, etc. Archiv fur Anat. 1885. mid Physiol.

Goette, A. Entwickelungsgeschichte der iinke. Leipzig. 1875.

Goronowitsch, N. Das Gehirn und die Kranialnerven von Acipenser ruthenus.

1885. Morph. Jahrb., Bd. 13.

1886. Untersuchungen fiber die erste Anlage der Kranialnerven bei Salmo fario. Societe Imperiale des Naturaliste, de Moscou, Tome 16.

Harrison, Ross Granville. Ueber die Histogenese des peripheren Nerven19 A I Systems bei Salmo salar. Archiv mikr. Anat. und Entw. Bd. 57.

Herrick, C. Judson. The cranial and first spinal nerves of Menidia: A con 1899. tribution upon the nerve components of the bony fishes. Jour. Comp. Neur. Psych., vol. 9, pp. 153-455.

1900. A contribution upon the cranial nerves of the codfish. Jour. Comp.

Neur. Psych., vol. 10, pp. 205-316.

1901. Cranial nerves and cutaneous sense organs of North American siluroid fishes. Jour. Comp. Neur. Psych., vol. 11, pp. 177-249.

1903. The organ and sense of taste in fishes. U. S. Fish Commission Bul letin for 1902, pp. 237-272.

1906. On the centers for touch and taste in the medulla oblongata of fishes,

Jour. Comp. Neur. Psych., vol. 16, no. 6.

His, W. Untersuchungen iiber die erste Anlage des Wirbelthiereleibes. Leipzig. 1868

1879. Die Anfange des peripherischen Nerven-Systems. Archiv f. Anat.

u. Entwickelungsgesch.

Hoffmann, C. K. Zur Entwickelungsgeschichte des Selachierkopfes. Anat. 1S94. Anz., Bd. 9, pp. 638.

Johnson, Alice J., and Sheldon, Lillian. Notes on the development of the 1886. newt. Quar. Jour. Mic. Sci., 5, No. 4, (n.s.), July, 1886.

Johnston, J. B. Hind brain and cranial nerves of Acipenser. Anat. Anz., Bd. 1898. 14, no. 22-23.

1902 a. The brain of Petromyzon. Jour. Comp. Neur. Psych., vol. 12.

1902/;. An attempt to define the primitive functional divisions of the central nervous system. Jour. Comp. Neur. Psych., vol. 12.


THE CRANIAL GANGLIA IN AMEIURUS 391

Johnston, J. B. Das Gehirn und die Cranialnerven der Anamnier. Merkel und 1902 c. Bonnet's Ergebnisse. Bd. 11.

1905 a. The morphology of the vertebrate head from the viewpoint of the functional divisions of the nervous system. Jour. Comp. Neur. Psych., vol. 15.

19056. The cranial nerve components of Petromyzon. Morph. Jahrb., Bd. 34, Heft 2.

1906. The nervous system of vertebrates. Philadelphia, P. Blakiston's Sons & Co.

Koltzoff. Entwickelungsgeschichte des Kopfes von Petromyzon planeri : Ein 1902. Beitrag zur Lehre iiber Metamerie des Wirbelthierkopfes. Moskau.

Kupffer, C. von. The development of the cranial nerves of vertebrates. Jour. 1891. Comp. Neur. Psych., vol. I.

1894. Studien zur vergleichenden Entwickelungsgeschichte von Ammo, coetes planeri. Miincheyi und Leipzig.

Landacre, F. L. On the place of origin and method of distribution of taste

1907. buds in Ameiurus melas. Jour. Comp. Neur. Psych., vol. 17, no. 1.

Lillie, F. R. The development of the chick. New York. 1908.

Locy, Wm. A. Contribution to the structure and development of the vertebrate j **

1895. head. Jour. Morph., vol. 2JL no. 3. 1 <*< J. 4


,!.


Marshall, A. M. On the early stages*! development of the nerves in birds.

1877. Jour, of Anat. and Phys., vol. 11.

1878. The development of the cranial nerves of the chick. Quar. Jour. Mic. Sci. ..

f 1881. On the head £eji£efs and associated nerves in Elasmobranchs. Quar.

Jour. Mic. Sci.

1882. The segmental value of the cranial nerves. Jour. Anat. and Phys.

Marshall A. M., and Spencer, W. B. On the cranial nerves of Scyllium. Quar. 1881. Jour. Mic. Sci.

Mitrophanow, P. Etude embryogenique sur les Selaciens. Arch, de Zool. 1893. Exper. et Gen., ser 3, t, 1, pp. 161-220. pi. 9-15.

1891. Sur la formation du systeme nerveux perif6rique des vertebres.

Comptes Rendus, Paris, Tome 113, no. 19, pp. 659-562.

Neal, H. V. The segmentation of the nervous system in Squalus acanthias. 1898. Bull. Mus. Com. Zool. Harvard College, vol. 31, no. 7,

Onodi, A. P. Ueber die Entwickelung des sympathischen Nerven-Systems. 1886. Archiv.fur Mikrosk. Anat.. vol. 27.


392 F. L. LANDACRE

Pollard, H. B. Lateral line of Siluroids. Zool. Jahrb. Bd. 5, pp. 525-550. 1892.

1895. The oral cirri of Siluroids and the origin of the head in vertebrates.

Zool. Jahrb., Bd. 8, pp. 379-424.

Platt, Julia B. A contribution to the morphology of the vertebrate head based

1891. on a study of Acanthias vulgaris. Jour. Morph., vol. 5, no. 1.

1895-96. Ontogenetic differeniation of the ectoderm in Necturus. Study II. On the development of the peripheral nervous system. Quar. Jour. Mic. Sci., vol. 38.

Rabl, C. Ueber die Metamerie des Wirbelthierkopfes. Verh. Anat. Gesellsch..

1892. vol. 6, pp. 104-135.

Streeter, G. L. The development of the cranial and spinal nerves in the occipi1904. tal region of the human embryo. Amer. Jour. Anat.

Strong, O. S. The cranial nerves of the amphibia; a contribution to the mor1895. phology of the vertebrate nervous system. Jour. Morph., vol. 10.

Whitman, C. O. The segmental sense organs of the leech. The Amer. Nat., vol. 1884. 18.

1889. Some new facts about the Hirundinea. Jour. Morph., vol. 2.

Wijhe, J. W. Van. Ueber die Mesodermsegments und fiber die Entwickelung des 1882. Nerven des Selachier-Kopfes. Amsterdam.

1886. Ueber die Kopfsegments und die Phylogenie des Geruchsorganes de« 

Wirbelthiere. Zool. Anz., Bd. 9.

Wilson, H. V. The embryology of the sea bass (Serranu satrarius). Bull, of 1891. the U. S. Fish Com. Vol. 9, Washington.

Wilson, H. V. and Mattocks, J. E. The lateral sensory anlage in the salmon 1897. Anat. Anz., Bd. 13, no. 24.

Wright, R. R. On the skin and cutaneous sense organs of Ameiurus. Proc. 1884 a. Canadian Institute, n.s. vol. 2.

1884 b. On the nervous system and sense organs of Ameiurus. Ibid.


ABBREVIATIONS

Au. G., Auditory ganglion. Ep., Epidermis.

Au. G., I and II. First and second divi- Ep. PI. I, First epibranchial placode

sions of the auditory ganglion. (epibranchial placode of the VII

Au. Ves., Auditory vesicle. nerve).

Br. I, II, III, IV, V, First to fifth bran- Ep. PI. II, Second epibranchial placode

chial nerves. (epibranchial placode of the IX

En., Endoderm. nerve).

D. L. VII, Dorso-lateral portion of the Ep. PI. Ill, Third epibranchial placode

lateralis VII ganglion. (first epibranchial placode of the

D. L. M., Dorso-lateral mass. X nerve).


THE CRANIAL GANGLIA IN AMEIURUS


393


Ep. PI. IV, Fourth epibranchial placode

(second epibranchial placode of

the X nerve). Ep. PI. V, Fifth epibranchial placode

(third epibranchial placode of

the X nerve). Ep. PI. VI, Sixth epibranchial placode

(fourth epibranchial placode of

the X nerve). Ep. G. II, Second epibranchial ganglion

(epibranchial ganglion of the IX

nerve). Ep. G. Ill, Third epibranchial ganglion

(first epibranchial ganglion of

the X nerve). Ep. G. IV, Fourth epibranchial ganglion

(second epibranchial ganglion of

the X nerve). G., Remnant of the lateral mass in the

region of the lateralis IX ganglion. Gass. G., Gasserian ganglion. Gen. Com. X., General communis or

visceral ganglion of the X nerve. Gen. G., Geniculate ganglion. J. G. X, Jugular or general cutaneous

ganglion of the X nerve. L. IX, Lateralis ganglion of the IX

nerve. L. G. X, Lateralis ganglion of the X

nerve. L. L. X, Lateral line nerve trunk of the

lateralis X ganglion. L. M., Lateral mass. L. M. G. VII, Lateral mass ganglion

(general visceral ganglion) of the

geniculate. Med., Medulla oblongata. Mes., Mesoderm. Mesec, Mesectoderm. N. C, Neural canal. Post PI., Postauditory placode. Pre. PI., Preauditory placode. R. X, Root of the X nerve. R. Au., Root of the auditory nerve. R. C. D. X., Ramus cutaneus dorsalis

vagi.


D. L., VII, Root of the dorso-lateral portion of the lateralis VII ganglion. Gass., Root of the Gasserian ganglion. Gen., Root of the geniculate ganglion. L. G. X, Lateralis root of the X nerve. S. J, Ramus supra-temporalis J, glossopharyngei, (Herrick, '01). V. L., VII, Root of the ventrolateral portion of the lateralis VII ganglion. A. A., Ramulus acusticus ampulse

anterioris. A. E., Ramulus acusticus ampullae

extrense. L., Ramulus acusticus lagense. N., Ramulus acusticus neglectus. P., Ramulus acusticus ampullae posteriori s. Sac, Ramulus acusticus sacculi. N., Ramulus acusticus recessus utriculi. T. Gass., Trunk of the Gasserian ganglion (supero-lateral strand of Wright). T. Gen., Trunk of the geniculate ganganglion of the VII nerve (infero-mesial strand of Wright). T. V. L. VII, Trunk of the ventrolateral portion of the lateralis VII ganglion (hyomandibular nerve). V. L. VII, Ventrolateral portion of the

lateralis VII ganglion. X, Mass of cells of unknown origin and whose fate is unknown, lying just anterior to the first epibranchial placode. V. L. VII, Ventrolateral portion of

the lateralis VII ganglion. X, Mass of cells of unknown origin and whose fate is unknown, lying just anterior to the first epibranchial placode. Y, Ventral portion of the lateral mass lying in the region of the IX and X ganglia.


THE JOURNAL OF COMPARATIVE NEUROLOGY AND PSYCHOLOGY, VOL 20, NO. 4.


394 F. L. LANDACRE


EXPLANATION OF FIGURES

All sections are cut 7 micra in thickness, and all figures and reconstructions are taken from transverse sections. Figures were drawn with a camera lucida with objective 4 mm. eye-piece, X 8 Spencer, reduction to J. Total magnification, 186 dia.

1 (A. neb., Stage I) is taken at a stage when the medullary plate is slightly differentiated into the neural cord (N. C.) and the lateral mass (L. M.). The section lies just posterior to the point of formation of the optic cup.

2 and 3 (A. neb.) are taken from the same embryo, Stage II. Fig. 2 is taken six sections posterior to the optic vesicle and lies in a region just anterior to the point at which the Gasserian ganglion will form later. The lateral mass here breaks down completely into mesectoderm.

3 is taken just posterior to the future position of the Gasserian ganglion, twentynine sections posterior to the optic vesicle. The greater portion of the lateral mass here breaks down into mesectoderm also.

4 to 9 (A. neb., Stage III) are all from the same embryo which is slightly older than that from which figs. 2 and 3 were taken. Fig. 4 lies four sections anterior to the auditory vesicle and shows the differentiation of the lateral mass into the preauditory placode and the dorsal portion of the lateral mass which later forms the lateralis VII ganglion.

5 shows the differentiation of the lateral mass into dorso-lateral mass and postauditory placode, four sections posterior to the auditory vesicle.

6 is taken through the middle of the auditory vesicle before the appearance of a definite cavity.

7 lies four sections posterior to the optic vesicle. It is taken through the same relative position as fig. 2, Stage I, and shows the transition of lateral mass into mesectoderm.

8 is taken through the regioninwhichthe Gasserian ganglion forms. The lateral mass does not break down into a loose mass of cells so early here as it does just anterior and posterior to the Gasserian ganglion.

9 lies five sections anterior to fig. 4 and nine sections anterior to auditory vesicle. It shows the reduction in size of the preauditory placode as one reads forward from the position of fig. 4.


THE CRANIAL GANGLIA IN AMEIURUS


395



396 F. L. LAX I) ACRE


EXPLANATION OF FIGURES

10 and 11 (A. neb. . Stage V) are from the same embryo. Fig. 10 is taken through the posterior end of the Gasserian ganglion at one of the earliest stages at which it can be definitely located. This section passes through the hyoid gill pocket (En.).

11 is taken through the posterior end of the lateralis VII ganglion and lies four sections anterior to the auditory vesicle.

12 (A. neb., Stage VII) is from an older embryo than that from which fig. 6 was taken. It lies in the mid-region of the vesicle and shows the absence of a lateral mass after the vesicle is formed.

13 is taken from the same embryo as that from which figs. 10 and 11 were taken (A. neb., Stage V). It lies four sections back of the anterior end of the auditory vesicle. The proliferation of the capsule cells to form the auditory ganglion is taking place here.

14 and 15 (A. neb., Stage IV) are from embryos of the same age. Fig. 14 lies two sections anterior to the anterior end of the auditory vesicle. The preauditory placode is here continuous with the vesicle; on the opposite side of the same embryo there is one section, fig. 15, intervening between the preauditory placode and the vesicle.

16 (A. neb., Stage IV) is from a slightly more developed embryo of the same age as figs. 14 and 15 and is taken through the middle of the preauditory placode. There are six sections in which the placode is absent between the posterior end of the placode and the anterior end of the vesicle.

17 (A. neb., Stage V) is taken from an older embryo than the one from which fig. 16 Is taken and lies nineteen sections anterior to the vesicle. The placode is here disintegrating and is entirely absent in the remaining eighteen sections. It is intact anterior to this point.

18 from the same embryo as fig. 17 shows the appearance of the preauditory placode at the posterior third of its contact with the hyoid gill pocket which extends over eleven sections. This is the last recognizable stage of the placode. It still retains a slight resemblance to the earlier stages.

19 to 24 (A. neb., Stage VI) are all from the same embryo. Fig. 19 is taken through the mid region of the hyoid gill pocket and shows the complete disappearance of the preauditory placode.

20, 21, 22 and 23 are consecutive sections, fig. 20 lying at the extreme posterior end of the point of contact of the hyoid pocket whth the epidermis. The small mass of cells (X) lies anterior to the position of the future epibranchial ganglion and does not seem to enter into its composition. These sections show the gradual thickening and irregular arrangement of the cells in the epidermis (epibranchial placode) which precedes the proliferation of cells mesially.


THE CRANIAL GANGLIA IN AMEIURUS


39:






u ' (9/ ■







19


20


a SL" Hi"







'•■■ ■ ^#?


o




8


S'F - 9 ^-~« 


op v




•^


9




398 F. L. LANDACRE


EXPLANATION OF FIGURES

24 lies four sections back of fig. 23, the epidermis is reduced in thickness and the placode is not so long dorso-ventrally. Just back of this the epidermis is of normal thickness.

25 to 30 (A. neb., Stage VII), illustrating the origin of the epibranchial portion of the geniculate ganglon, are all from the same embryo which is slightly older than the one from which figs. 19 to 24 were taken. Figs. 25 to 2S are consecutive. Figs. 29 and 30 are consecutive, one section intervenes between figs. 28 and 29. Fig. 25 lies near the posterior limit of the contact of the hyoid pocket with the epidermis and fig. 26 at the extreme posterior limit. Active cell division is taking place here. In fig. 27 the ganglionic mass is proliferated mesially over the hyoid pocket which is no longer in contact with the epidermis. In fig. 28 the ganglionic mass is still purely placodal in origin, but in fig. 29 the placodal ganglion is in contact with a slightly delimited mass (L. M. G. VII) derived from the lateral mass. In fig. 30 the lateral mass portion of the ganglion predominates. A few sections posterior to this point the ganglion is entirely of lateral mass origin and the epidermis is of normal thickness.

31 to 39 (A. melas, 86 hours) illustrate the relations of the Gasserian ganglion to the geniculate and the dorso-lateral and ventro-lateral lateralis ganglia of the VII nerve. Figs. 37, 38 and 39 also show this relation to the auditory ganglion and vesicle. Preceding the stage of 86 hours it is not possible to differentiate between the two lateralis ganglia of the VII nerve and following this stage the ganglia soon becomes condensed and it is difficult to unravel them.

31 is taken throughthe trunk of the nerve (supero-lateral strand of Wright) of the Gasserian ganglion.

32 is taken just anterior to the point of origin of the nerve (infero-mesial strand of Wright) of the geniculate ganglion. These two strands combine in a later stage some distance from the ganglia and then split into the maxillary and mandibular nerves.

33 is taken through the root of the Gasserian ganglion.


THE CRANIAL GANGLIA IN AMEIURUS


399



. % tea


^m>


-$■


v^S'


m „ ^- -^U 77 © ' *1 /-■ .


UK





27


32


C?


J


■' mJ> If


TO


a



28.




®

Ei- Pl 1





?e>


31.

,# !





,J m.m-,


& 'Mm fcgSMp


<*>:


& .


e I


33.


efm




w^m '?

to <5e> ® °~ s ' >e


24.

I (2

to


•■©.'3




\


%!>!


$&4>'<X




<*>>


'00 mi


s#>m




400 F. L. LANDACRE


EXPLANATION OF FIGURES

34 is taken through the anterior end of the dorso-lateral lateralis VII ganglion and the root of the Gasserian.

35 is taken through the trunk (hyomandibular nerve) of the ventral lateralis ganglion of the VII and the root of the geniculate ganglion.

36 is taken through the root of the ventral lateralis ganglion.

37 is taken through the extreme anterior tip of the auditory ganglion. The dorsal lateralis VII runs through figs. 34 to 37.

38 and 39 are taken through the anterior and median portions of the auditory ganglion. The roots of the dorso-lateral, ventro-lateral and geniculate ganglia appear in figs. 37, 38 and 39.

40 to 43 (A. neb., Stage VII) are taken from the same embryo as figs. 25 to 30. Figs. 40 and 41 are consecutive sections and show the only recognizable trace of a condensation of the cells derived from the lateral mass in the region of the IX nerve, except the ventral portion of the lateral mass (Y) which does not enter into the IX ganglion.

40 is taken one section posterior to the auditory vesicle.

41 is taken two sections posterior to the auditory vesicle.


THE CRANIAL GANGLIA IN AMEIURUS


401



- -R G »"ss





U L.vnRCsv R V L vii. V L.-vii T V I vii


35. ~ "TE*T ^ 4N£#f







--D L -B.G




11 W^M&Smk £%k. w*- - a)



R.D.L vii

R o«« ■

R.V r. v,i


Au.O |.


d 00




dF




«$y*jm&>





402 F. L. LANDACRE


EXPLANATION OF FIGURES

42 is taken five sections posterior to the auditory vesicle and lies just back of the point where the root of the IX later appears. The lateral mass is here completely converted into mesectoderm except the ventral portion (Y).

43 is taken through a condensation of lateral mass cells at the point where the root of the X and the jugular ganglion later appear. This mass cannot be located in later series for some time.

44 to 52 illustrate the detachment of the postauditory placode from the auditory vesicle, its migration and the formation of the lateralis X ganglion.

44 and 45 (A. neb., Stage IV) are consecutive sections. Fig. 44 being taken just at the point where the auditory vesicle passes into the placode by the disappearance of the dorsal half of the vesicle. Fig. 45 being near the anterior end of the placode, three sections posterior to the auditory vesicle.

46 is taken three sections posterior to the auditory vesicle and is from a slightly less developed embryo of the same age as figs 44 and 45. The vesicle passes gradually into the placode in this series.

47 to 51 (A. neb., Stage VII) show the characteristic appearance of the lateralis X ganglion as it is proliferated from the placode before the placode has moved beyond the posterior limit of the ganglion. Fig. 47 is twenty sections posterior to the vesicle ; fig. 48 twenty-three sections ; fig. 49 twenty-five sections ; fig. 50 twentyseven sections; and fig. 51 twenty-nine sections posterior to the auditory vesicle. Fig. 50 is at the extreme posterior limit of the lat. X ganglion.

52 (A. neb., Stage IX) shows the usual appearance of the lateralis X ganglion in a somewhat older series.

53 to 60 (A. melas) illustrate the formation and detachment of theepibranchial placode of the IX ganglion.

53 (A. melas, 56* hours) shows the appearance of the placode when it can first be detected.

54 (A. melas, 69 hours) shows the extent of the thickening of the placode while the whole placode is still in contact with the epidermis.

55 to 59 are from the same embryo (A. melas, 81 hours) at a stage when the posterior end of the placode is detached from the epidermis and has formed an attachment to the root of the IX nerve which contains motor fibers and lateralis ganglion cells.

55, 56 and 57 are consecutive. Figs. 55 and 56 show the appearance of the ganglionic mass just before it becomes completely detached. The anterior end of the ganglion is attached, for some time after the posterior end is free, apparently by the growth dorsally and mesially of the ganglionic mass.


THE CRANIAL GANGLIA IN AMBIURUS


403



^z&y o- o- ^ ^ ,


-; " ■&£> g'rk&'^bj






404 F. L. LANDACRE


EXPLANATION OF FIGURES

57 is taken one section posterior to fig. 56 at the point where the ganglionic mass is free from the epidermis.

58 and 59 are taken through the root of the epibranchial ganglion of the IXth just before it comes into contact with the lateralis and motor portions of the IX, as shown in fig. 59.

60 (A. melas, 93 hours) is taken through the epibranchial ganglion at the nearest point of approach to the epidermis. This is the first stage in which the ganglion is completely detached from the epidermis. Its clean cut boundaries during all of its development indicate that no other cells than those derived from the placode are incorporated in the ganglion.

61, 62 and 63 (A. neb.) illustrate three stages in the formation of the lateralis IX ganglion of which fig. 63 (A. neb., Stage V) is the earliest. The ganglion varies a great deal in appearance and somewhat in position. Fig. 62 (A. neb., Stage IX) is older than fig. 61 (A. neb., Stage VIII). Fig. 63, while taken from the posterior end of the vesicle, differs totally from the appearance of this region before and after the formation of the ganglion.

64 to 71 show the formation of the first two epibranchial ganglia of the X.

6-4 (A. melas) 69 hours shows the earliest recognizable trace of the third epibranchial placode. Figs. 65 to 71 are from the same embryo (A. melas, 93 hours).

65 is through the anterior portion of the placode.

66 is through the middle region just before the detached portion is reached.

67 is just backof the point of detachment.

68 is through the root of the third epibranchial ganglion and the fourth epibranchial placode. This probably is not the earliest trace of the fourth epibranchial placode, however. The preceding series is defective at this point.


THE CRANIAL GANGLIA IN AMEIURUS


405


4 /- 1



40() F. L. LAND ACRE


EXPLANATION OF FIGURES

09 (A. melas, 93 hours) is taken through the detached portion of epibranchial ganglion IV, also shows the root of epibranchial ganglion III.

70 (A. melas, 93 hours) is taken through the fourth epibranchial ganglion just before it comes into contact with the lateral mass ganglion of the X (general communis X). In fig. 70 the root of the third epibranchial ganglion has become so attenuated that it cannot be recognized with certainty, alhough the two cells marked Ep. gl. Ill appear to be the posterior extension of this ganglion.

71 (A. melas, 93 hours) shows the point of union of the fourth epibranchial ganglion with the lateral mass of the X (Gen. vis. X).

7(i. 72, 73 and 74 illustrate the conditions under which the fifth and sixth epibranchial placodes arise. In fig. 70 (A. melas, 93 hours) the fifth epibranchial placode appears before the lateral mass ganglion of the X comes into contact with the epidermis. In fig. 72 (A. melas, 99 hours) the attachment is quite like that in fig. 73 (A. melas 105 hours). In both these cases the attachment of a lateral mass (neural crest) ganglion to the epidermis is quite evident and corresponds to the oft repeated descriptions in the literature of the contact formed between neural crest ganglia and the epidermis. Nothing resembling this occurs in Ameiurus except in the fifth and sixth epibranchial placodes.

74 (A. melas, 113 hours) is taken through the attachment of the general communis X to the sixth epibranchial placode.

75 to 78 illustrate the formation of the jugular ganglion of the Xth. Fig. 75 (A. melas, 69 hours) shows the slight condensation of cells in the future position of the jugular ganglion before the appearance of the root of the X.

76 (A. melas, 81 hours). The cells which later form the jugular ganglion inclose the fibrous root of the X.

77 (A. melas, 113 hours) shows the first decided increase in size of the jugular ganglion when it begins to mass itself into definite areas of cells such as appear in fig. 78 (138 hours) where the ganglion is broken into small masses by the fibrillated root of the X.


THE CRANIAL GANGLIA IN AME1URUS


407






vi - •


70.

f




73.


4

©>


&J

fi


i-v"*tf»>^ - ■ •


m &.Vv9>Q©


mm






7L



^^■"



llllisQl




/



^


M^i^


75. 6


^§V©



76.




(9 * C^"


<3>



© 0® 9© ©


78.

M f

riiUT

■' si i?

% m4 M


•0


©0


'0




Ul 1


40S F. L. LANDACRE


EXPLANATION OF FIGURES

79 to 83 are reconstructions of the cranial ganglia of A. melas and are true in two dimensions, the anterior-posterior and the dorso-ventral. The reconstructions were made by projecting the section of the ganglion on paper to determine the vertical length of a given section. The lens of the eye, which is an almost perfect circle in these diameters, was used as a basis for determining the ratio of longitudinal to vertical dimensions. The reconstructions give an approximately exact picture of the lateral view of the ganglia as seen on a flat surface. The figures give no idea of the relative thickness of the ganglia and many of the roots appear as large as the ganglia while in reality they are quite thin, sometimes not more than one cell thick, cf. the roots of the II, III and IV epibranchial ganglia, figs. 79, 80 and 81. The object in making these reconstructions was to trace the embryonic ganglia up to a stage where they could be positively identified as the definitive ganglia of the adult.

79 is a reconstruction of the ninth and tenth ganglia of A. melas, SI hours; only three epibranchial placodes are present at this stage.

SO is a reconstruction of the ninth and tenth ganglia of A. melas, 93 hours. Four epibranchial placodes are here present.

81 is a reconstruction of the eighth, ninth and tenth ganglia of A. melas 138 hours. All placodes have disappeared some hours previous to this stage.


THE CRANIAL GANGLIA IN AMEIURUS


409



THE JOURNAL OF COMPARATIVE NEUPOLOOY AND PSYCHOLOGY, VOL. 20, NO. 4.


410 F. L. LANDACRE


EXPLANATION OF FIGURES

82 is a reconstruction of the eighth, ninth and tenth ganglia of A. melas, 171 hours.

83 is a reconstruction of the fifth, seventh and anterior portion of the eighth ganglia of A. melas. 86 hours. The roots of the V and VII ganglia are slightly schematized beyond their origin from the ganglia for the sake of clearness.


THE CRANIAL GANGLIA IN AMEIURUS


411



oeneru/ so/naT/c enera/ v/^cera/ oec/a/ somaT/c

soec/a/ v/sce/a/


STANDARD SIZES FOR ILLUSTRATIONS

IN THE JOURNALS PUBLISHED BY

THE WISTAR INSTITUTE OF ANATOMY AND BIOLOGY

SUGGESTED BY PROFESSOR SIMON HENRY GAGE

AND APPROVED BY THE ADVISORY BOARD OP THE WISTAR INSTITUTE AT ITS LAST MEETING

(APRIL 1910)

STANDARD MAGNIFICATIONS

It is desirable that a uniform system of magnifications should be followed in making illustrations for publication.

This system has been adopted: 1, 2\, 5, 10, 15, 20, 25, 30, 35, 4.0, 45, 50, 60, 70, 75, 80, P0, 100, 125, 150, 175, 200, 250, 300, 350, 400, 450, 500, 600, 700, 800, 900, 1000, 1250, 1500, 2000.

Authors are requested to make their drawings of such size that when reduced for publication the magnification will correspond exactly with one of the numbers of this system.

The relative size of every published figure should be given thus: State whether it is (1) of the same size as the object; (2) magnified,

as indicated above; or (3) smaller than the object, as Jr, i, j l , etc. natural

size.

DRAWINGS AND THEIR REDUCTIONS

1. If there is to be no change in size by the engraving process, the drawing must be of the same size as the finished engraving.

2. If the drawing is to be reduced f, then it must be twice the size of the engraving, (i.e., every line in the drawing must be twice the length it will appear in the engraving — twice the diameter).

3. If the drawing is to be reduced f, it must be made If times the size of the engraving.

4. If the drawing is to be reduced \, it must be made 1| times the size of the engraving.

Examples. Suppose the illustration (as it will appear in one of the journals) is to be at a magnification of 40. The size of the drawing for each of the above would then be as follows:

(1) Full size, or no reduction — each line of the drawing must lie 40

times the length of the corresponding line in the object.

(2) For | reduction- — each line must be 80 times the length of the

corresponding line of the object.

(3) For I reduction — each line must be 60 times the length of the

corresponding line of the object.

(4) For I reduction — each line must be 50 times the length of the

corresponding line of the object.

HOW TO MARK A DRAWING TO BE REPRODUCED.

(1) Full size (i.e., no reduction), (2) \ off; (3) | off; (4) \ off; (5) \ off.


THE MORPHOLOGY OF THE FOREBRAIN IN AMPHIBIA

AND REPTILIA

C. JUDSON HERRICK

From the Anatomical Laboratory of the University of Chicago

EIGHTY-FOUR FIGURES

Introduction 413

Amphibia 415

Urodela 416

Amblystoma 416

Othei Urodela 435

Anura 438

Comparisons with fishes 449

Reptilia 452

Discussion 466

Diencephalon and telencephalon 466

The paraterminal body , 483

The morphology of the cerebral cortex 486

The subdivision of the prosencephalon 492

Summary 498

INTRODUCTION

Notwithstanding the publication of a great mass of descriptive detail regarding the structure and morphology of the forebrains of lower vertebrates, it is very difficult to form a clear picture of the fundamental morphological features of the vertebrate cerebral hemisphere. This wealth of observation has stubbornly resisted correlation and the morphological fruits of these arduous labors have until very recently, it must be confessed, been disappointingly meager. We have, however, now reached a point where effective correlation has begun to take form and within the bewildering complexity of detail characteristic of individual species, it is possible to see a common morphological pattern which is sur THE JOURNAL OF COMPARATIVE NEUROLOGY AND PSYCHOLOGY, VOL. 20, NO. 5.


414 C. JUDSON HERRICK

prisingly constant and very simple. In this contribution I have brought together certain features of embryonic and adult brains of amphibians and reptiles which illustrate the fundamental simplicity of this pattern

Johnston has recently presented strong evidence that the telencephalon must be regarded as the terminal segment of the neural tube, a view confirming (with some modifications) the original teachings of His as opposed to the usage in the BNA tables In the latter the diencephalon is regarded as extending to the extreme rostral end of the primary neural tube, thus comprising the whole of the unpaired ventricle of this part of the brain and its lateral walls, including the lamina terminalis, while the telencephalon is regarded as limited to the secondarily evaginated parts of the neural tube termed the cerebral hemispheres, viz., the lateral ventricles and their massive walls. The usage of His and Johnston implies that the rostral part of the third ventricle, bounded behind by the velum transversum above and the chiasma-ridge (Johnston) below, and its walls are to be regarded as telencephalon . medium, while the evaginated hemispheres constitute the telencephalon laterale. Johnston has further shown that in lower vertebrates there has been a progressive tendency as we pass up the developmental series (both ontogenetic and phylogenetic) for more and more of the telencephalon medium to be evaginated through the interventricular foramina into the hemispheres.

These considerations have an obvious bearing on the problem of the relation of the cerebral cortex to the primordial tissues from which it has been differentiated. With a view to the contribution of further data for the solution of this problem, I have examined the embryonic and adult brains of a series of types of lower vertebrates, the first results of which are presented in this paper.

I shall discuss the brains of fishes only incidentally and devote my attention chiefly to the amphibians and reptiles, whose cerebral hemispheres have evaginated so far from the primordial neural tube as to present a form approximating more closely to the mammalian conditions and readily leading up to them.


MORPHOLOGY OF THE FOREBRAIN 415

Our immediate problem, then, is the relations of the first recognizable primordia of the cerebral cortex to the other elements of the evaginated cerebral hemisphere and of all of these structures to the more ancient tissues of the telencephalon medium and diencephalon.

My indebtedness to the published works of Johnston will be evident to the reader throughout this paper. I have received still greater assistance from many extended conferences with Professor Johnston, in which he has freely shared with me his unpublished observations and stimulatng suggestions. The full extent of this obligation it is not necessary, nor indeed possible, for me to indicate here. It should, moreover, be added that, while many parts of this discussion have been greatly influenced and I trust improved by these conferences, the responsibility for the morphological views here expressed is wholly my own.

AMPHIBIA

I have studied an extensive series of sections of larval and adult Amblystoma, Necturus and various species of frogs, prepared by different methods, including the silver methods of Golgi and Ramon y Cajal, the method of Weigert, a toluidin blue modification of Nissl's method and various general embryological methods. Most of this material, except the larval Necturus which I studied through the courtesy of Professor Minot in the Harvard Embryological Collection, was prepared by Mr. P. S. McKibben of the Department of Anatomy, University of Chicago, to whose kindness and skill I am greatly indebted. I have also examined a series of cross sections through the head of Petromyzon (Ichthyomyzon concolor) prepared and kindly loaned to me by Dr. Charles Brookover.

In the Amphibia the wall of the cerebral hemisphere is naturally divided into five parts. Not to prejudice the morphological significance of these parts at the start, I shall call them simply olfactory bulb, ventro-medial, ventro-lateral, dorso-lateral and dorso-medial parts. They are especially distinct in the adult frog and are termed by Gaupp respectively lobus olfactorius,


416 C. JUDSON HEBPICK

eminentia septalis, ganglion basale, format io pallii lateralis and eminentia pallii medialis. The two ventral parts are Gaupp's pars inferior s. subpallialis hemisphaerii and the dorsal parts are his pars pallialis (Gaupp, '99, fig. 29, p. 107). The dorsal and ventral parts are sometimes separated by well marked total fissures or by a conspicuous difference in internal structure. The fissure which separates the two lateral parts is the sulcus limitans lateralis of Gaupp and is incompletely homologous with the fissura endo-rhinalis (Turner) of other classes of vertebrates. The two medial parts of the frog are also separated by a total fissure, the fissura septo-corticalis of Kappers. Gaupp names this fissure the sulcus intermedins on the ventricular side and fissura arcuata on the superficial side. For reasons which will appear in the subsequent discussion, I name it fissura limitans hippocampi. The positions of both fissures in the frog are marked internally by a characteristic disposition of cells and fibers termed the zona limitans (lateralis et medialis), and the zona limitans may be presentas a useful landmark in cases where the corresponding fissure is not externally evident.

URODELA

Ainbly stoma

We will first examine a series of transverse sections through the brain of a specimen of larval Amblystoma 17 mm. long and about 35 days of age after fertilization. The relations are very similar to those of the larvae of Necturus described b} T Warren ('05, 18 mm. stage) and by Kupffer ('06, 24 mm. stage) though with a less pronounced flexure in the thalamus region.

At the level of the olfactory bulb (fig. 1) the wall of the hemisphere is massive on all sides and the five parts as defined above are not separate. Secondary olfactory fibers (tractus olfactorius) are present around almost the whole periphery. The olfactory bulb is limited to the lateral aspect of the hemisphere.

A short distance caudal to the bulb (fig. 2) the greater part of the medial wall becomes membranous. This septum ependymale separates the dorso-medial from the ventro-medial part of


MORPHOLOGY OF THE FOREBRAIN 417

the hemisphere, both of the latter being relatively small. The medial olfactory tract divides into a dorsal part passing into the massive dorso-medial wall and a ventral part for the medial wall below the septum ependymale. The former joins the fimbria complex. The lateral olfactory tract is also divided into dorsal and ventral parts, the former {tr. olf. lat.) running separately to the pars dorso-lateralis of the hemisphere and the ventral being confused with the lateral forebrain tract and ending in the pars ventro-lateralis. The tract marked tr. o. is a mixture of the ventral parts of both the medial and lateral olfactory tracts and the medial and lateral components of the basal forebrain bundles, all of which are separate in the frog.

Midway of the hemisphere the septum ependymale is interrupted by the interventricular foramen and from this point caudad the ventriculus lateralis communicates widely with the ventriculus impar of the forebrain (fig. 3) The wide membranous roof over the ventriculus impar is a choroid plexus which is extended laterally to form the plexus lateralis Farther caudad this membrane is evaginated dorsally to form the paraphysis and backward into the third ventricle as velum transversum and diencephalic plexus, both of which at this age are very small. The line of contact of the roof membrane with the massive dorso-medial wall of the hemisphere is the taenia fornicis. The caudal border of this membrane is attached to the thalamus along the taenia thalami, where for a short distance the taenia fornicis and the taenia thalami come together (fig. 4). At the line of contact of the hemisphere and the thalamus there is a membrane stretching across from the taenia fornicis et thalami of one side to the other. This is the locus of the velum transversum (fig. 4). Passing forward from this point are three separate membranes: (1) the forebrain roof and the plexus lateralis already described; (2) the paraphysis (figs. 2 and 3); (3) the plexus chorioideus ventriculi tertii or diencephalic plexus. For further details regarding these membranes I refer to the excellent account of Necturus by Warren ('05).

At the levels of figs. 2 and 3 the pars ventro-medialis is very small with no recognizable pre-commissural body (nucleus media


418 C. JUDSON HERRICK

nus septi), the two lateral parts are imperfectly separated by a shallow ependymal groove which is the precursor of the fissura endorhinalis, and the dorso-medial part is sharply inflected from the dorso-medial angle to the taenia fornicis. The pars dorsomedialis becomes the primordium hippocampi of the adult.

The rostral ends of all four parts of the hemisphere are reached by fibers from the olfactory bulb and fibers of the tractus olfactorius dorso-lateralis follow the whole length of the pars dorsolaterals and reach the posterior pole. The basal forebrain bundle contains two chief components which characterize respectively the ventro-lateral and the ventro-medial parts of the hemisphere, as seen best in the adult frog. The mingling of these components in urodeles is correlated with the imperfect separation of the two ventral parts of the hemisphere. The lateral and medial forebrain tracts can, however, be distinguished in Cajal preparations by the presence of much coarser fibres in the former (fig. 3). The medial forebrain tract connects chiefly with the hypothalamus, the lateral with the thalamus and mid-brain. Both decussate partially in the anterior commissure (fig. 5).

The taenia fornicis is accompanied by a mixed fiber tract whose composition varies in different parts and which I term the fimbria complex. At the rostral end it receives the dorsal component of the tractus olfactorius medialis (figs. 1 and 2) and comparison with the adult and with Anura suggests that here also it probably contains elements of the primordial columna fornicis system passing rostrally of the interventricular foramen and anterior commissure between the primordium hippocampi and the pars ventromedialis. Farther back this tract contains fibers for the commissura pallii anterior (corn, hippocampi), others for the stria medullaris and perhaps thalamic connections, all of which leave the primordium hippocampi in the region of the posterior pole caudal to the foramen interventriculare (see pp. 427 ff.).

The dorso-medial and dorso-lateral parts of the hemisphere converge at the posterior pole, where their distinctive characteristics are lost. This applies to the adults also of both Urodela and Anura.

The relations of telencephalon and diencephalon are difficult


MORPHOLOGY OF THE FOREBRAIN 419

to determine from sections in these specimens, on account of the strong diencephalic flexure. The caudal lip of the velum transversum passes over into the diencephalic plexus and, farther laterally and ventrally, into a ridge which forms a massive connecting bridge between the thalamus and the hemisphere which I shall term the eminentia thalami. This is bounded dorsally and ventrally by sharp ependymal grooves (fig. 5) which converge anteriorly to the interventricular foramen. The dorsal one of the these I shall term the sulcus diencephalicus medius, the ventral one the sulcus diencephalicus ventralis. The sulcus medius extends caudad through the whole length of the diencephalon, turning ventrally behind to join the sulcus limitans in front of the tuberculum posterius. As we shall see beyond (pp. 431 and 469) the sulcus medius is functionally an extension of the sulcus limitans (which ends in the preoptic recess), and the two sulci together in the diencephalon are sometimes termed sulcus Monroi. Kupffer ('06, p. 181, fig. 193) designates in Necturus the sulcus ventralis as sulcus Monroi, but this is evidently inappropriate, for this ventral sulcus separates the hypothalamus from the eminentia thalami (see pp. 431 and 469 ft\). Immediately behind the anterior commissural ridge in Amblystoma the ventral sulcus divides. One part follows the caudal border of the ridge into the preoptic recess. This is the ventral part of Kupffer's sulcus interencephalicus anterior (cf. '06, p. 175, fig. 187). The other part continues caudad dorsal to the preoptic nucleus and chiasma ridge to terminate blindly in the caudal part of the hypothalamus. It marks the boundary between the preoptic nucleus and hypothalamus and the thalamus in front and at its posterior end separates the hypothalamus proper from the tuberculum pOsterius, which probably belongs morphologically with the pars ventralis of the mesencephalon .

Still farther dorsal is a short sulcus extending (morphologically) caudad and dorsad from the interventricular foramen and forming the ventral boundary of the habenula. It is the sulcus diencephalicus dorsalis.

These relations come out much more clearly in older larvae and adults after the straightening *of the diencephalic flexure.


420 C. JUDSON HERRICK

They conform in morphological type to those figured by Sterzi in Acanthias ('09, p. 577, fig. 232).

Sections of much younger Amblystoma larvae of 10 mm. (about 15 days) so oriented as to cut the brain horizontally in the region of the velum trans versum (figs. 6 and 7) illustrate clearly the relations of the velum (r) to the eminentia thalami and adjacent parts and the communication of the diencephalic and telencephalic ventricles through the wide aula in front of the velum.

The only massive connection between the hemispheres and the diencephalon at this age is by way of the eminentia thalami or structures lying farther ventrally which connect the ventral parts of the hemisphere with the ventral nuclei of the thalamus below the sulcus medius. The dorsal parts of the hemisphere are not well differentiated at this age, but comparison with older stages of Amblystoma and other Amphibia furnishes abundant confirmation of the statement last made.

The sections figured are so inclined as to show clearly that the connection between the dorsal part of the posterior pole and the wall of the thalamus is at this age wholly membranous (figs. 6 and 7, z). That is, it is comparable with the posterior chorioidal fold which is so conspicuous a feature of the brains of all embryonic reptiles and mammals. The fact that mitotic figures are more abundant around this angle of the lateral ventricle than elsewhere suggests that this is the point of most rapid growth at this age. In immediately following stages this fold is all incorporated into the extensive lateral plexus whose earliest rudiment is seen in fig. 6 at y (cf. fig. 3) and the massive tissue of the posterior pole rests in immediate contact with the eminentia thalami (figs. 3-5). There is at no stage in any of the Amphibia which I have examined a direct massive connection between the dorsal parts of the cerebral hemisphere and the pars dorsalis thalami (dorsally of the sulcus medius) or the epithalamus. Here, as in Anniota, the important fibrous connections between these dorsal parts of the telencephalon and diencephalon all cross the dorsal barrier interposed by the velum transversum and di-telencephalic fissure in the massive ventral parts. See p. 430 and the discussion on pp. 474 and 486.


MORPHOLOGY OF THE FOREBRAIN 421

The following description is based upon Amblystoma larvae between 30 mm. and 40 mm. in length, specimens taken at the time of metamorphosis and adults.

In all of these specimens the olfactory bulb (consisting of the glomeruli, mitral cells and granule cells) is confined to the lateral and extreme rostral parts of the hemisphere. In reading a series of cross-sections backward, as soon as the lateral ventricle appears the olfactory bulb is found to lie wholly laterally of it; but rostral to this level the layer of granule cells borders the whole median surface of the section. This short region where the granular layers of the two hemispheres are closely approximated (fig. 8), marks the site of the interbulbar union of the anuran brain. This figure from an adult brain shows medullated and unmedullated fibers from the mitral cells passing through the granular layer to accumulate on the median border of the hemisphere (fig. 9). Fibers of the tractus olfactorius medialis at the levels of these figures come from both the dorsal and ventral parts of the olfactory bulb, but farther caudad from the ventral part only.

In the adult at the rostral end of the olfactory bulb the mitral cells are separated from the glomeruli by a wide molecular layer. There are clusters of cells among the glomeruli which correspond with the subglomerular cells of Rubaschkin's description ('03) and probably with the periglomerular neurones of Cajal. Farther caudad the mitral cell layer is less compact and its cells spread throughout the molecular layer.

Beginning at the rostral tip of the lateral ventricle (fig. 9) the median wall of the hemisphere is occupied by an extensive and undifferentiated secondary olfactory nucleus, which I term the nucleus olfactorius anterior, and which as we pass caudad spreads through the medial and dorsal walls (fig. 10), the olfactory bulb occupying the whole ventro-lateral wall. Medullated and unmedullated secondary olfactory fibers pass from the olfactory bulb into the dorsal border of the anterior olfactory nucleus and continue caudad in this relation as tractus olfactorius dorso-lateralis. In a similar way medullated and unmedullated fibers curve around the ventral angle of the lateral ventricle to form the dorsal and ven


422 C. JUDSON HERRICK

tral divisions of the tractus olfactorius medialis. The ventral niedullated tract arises only from the rostral end of the bulb. All of its fibers, which are few in number, terminated soon in the nucleus olfactorius anterior. The extent of distribution of the unmedullated fibers I have not determined, for the dorsal division enters the rostral end of the primordium hippocampi where its fibers are mingled with those of the fimbria complex and the ventral division is mingled with the median forebrain tract.

The tractus olfactorius dqrso-lateralis arises from the whole length of the dorsal border of the olfactory bulb. The total number of niedullated fibers is, accordingly, quite large (figs. 8 to 11). These niedullated fibers are, however, all short, ending in the adjacent gray of the nucleus olfactorius anterior and pars dorso-lateralis of the hemisphere. The niedullated tract does not increase in size as we approach the caudal end of the bulb and all its fibers terminate a short distance farther caudad (fig. 12). The accompanying unmedullated fibers doubtless extend farther caudad and reach the posterior pole as in the larva and the Anura, though my preparations do not demonstrate this in adult Amblystoma.

The tractus olfactorius ventro-lateralis arises from the caudal end of the olfactory bulb (corresponding with the bulbulus accessorius of the frog) and, as in the frog, passes directly back close to the ventricular ependyma to end in a cellular thickening at the caudal end of the pars ventro-lateralis opposite the anterior commissure, which corresponds with the so-called corpus striatum of the frog.

As we approach the caudal end of the olfactory bulb (fig. 11) the medial wall of the hemisphere becomes . specialized into two structurally defined regions, the primordium hippocampi above and the nucleus post-olfactorius below, the latter corresponding to the eminentia post-olfactoria (Gaupp) of the frog brain and probably to the tuberculum olfactorium of mammals. The dorsal wall remains undifferentiated and is continued caudad into the pars dorso-lateralis of the hemisphere. The latter is, accordingly, to be regarded as the direct continuation of the nucleus olfactorius anterior.


MORPHOLOGY OF THE FOREBRAIN 423

The dorso-medial wall of the hemisphere from this point caudad to the posterior pole has the same structure and fiber connections as in the frog. Its cells are not arranged, as elsewhere in the hemisphere, in the form of primitive ventricular grey, but are scattered uniformly through the substance of the wall. Though no true cortex is formed here, the comparative anatomy of this part makes it plain that the cortex hippocampi of Amniota is differentiated within this region. Accordingly it is properly termed primordium hippocampi. From its rostral end a few medullated fibers accumulate close to the medial surface and descend to the ventro-medial angle, where they turn caudad and accompany the ventral forebrain tract to the hypothalamus (fig. 11). This tract is present also in adult Necturus and here, though the number of medullated fibers is still less than in Amblystoma, they run more separate from the other medullated fibers in the ventral forebrain tract so that the whole course can be read with ease. This is clearly the columna fornicis (see fig. 22). Many unmedullated fibers accompany this medullated tract between the primordium hippocampi and the pars ventromedialis hemisphaerii, but whether any of these extend to the hypothalamus and thus form a part of the columna fornicis is not clear. Similar unmedullated fibers are abundant between the primordium and the nucleus medianus septi for the whole extent of the latter. They are probably for the most part short association fibers. There is also a thin superficial layer of association fibers running around the dorsal angle of the hemisphere between the lateral and medial parts. This extends the whole length of the hemisphere. Accompanying the columna fornicis are many unmedullated fibers and a few medullated fibers between the pars ventro-medialis of the hemisphere and the hypothalamus, the whole complex being the medial forebrain tract.

Extending caudad from the olfactory bulb is the pars ventrolateralis hemisphaerii. This contains, in addition to the tractus olfactorius ventro-lateralis already referred to, medullated and unmedullated ascending and descending fibers between the hemisphere and the thalamus, the lateral forebrain tract.

A cross-section taken behind the olfactorv bulb of adult Am.


424 C. JUDSON HERRICK

blystoma resembles closely a corresponding section of the frog, save for the absence hereof a clearly defined zona limitans lateralis (fig. 12). The pars ventro-medialis is thickened by the enlarged nucleus medianus septi.

As we approach the foramen interventriculare the medial wall below the primordium hippocampi, containing the nucleus medianus septi, becomes thinner (fig. 13) in the site of the larval septum ependymale, and for about 100 micra immediately rostral to the foramen the larval septum ependymale is preserved (fig. 14). No cells of the nucleus medianus extend caudad above the foramen.

The wide larval septum ependymale is almost obliterated by the growth into it from below and from in front of cells of the nucleus medianus septi. In larvae of 35 mm. this movement is in process, as shown by fig. 24, which illustrates a cross-section taken in a plane corresponding to fig. 13 of the adult.

At the level of the interventricular foramen the pars ventromedialis is reduced in size and no clearly defined nucleus medianus septi is here present (fig. 15) in either larvae or adults. The pars dorso-medialis (primordium hippocampi) is in this region large and well differentiated in the adult, but in young larvae it is smaller than at its rostral and caudal ends. Compare fig. 15 (adult) with fig. 4, a section of the 17 mm. larva taken from the same part of the hemisphere. A corresponding section from a larva of 35 mm. is substantially the same as that of the 17 mm. specimen. The ontogeny shows, in fact, that the primordium hippocampi develops wholly within the dorso-median wall of the hemisphere and that its histological differentiation begins at its rostral and caudal ends, the middle part becoming differentiated at a later stage. The morphological significance of this fact will be commented upon in the discussion on page 487. The composition of the fimbria complex is as already described for the 17 mm. larva (p. 418). It receives secondary olfactory fibers from the tractus olfactorius dorso-medianus. The medullated columna fornicis fibers are confined to its rostral end. The unmedullated fibers passing between the primordium hippocampi and the nucleus medianus septi connect with all parts of the primordium in


MORPHOLOGY OF THE FOREBRAIN 425

the frog and probably also in Amblystoma. The fibers of the commissura pallii anterior (com. hippocampi) also reach all parts of the primordium.

Farther caudad in the fimbria complex the commissural fibers predominate. Unmedullated fibers of the commissura pallii anterior come in about equal numbers from the primordium hippocampi in front and from the posterior pole. The commissure passes down behind the interventricular foramen to cross in the dorsal part of the commissural ridge of the lamina terminalis (fig. 16). It is the "dorsal commissure" or the dorsal component of the anterior commissure of the older literature and the " corpus callosum" of Osborn and his followers.

The "dorsal commissure" and the ventral or anterior commissure are not so clearly separate in urodeles as in the frog and comparison with the latter type shows that some of the components of the dorsal commissure of urodeles are dissociated from it in Anura. Morphologically the dorsal commissure, or commissura pallii anterior, should be defined as containing only fibers which connect with the dorsal part (pars pallialis) of the hemisphere, and all fibers related only with the ventral parts of the hemisphere should be classed with the anterior commissure regardless of their topographic position in the commissural complex at the median plane.

In the region where the dorsal commissural fibers descend into the commissure ridge of the lamina terminalis an important component of the fimbria complex continues directly caudad to enter the stria medullatis of the same side. The relations here are very intricate, as will appear beyond. This connection lies dorsally of the interventricular foramen, but ventrally of the sulcus medius, i.e., intheparsventralis thalami, and it occurs in the same relations in Anura also. In Amnio ta the configuration of the parts is so modified, chiefly by the posterior chorioidal fold, as to render such a connection impossible.

At the point where the taenia fornicis joins the taenia thalami the dorsal parts of the hemisphere are cut off from the pars dorsalis thalami and epithalamus by the di-telencephalic fissure and the dorsal plexuses (paraphysis, etc. For the relations of


426 C. JUDSON HERKICK

these plexuses see beyond, p. 431). Here in the eminentia thaiami, which borders the taenia (figs. 16 to 19), many fiber systems are crowded together.

The observed relations are these, the following tracts being all unmedullated except as specified. The fimbria complex, bearing a fewmedullated fibers (fig. 15) connected with the primordium hippocampi, divides into two parts, each with medulla ted fibers, one entering the commissura pallii anterior, the other the stria medullaris (fig. 16). Medullated fibers pass between both of these subdivisions and the adjacent grey of the pars ventrolateralis hemisphaerii (striatum complex). Both subdivisions have unmedullated connections with the posterior pole. Farther back (figs. 18 and 19) the stria medullaris has an unmedullated connection with the eminentia thalami and a strong medullated and unmedullated connection with the rostral end of the pars ventralis thalami. Golgi preparations of adult Necturus show that some at least of the fibers between the stria medullaris and the pars ventro-lateralis hemisphaerii, pars ventralis thalami and eminentia thalami end in these parts by free arborizations. In view of the fact that these are all centers of efferent discharge, I interpret these fibers as conducting downward from the habenula to the somatic motor correlation centers.

There is also a strong connection between the stria medullaris and the preoptic nucleus, which is very large in urodeles. This tract runs chiefly external to the lateral forebrain tract but partly internal to it. A few medullated fibers are found in the latter path (fig. 19). This tractus olfacto-habenularis lateralis et medialis reaches as far forward as the pars ventro-medialis hemisphaerii and possibly as far back as the hypothalamus. Johnston describes in Necturus a tract from the "medial olfactory nucleus" (nucleus medianus septi) to the commissura pallii anterior. I find the tract in Necturus and in the frog in the same relations as figured by Johnston ('06, p. 306, fig. 150). Golgi preparations of adult Necturus show that these fibers arise chiefly from the bodies of cells bordering the recessus superior. These cells lie very close to the ventricle and look like ependyma cells, though probably they should not be classed as such. This tract doubt


MORPHOLOGY OF THE FOREBRAIN 427

less occurs in Amblystoma also, though I have not been able to demonstrate it. Some of its fibers cross in the commissure, but most of them ascend by way of the stria medullaris to the habenula of the same side. It is therefore a component (partly decussating?) of the tractus olfacto-habenularis. I have not been able to find the "tractus lobo-epistriaticus" described by Johnston ('06, p. 309) as passing from the hypothalamus to the primordium hippocampi by way of a decussation in the commissura pallii anterior. These fibers may be the same as the ones which I interpret as tractus cortico-thalamicus, or I may have overlooked them.

On the basis of the relations just described and of the study of an extensive series of brains of Necturus and the frog I interpret the composition of the fimbria and stria medullaris systems of fibers in urodeles as expressed in the following summary (see also fig. 22).

1. The fimbria complex at its rostral end receives fibers of the tractus olfactorius dorso-medialis, is connected by association fibers (ascending and descending) with the nucleus medianus septi and gives rise to the columna fornicis. At its caudal end it contains fibers of the commissura pallii anterior, tractus cortico-habenularis medialis and tractus cortico-thalamicus. The hippocampal commissure is divided into two parts, the commissura pallii anterior and posterior.

2. The commissura pallii anterior includes commissural fibers between the dorso-medial parts of the hemispheres and decussating fibers of the tractus cortico-habenularis medialis and cortico-thalamicus.

3. The stria medullaris is a very complex tract bordering the taenia thalamia. It includes a part of the course of all of the following tracts.

4. The commissura pallii posterior. These are unmedullated fibers which arise from the ventral surface of the posterior pole of the hemisphere and pass directly medial-ward to join the stria medullaris, within which they ascend to the commissura superior and thence pass to the posterior pole of the other hemisphere. They are homologous with the similar tract described by Elliot Smith ('03) in reptiles under the name, commissura aberrans.


428 C. JUDSON HERRICK

5. The tractus olfacto-habenularis system. Primitively, as in cyclostomes, this tract runs in diffuse formation from practically all parts of the secondary olfactory center to converge in the stria medullaris and reach the habenula, part of its fibers first decussating in the superior commissure. In the Amphibia, with the further differentiation of the caudal part of the telencephalon, the relations become very complex, though the same in principle. There are five components (paragraphs 6 to 10 below) of this system, of which the largest comes from the nucleus preopticus and constitutes the tractus olfacto-habenularis in the restricted sense.

6. The tractus olfacto-habenularis lateralis arises chiefly from the anterior part of the preoptic nucleus (see p. 432) and passes upward into the stria medullaris laterally of the lateral forebrain tract (fig. 19).

7. The tractus olfacto-habenularis medialis arises chiefly from the pars magno-cellularis of the preoptic nucleus and ascends internal to the lateral forebrain tract (fig. 18).

8. The tractus septo-habenularis arises in the nucleus medianus septi (it is much larger in Necturus) and passes backward close to the ventricular ependyma to cross the dorsal surface of the anterior commissure and enter the stria medullaris in company with the tractus cortico-habenularis medialis.

9. The tractus cortico-habenularis lateralis is one of the largest components of the stria medullaris. It passes from the posterior pole in company with the commissura pallii posterior (fig. 18).

10. The tractus cortico-habenularis medialis consists of a few fibers which pass from the caudal end of the primordium hippocampi directly into the stria medullaris. Some of its fibers probably cross in the commissura pallii anterior.

All of the components of the tractus olfacto-habenularis just enumerated (numbers 6 to 10) occur in the frog and all but the last in the reptiles.

11. Tractus cortico-thalamicus. This is a sparse collection of medullated fibers accompanying the tractus cortico-habenularis medialis to the stria medullaris; but instead of turning dorsally into the habenula, they continue backward into the thalamus


MORPHOLOGY OF THE FOREBRAIN # 429

and apparently reach the pars ventralis thalami above and behind the optic chiasma. A part of this system decussates in the commissura pallii anterior. I have not been able to trace any of its fibers into the hypothalamus. This tract was recognized in the frog by P. Ram6n y Cajal and named fornix longus. If any fibers of this tract reach the hypothalamus, these would be able to serve as an aberrant columna fornicis, connecting with the caudal part of the primordium hippocampi, just as the typical columna fornicis connects with its rostral end (see fig. 22); but neither my preparations nor P. Ram6n's figure ('96, p. 249) give any evidence of hypothalamic connections.

12. Tractus habenulo-striaticus. Scattered medullated fibers leave the stria medullaris to spread through the grey matter of the caudal end of the pars ventro-lateralis hemisphaerii laterally of the anterior commissure This is the region designated corpus striatum by some recent authors Though not fully homologous with the mammalian striatum, it is one of the sources of that structure; accordingly I term the fibers tractus habenulo-striaticus.

13. Tractus habenulo-thalamicus. Similarly medullated and unmedu Hated fibers in larger number enter the rostral end of the pars ventralis thalami.

14. Tractus thalamo-habenularis. These fibers pass in diffuse formation between the pars dorsalis thalami and the habenula and the most rostral members of the group are for a short distance joined to the stria medullaris.

In the 17 mm. specimen we found the posterior pole of the hemisphere solid and extending but a short distance behind the level of the anterior commissure (fig. 5). At 32 mm. the same is true and the diencephalic flexure still obscures the relations between diencephalon and telencephalon somewhat (figs. 25 and 26). In the adult this flexure has disappeared and the posterior pole has grown far backward laterally of the thalamus (figs. 16 to 20). The primordium hippocampi forms the median wall of the posterior pole and the pars dorso-lateralis the lateral wall, but its caudal end is a relatively undifferentiated tissue which is in contact with the caudal part of the striatum complex below and which has the following characteristic fiber connections: tractus olfac THE JOURNAL OF COMPARATIVE NEUROLOGY AND PSYCHOLOGY, VOL. 20, NO. 5.


430 % C. JUDSON HERRICK

torius dorso-lateralis on the lateral border, thalamic connections by way of the lateral forebrain bundle and striatum complex, commissural fibers in both the commissura pallii anterior and posterior, epithalamic connections by way of the stria medullaris.

The two dorsal parts of the hemisphere, then, terminate in the posterior pole. They do not directly connect with any diencephalic structures and their functional relations with the latter are all effected by tracts which cross the di-telencephalic fissure in the ventral part of the brain tube, i .e., ventrally of the sulcus medius, either by way of the basal forebrain bundles or by way of the stria medullaris. Although the latter receives fibers directly from the caudal end of the primordium hippocampi dorsal and caudal to the interventricular foramen (a condition which does not prevail in mammals), nevertheless these connecting fibers pass through the eminentia thalami (figs. 17 and 18) which corresponds in position to the ventral part of the lateral thalamic wall (see pp. 476, ff) and not directly from dorsal telencephalic to dorsal diencephalic centers. We shall find that the same condition prevails in the frog, though the eminentia thalami is reduced there to a small vestige which is crowded far dorsally. This vestige in the frog is the nucleus supracommissuralis of the literature, which I shall term the nucleus of the commissura hippocampi. For further discussion of this nucleus see p. 440.

The ventro-median part of the hemisphere merges with the nucleus preopticus and this with the hypothalamus, these forming a continuous column of grey, with the median forebrain tract connecting all parts. This tract corresponds with the tractus olfacto-hypothalamicus medialis of fishes and contains both descending and ascending fibers.

The ventro-lateral part of the hemisphere is characterized by the lateral forebrain tract and is continued backward into the prominentia fascicularis (Gaupp) of the thalamus, this prominence being much more evident in the Anura than in urodeles. The coarse fibers of this tract are clearly seen to reach all parts of the ventro-lateral and dorso- lateral parts in the mid-region of the hemisphere; they also reach the posterior pole. A larger proportion of the fibers of this tract decussates in the anterior commissure


MORPHOLOGY OF THE FOREBRAIN 431

in these specimens than in the frog. The lateral forebrain tract is the great conduction path between the lateral parts of the hemispheres and the thalamus.

As is well known, the choroid plexuses of the region are very complex in adult urodeles ; but their development in Amblystoma, Necturus (Warren, '05) and Salamandra (KupfTer, '06, p. 171) shows that the lateral plexuses, median telencephalic plexus (auliplexus, of Kingsbury, '95, fig. 4, and Fish, '95; fig. 3) and paraphysis sensu stricto develop from the telencephalic lamina of the velum transversum in the typical manner, while the diencephalic plexus arises from the caudal lamina of the velum. Fig. 16 is taken at the point of union of the median telencephalic and diencephalic plexuses, i.e., at the dorsal boundary between telencephalon and diencephalon. The eminentia thalami is developed wholly caudal to this level. The di-telencephalic boundary curves downward and backward rostral to this eminence along the line of the sulcus ventralis in figures 17 to 21 to terminate far caudad behind the optic and post-optic commissure ridge. This very anomalous relation grows out of the exaggerated size of the preoptic and supraoptic nucleus in the telencephalon medium of urodeles.

The plan of the diencephalon appears with diagrammatic simplicity in cross sections of the adult brain, where the embryonic diencephalic flexure no longer complicates the matter. The three longitudinal sulci found in the 17 mm. larva (see p. 419) are preserved. The sulcus diencephalicus ventralis passes backward from the interventricular foramen and separates the preoptic nucleus and hypothalamus from the thalamus.

The sulcus medius passes backward from the foramen and separates the thalamus (medithalamus) into dorsal and ventral parts. It becomes less distinct farther caudad and, as in the larva, joins the sulcus limitans. The rostral end of the pars ventralis thaiami is somewhat enlarged and separated from the remainder by a wide groove. This is the eminentia thalami (figs. 17, 18 and 22) which is larger in young larvae than in the adult.

The sulcus diencephalicus dorsalis separates the epithalamus from the thalamus. It is divided into two segments, the rostral,


432 C. JUDSON HERRICK

one of which is the subhabenular sulcus which curves dorsally around the caudal end of the habenula. The other segment begins under the caudal end of the habenula and extends backward through the whole remaining length of the diencephalon. The tissue dorsally of it apparently corresponds with the "posthabenulare Zwischenhirngebiet" described by Goldstein ('05) in teleosts. Its morphology is obscure. The pars dorsalis thalami does not extend as far rostrad as the pars ventralis. Accordingly, the sulcus medius in the rostral part of its course (figs. 18 and 19) separates the pars ventralis (eminentia thalami) from the epithalamus instead of from the pars dorsalis thalami, as farther caudad (see fig. 22). The sulcus diencephalics dorsalis is interrupted by the commissura posterior, behind which the tectum mesencephali for a short distance lies dorsally of the thalamus in the same relations as the post-habenular intermediate region does farther forward (fig. 21). A short distance farther caudad the pars dorsalis thalami also disappears or is merged with the tectum mesencephali and the pars ventralis thalami passes over into the pedunculus cerebri region under the tectum.

The sulcus limitans extends forward from the midbrain into the preoptic recess, its whole diencephalic extent being preserved in the adult Amblystoma (see fig. 22) . In the adults of some other amphibians it seems to be obliterated where it crosses the qaudal end of the pars ventralis thalami. In fig. 22, which is drawn from sagittal sections of adult Amblystoma, there are seen two short sulci running ventrally from the interventricular foramen across the preoptic nucleus. The more rostral one follows the caudal border of the anterior commissura ridge and is the ventral part of Kupffer's sulcus interencephalicus anterior, as we found it in the 17 mm. larva. The other vertical sulcus divides the preoptic nucleus into rostral and caudal portions (figs. 18 and 19), which I shall term the pars anterior and pars magnocellularis of the preoptic nucleus. The latter is homologous with the nucleus magnocellularis of fishes and is intimately related with the corpus striatum complex and the rostral end of the pars ventralis thalami. The tractus olfacto-habenularis medialis arises chiefly from this nucleus, the tractus olfacto-habenularis lateralis from the rostral


MORPHOLOGY OF THE FOREBRAIN 433

part of the preoptic nucleus. I find the same relations in the frog.

The pars ventro-lateralis of the hemisphere contains elements which give rise to the corpus striatum of mammals. From it (and probably also from the pars dorso-lateralis) fibers arise which terminate in the pars ventralis thalami. This component of the lateral forebrain tract is the tractus strio-thalamicus, which partially decussates in the anterior commissure. Its fibers, which are partly medullated, terminate chiefly at the transverse level of the post-optic commissure, though some continue farther caudad. In the caudal part of the pars ventralis thalami they are joined by many other medullated fibers from this part to form the tractus thalamo-bulbaris et spinalis, and by the tractus mammillo-bulbaris from the hypothalamus. These fibers enter the strong ventro-lateral descending tract of the midbrain and oblongata. The tractus tecto-spinalis is added in the midbrain and the whole system evidently corresponds with the descending pedunculus cerebri tracts of mammals. From this it follows that the ventral part of the hemisphere and the ventral part of the thalamus are of common type, both being fundamentally centers of efferent discharge and the correlations directly connected therewith.

The great lemniscus system of the medulla oblongata passes forward into the dorsal part of the midbrain, i.e., above the sulcus limitans, where the larger part of its fibers terminate in the tectum mesencephali (primordial colliculus inferior). This portion of the lemniscus of Necturus is shown in Kingsbury's figures ('95). Some of its fibers, however, continue upward to end farther forward in the tectum and in the pars dorsalis thalami. These are joined by other fibers from the tectum mesencephali to the thalamus. A strong medullated tract associated with the rostral end of the lemniscus passes from the tectum mesencephali through the lateral part of the thalamus to enter the post-optic commissure. Other similar fibers enter this commissure from nearly all portions of the pars dorsalis thalami.

A large collection of medullated and unmedullated fibers passes from the rostral end of the pars dorsalis thalami to the grey nuclei


434 C. JUDSON HEltRICK

in the caudal end of the pars ventralis thalami from which the tractus thalamo-spinalis arises. This is the tractus thalamopeduncularis. An unmedullated tract passes from the rostral end of the pars dorsalis thalami into the hypothalamus, the tractus thalamo-mammillaris. A much larger tract of unmedullated fibers passes from the middle and rostral part of the pars dorsalis thalami to the lateral forebrain tract, and thence to the cerebral hemisphere. This is the precursor of the tractus thalamo-corticalis of mammals and may receive the same name here, though its termination in the hemisphere cannot be given more precisely than to say that it reaches its lateral parts.

The pars dorsalis thalami is evidently a receptive center for somatic (exteroceptive) impulses. These come from the spinal cord and oblongata by way of the lemniscus and from the tectum mesencephali by fibers associated with the lemniscus. Optic impulses are also received directly by collaterals from the tractus opticus ending in the corpus geniculatum laterale. This connection has been frequently described in fishes, where the geniculate body is very small. Johnston mentions it in Acipenser ('01, p. 57) though the lateral geniculate body is not here separable from the nucleus anterior of the pars dorsalis thalami, a condition which prevails to some extent in Amphibia also, where in young larvae the optic connection is confined to the posterior part of the thalamus. In adult Amblystoma the lateral geniculate body covers more than two-thirds of the pars dorsalis thalami and in the frog it has extended forward so as to cover almost its whole lateral surface. I have observed collaterals from the tractus opticus ending in relation with these cells. From this it follows that the pars dorsalis thalami receives optic fibers along with those of the other exteroceptive senses — tactile, somaesthetic and acoustic — received from the lemniscus, and that all of these sensory elements may be represented in the sensory radiations which enter the telencephalon from this part by way of the lateral forebrain tract.

From the preceding description it appears that in Amblystoma the massive lateral wall of the diencephalon on each side is divided longitudinally into four parts which bear simple and obvious rela


MORPHOLOGY OF THE FOREBRAIN 435

tions to corresponding parts of the homolateral cerebral hemisphere. The development shows that these parts in both cases are differentiated from the primitive central grey of the early embryo by a more rapid cellular proliferation than occurs in the intervening sulci; i.e., they are functionally defined. Accordingly, we find them related by important fiber tracts which clearly demonstrate the underlying functional motives of the differentiation.

The hypothalamus is continued forward directly into the pars ventro-medialis hemisphaerii, the whole column being an olfactovisceral center, the rostral end of which is receptive and the caudal end efferent.

The pars ventralis thalami is continued forward directly into the pars ventro-lateralis hemisphaerii, the whole column in both the telencephalon and diencephalon being the primary pathway of efferent discharge into the somatic motor centers.

The pars dorsalis thalami is the receptive center for somatic sensory (exteroceptive) impressions and is functionally related with the lateral wall of the hemisphere, probably (on comparative grounds) chiefly with its dorsal part.

The epithalamus corresponds in position with the pars dorsomedialis hemisphaerii (primordium hippocampi) ; but the morphological problems involved here are not as simple and self-explanatory as in the other cases and can best be discussed on a later page (p. 468, ff.).

Other Urodela

The brain of Desmognathus fusca, as described by Fish ('95) is very similar to that of Amblystoma. The figures show that both primordium hippocampi and nucleus medianus septi are less highly developed and the membranous septum ependymale is extensive. In these respects adult Desmognathus seems to be intermediate between the 17 mm. and the 35 mm. Amblystoma larvae. No medullated fibers were found in the cerebral hemisphere except in the lateral forebrain tract. The three diencephalic sulci are present as I have described them in Amblystoma,


436 C. JUDSON HEKRICK

their positions being indicated in the median dissection of the brain shown in Fish's fig. 3 by projections of the choroid plexus which lie within them.

In Amphiuma (Muraenopsis) the pars ventro-medialis is also feebly developed (Osborn, '83). There is no septum ependymale, the brain being greatly flattened dorso-ventrally. This brings the ventral border of the pars dorso-medialis far ventrad (fig. 23). A small precommissural body is seen around the ventro-medial angle of the lateral ventricle. The dorso-medial and dorso-lateral parts are separated by a deep ependymal groove, while the boundary between the two ventral parts is not so clearly marked.

The 10 mm. larva of Diemyctylus viridescens (Mrs. Gage, '93) in the matters here under discussion is very similar to the 17 mm. larva of Amblystoma, and the adult Diemyctylus to the adult Amblystoma.

In the brain of the adult Plethodon glutinosus the figures of Dodds ('07) show relations of the walls of the hemisphere which .are similar in essential points to those of adult Amblystoma.

In none of the species which I have hitherto described do the cells of the nucleus medianus septi extend dorsally above the interventricular foramen.

Through the kindness of Dr. J. B. Johnston I have examined sections of the brain Cryptobranchus and find the relations very similar to those of Necturus. The nucleus medianus septi is larger than in Amblystoma and extends for a short distance backward dorsally of the interventricular foramen.

I have studied the conditions in adult Necturus and in larvae, the latter from specimens in the Harvard Embryological Collection. Kingsbury ('95) has given an excellent series of figures of the adult brain, and comparison of these with the preceding figures of adult Amblystoma will render a detailed discussion of the conditions in Necturus unnecessary.

In adult Necturus the septum ependymale has become massive except for a very short remnant immediately rostral to the interventricular foramen. The nucleus medianus septi is moderately developed in the ventro-median part of the hemisphere. As we approach the lamina terminalis, it is divided into a ventral and a


MORPHOLOGY OF THE FOREBRAIN 437

dorsal segment by the vestige of the septum ependymale, the dorsal segment being separated from the overlying primordium hippocampi by a deep ependymal groove which represents an incomplete fissura limitans hippocampi (see Kingsbury, '95, figs. 32 and 31). The nucleus medianus septi extends for a very short distance backward above the interventricular foramen, this part being homologous with the pars fimbrialis septi (Kappers, '08) of the frog. In my specimens of Necturus neither the eminentia thalami nor the diencephalic sulci are so distinct as in Amblystoma. The other features of this brain so far as noted agree substantially with those of Amblystoma.

The preparations in the Harvard Embryological Collection show that in the Necturus larva of 29.6 mm., as the nucleus medianus septi invades the septum ependymale, it pushes farther backward in the dorsal border of the septum than in the ventral. This is in contrast with the condition found in Amblystoma, as will be seen by a comparison of fig. 24 with fig. 27. In the latter figure the right side is farther caudad than the left, by reason of slight obliquity of the section. The nucleus medianus fills the whole septum on the left side, but on the right has grown back into its dorsal part only. This is the beginning of the pars fimbrialis septi. Two sections (28/*) farther back the septum is wholly membranous (fig. 28), and the interventricular foramen appears in the next section caudad.

In Kupffer's account ('06) of the larval Necturus 24 mm. long it is evident from the preceding considerations that he has incorrectly identified some of the structures in the median wall of the hemisphere. In his fig. 191 he designates rostrally of the interventricular foramen a sulcus intermedins said to be equivalent to the fissure so named by Gaupp in the frog and to separate the eminentia pallialis medialis (my primordium hippocampi) from the eminentia septalis (my pars ventro-medialis) . This is impossible, for the part here marked eminentia septalis gives rise to the fibers of the hippocampal commissure and conforms in every other respect to the pars dorso-medialis and not to the pars ventromedialis; i.e., it is equivalent to the primordium hippocampi of the frog and not to the eminentia septalis, as Kupffer supposed.


438 C. JUDSON HERRICK

My study of larval and adult Necturus shows clearly that Kupffer's eminentia pallii medialis corresponds with a part of the undifferentiated dorsal wall of the hemisphere and not with the eminentia pa llialis medialis (Gaupp) of the frog; that his sulcus intermedius is a furrow within the pars dorso-medialis and does not correspond with the indentation so named by Gaupp in the frog, the latter being the sulcus limitans hippocampi of my description ; and that his eminentia septalis is Gaupp 's eminentia pallialis medialis and my primordium hippocampi. The ependymal sulcus designated sulcus intermedius by KupfTer in the larva is present in the adult (see Kingsbury, '95, figs. 29, 30, 31) in the same relations, separating the primordium hippocampi from an undifferentiated portion of the pars dorso-medialis. The sulcus limitans hippocampi (sulcus intermedius Gaupp) is seen ventrally of this in Kingsbury's fig. 31.

ANURA

In the half-grown frog tadpole, as in the adult, the olfactory bulbs are fused at their tips, the fusion involving the glomerular and mitral cell layers only. These layers extend to the extreme rostral end of the bulbs and are not confined to their lateral surfaces, as in urodeles and younger frog larvae.

As compared with the urodeles, the ventro-median part of the hemisphere is greatly enlarged and a dense accumulation of cells is found in its dorsal part. This is the precommissural body, or nucleus medianus septi (fig. 29). The region corresponding to the septum ependymale of urodeles is very massive and the lamina terminalis is greatly thickened (fig. 30). The nucleus medianus septi extends dorsal and caudal to the interventricular foramen, though not so extensively as in the adult (fig. 31). This is the pars fimbrialis septi of Kappers ('08).

The walls of the hemisphere are entirely massive rostrally of the interventricular foramen. Caudad, however, the roof is even more widely membranous than in urodele larvae. The foramen is very wide and, as in the adult, there is no plexus lateralis. The membranous roof of the forebrain ventricle is attached to the


MORPHOLOGY OF THE FOREBRAIN 439

massive wall of the hemisphere by the taenia fornicis which is directly continuous caudad with the taenia thalami. From this roof membrane the paraphysis and plexus chorioideus of the third ventricle are developed. The frog larva figured by Kupffer ('06) is apparently older than the specimens here figured and resembles the adult frog more closely.

The four fundamental parts of the wall of the hemisphere of the adult frog are here imperfectly separate, as in the urodeles. The cells of the dorso-median part have the diffuse arrangement characteristic of this region in the adults of both the frog and urodeles. The relations of the commissura pallii anterior to this region (fig. 32, commis. hippocampi) fix its homology as priniordium hippocampi. In the region of the lamina terminalis and farther rostrad the primordium hippocampi is clearly separated from the precommissural body by a cell-free zona limitans (figs. 29 and 30). An ependymal sulcus in fig. 29 marks the position of the fissura limitans hippocampi of the adult, a fissure which is termed fissura arcuata by Gaupp, though incorrectly,, as will appear beyond, for the latter fissure is absent in both larval and adult Amphibia.

The relations at the posterior pole are very similar in these larvae to those of Amblystoma of 30 mm. to 40 mm., described above. The lateral ventricle is extended into it for a short distance caudad of the lamina terminalis and the primordium hippocampi tissue is also extended farther caudad. The commissura pallii anterior is derived chiefly from the primordium hippocampi, but partly from the posterior pole, and at the point where it turns downward from the latter to pass behind the interventricular foramen into the lamina terminalis it comes in contact with the stria medullaris, as in urodeles. At this point of contact there is developed a small dense nucleus, which here bears the same relation to the taenia fornicis as does the caudal end of the pars fimbrialis septi farther rostrad (cf. figs. 31 and 32). These nuclei are, however, not continuous and are of quite different origins. This nucleus is termed by Gaupp in the adult frog nucleus supracommissuralis. I term it for reasons which will appear beyond the nucleus of the commissura hippocampi. In the tadpole it


440 C. JUDSON HERRICK

is relatively larger than in the adult and increases rapidly in size farther caudad. It lies ventrally of the sulcus diencephalics medius, is the equivalent of the eminentia thalami of urodeles and is directly continuous, as recognized by Gaupp ('99, p. 82), with the pars ventralis thalami (pars medialis, Gaupp). Its fiber connections I have not determined with certainty in the frog, but its uniform position at the union of the fimbria complex and the stria medullaris suggests that it is related to both of these tracts (see especially fig. 33), as indeed I find to be the case in reptiles. In the urodeles numerous fibers pass from the equivalent structure (eminentia thalami) into the stria medullaris (fig. 18). The fibers of the tractus cortico-habenularis medialis pass directly through it and some of these fibers may end here. The same is true of fibers of the tractus olfacto-habenularis from the preoptic nucleus.

Some of my preparations of adult frog brains by the silver reduction method of Ram6n y Cajal show very delicate fibers passing between this nucleus and the commissura pallii anterior. These appear to be dendrites of the cells of the nucleus and collaterals of the commissural fibers. Golgi preparations of adult Necturus show clearly short collaterals from the commissural fibers coming into relation with the cells of this nucleus. This indicates that the nucleus constitutes, in part at least, a station for nervous impulses passing between the primordium hippocampi and the habenula. The fibers from the commissura pallii anterior (com. hippocampi) thus constitute a tractus cortico-habenularis medialis cruciatus, with a synapse interpolated in this path at the supracommissural nucleus. The crossed and uncrossed path together, in this case, are strictly comparable with the tractus cortico-habenularis of reptiles and mammals save that in Amphibia they pass (like the commissura hippocampi itself) dorsad and caudad of the interventricular foramen to reach the stria medullaris, whereas in Amniota the path goes rostrad and ventrad of the foramen. We shall find the relations of these tracts in reptiles very instructive in this connection. (See p. 461.)

The relations of the fimbria complex, including the commissura hippocampi, to the stria medullaris at this point areessen


MORPHOLOGY OF THE FOREBRAIN 441

tially as described for Amblystoma (p. 424 ff.). Associated with the stria medullaris, are commissural fibers between the posterior poles of the hemispheres running by way of the commissura superior. This is the commissura pallii posterior.

On account of the absence of the diencephalic flexure in these brains the configuration of the thalamic structures bordering the telencephalon is very different from that of larval urodeles, but is the same in principle. The median and lateral ventral parts of the hemisphere pass back into the nucleus preopticus and prominentia fascicularis as before. The whole of the large nucleus preopticus, which belongs to the unevaginated forebrain or telencephalon medium, is overlapped dorsally by diencephalic structures, the sulcus diencephalicus ventralis forming the boundary between them (see figs. 33 to 38). As in the adult frog, this nucleus is very large and shows considerable internal differentiation. Correlated with the straightening out of the diencephalic flexure and the further evagination of the cerebral hemispheres in the anuran larvae we find, as mentioned above, that the eminentia thalami at the caudal border of the interventricular foramen is reduced to a small vestige, the nucleus of the hippocampal commissure (figs. 32 and 33).

Immediately caudal to the foramen the pars ventralis thalami enlarges to assume the same relations as in urodeles, being bounded by the sulcus medius and sulcus ventralis.

The ventro-lateral part is extended for a considerable distance behind the posterior pole of the hemisphere into the unevaginated telencephalon medium dorso-laterally of the preoptic nucleus (figs. 33, 34, 35). This tissue, including a part of the primordial corpus striatum, is in the course of further development almost all evaginated into the hemisphere proper, as shown by the adult configuration.

The habenula and superior commissure are placed far forward, as in urodeles, and the pars dorsalis thalami lies almost wholly caudal to these structures. The two segments of the sulcus diencephalicus dorsalis separate the epithalamus from the thalamus as described on p. 431 for urodeles.

The corpus geniculatum laterale appears near the rostral end


442 C. JUDSON HERRICK

of the thalamus and accompanies the lateral border of the nucleus dorsalis thalami for its entire length. It also overlaps to some extent the pars ventralis thalami (figs. 36 to 39).

Gaupp in 1899 summarized the most important data upon the structure of the brain of the adult frog in the literature up to that date, together with a wealth of new observations, and his excellent account should be the point of departure for all subsequent work. I have examined a very large number of brains of different species of frogs prepared by diverse methods, including those of Weigert, Nissl, Golgi and the silver reduction method of Ram6ny Cajal. The details of my observations must be reserved for later publication and we shall here consider only such general features as have a direct bearing upon the problems of the subdivision of the hemisphere and the morphological relations of these parts. The necessary figures will be found in the works of Gaupp ('99), Kupffer ('06), Kappers ('08), B. Haller ('08), Snessareff ('08) and the brothers Ramon y Cajal. See also fig. 40.

In the adult frog, as already stated, the fundamental parts of the hemisphere have attained the typical amphibian form. The shape of the hemisphere has considerably changed, as compared with urodeles and larval anurans, especially by the further evagination of structures found in earlier stages in the telencephalon medium, by the consequent contraction of the interventricular foramina and by the relative increase in the dorso-ventral diameter of the hemispheres. The lateral ventricles are dilated vertically and contracted laterally, thus giving to the cross-section of the hemisphere dorsal and ventral angles which form sharp boundaries between the lateral and medial parts. We shall next review the form and distinguishing characters of the five parts of the cerebral hemisphere of the frog.

The peculiar fusion of the olfactory bulbs seems to be confined to the bulbs themselves, as in the tadpole, though the participation of a small amount of secondary olfactory tissue (lobus olfactorius anterior) cannot certainly be excluded. Some fibers in diffuse formation cross the median plane in this region.

The pars ventro-medialis hemisphaerii (epistriatum, P. Ramon y Cajal; eminentia septalis, Gaupp; pars septalis hemisphaerii,


MORPHOLOGY OF THE FOREBRAIN 443

Kupffer) is included between the ventral angle of the hemisphere and the fissura limitans hippocampi. It is bounded rostrad by the olfactory bulb and caudad by the lamina terminalis. Within this part there is a highly differentiated cellular mass, the corpus precommissurale or nucleus medialis septi, which extends backward dorsally into the lamina terminalis and both below and above the interventricular foramen. These cells form the "bed" of the anterior and dorsal (or anterior pallial) commissures. The portion of the nucleus medianus septi which extends caudad above the foramen is very much enlarged in some species of frogs. It is well named by Kappers ('08) the pars fimbrialis septi, for its cells are in intimate relation with the fimbria fibers passing between the dorso-median and dorso-lateral parts and the ventromedian part which form one component of the complex termed tractus olfactorius septi.

B. Haller ('08, p. 373) designates this nucleus "Ammonskern" and considers it a part of the "Ammonswulst" or primordial hippocampus. He cites Edinger as teaching that the Ammonskern alone is the precursor of the hippocampus, but in the last edition of Edinger's Vorlesungen the pars dorso-medialis is designated cortex ('08, p. 306, fig. 276). The ventro-median part receives fibers from the olfactory bulb in front and is broadly connected with the hypothalamus through the median forebrain bundle, both by ascending and by descending tracts which partially decussate in the anterior commissure. It also has broad fibrous connections with the other parts of the hemisphere.

The pars ventro-lateralis is included between the ventral angle of the hemisphere and the fissura endorhinalis (sulcus limitans lateralis, Gaupp). It is bounded rostrad by the olfactory bulb ; the caudal boundary is fixed dorsally by the so-called corpus striatum laterally of the lamina terminalis, while ventrally it passes over without interruption into the prominentia fascicularis (Gaupp) of the thalamus, which is a part of my pars ventralis thalami. In front it receives fibers from the olfactory bulb which for the most part come from the bulbulus accessorius and join to form a compact fascicle of unmedullated fibers which passes backward close to the ventricle and immediately ventral to the


444 C. JUDSON HERRICK

zona limitans lateralis. These fibers have been variously interpreted by different authors. I have Golgi impregnations which show without ambiguity the whole course of this tract. Its fibers reach the so-called corpus striatum where they break up into a dense neuropil among the dendrites of the "striatum" cells. Fibers emerge from this neuropil and decussate in a separate (unmedullated) slip of the anterior commissure which connects the "striata." Others connect with the rostral end of the hypothalamus, where they decussate wholly or partially in the postoptic commissure system. The latter are probably both ascending and descending fibers, and the ascending fibers correspond to the tractus pallii of selachians, as has been pointed out to me by Professor Johnston. There are very numerous unmedullated fibers which leave the neuropil in the "striatum" to arborize freely in the adjacent lateral wall of the hemisphere and posterior pole. These come chiefly from the hypothalamic tract. Medullated fibers leave the nucleus for the pars ventralis thalami, some of which decussate in the anterior commissure. These form part of the tractus strio-thalamicus. Finally, there is the medulla ted and unmedullated connection with the stria medullaris which I interpret as tractus habenulo-striaticus (see p. 429). These relations are shown diagrammatically in fig. 41.

There is no differentiated corpus striatum in the Amphibia in the sense in which this term is used in mammals. The elements from which it is to be differentiated are present in the pars ventrolateralis.

The lateral forebrain tract is related to this part as in urodeles, its ascending fibers including strong tracts from the colliculus inferior and from the pars dorsalis thalami and corpus geniculatum laterale. These fibers terminate in both the ventral and dorsal lateral parts of the hemisphere and partially decussate in the anterior commissure.

The pars dorso-lateralis (formatio pallialis lateralis, Gaupp) lies between the fissura endo-rhinalis and the dorsal angle of the hemisphere. It extends from the olfactory bulb to the posterior pole, of which it forms the lateral wall. Tt receives a strong tract from the olfactory bulb and is broadly connected by very com


MORPHOLOGY OF THE FOREBKATN 445

plex tracts with the other three parts of the hemisphere. Its connections show clearly that it corresponds, at least in pare, with the nucleus sphaericus of the reptiles and the lobus pyriformis of the lower mammals.

The pars dorso-medialis (eminentia pallialis medialis, Gaupp ; septum, P. Ramon y Cajal; primordium hippocampi, Kappers). In the frog the whole of this part becomes primordium hippocampi but in some of the urodeles it seems probable that only its medial portion should be so designated. It extends the entire length of the hemisphere. In front of the lamina terminalis it is bounded ventrally by the fissura limitans hippocampi; behind this level it forms the whole median wall of the hemisphere. For its entire length it forms the thickest part of the wall.

At its rostral end the primordium hippocampi is enlarged and here it receives secondary olfactory fibers from the olfactory bulb, and is connected by a great mass of unmedullated ascending and descending fibers with the underlying nuclei of the septum. All portions of the primordium are connected by strong association tracts with the dorso-lateral part of the hemisphere. The fibers of the commissura pallii anterior arise throughout its mass and P. Ramon y Cajal ('05, fig. 4) figures a Golgi impregnation of this commissure which shows that the free terminal arborizations of its fibers after decussation also reach the lateral wall of the hemisphere directly.

The cells of the primordium hippocampi are arranged diffusely, being scarcely more dense next to the ventricle than elsewhere; but there is no true cortex developed except perhaps atthedorsomedian angle of the hemisphere. P. Ramon y Cajal ('05, p. 185 and fig. 7) describes here a superficial layer of tangential cells which have well differentiated axons leaving the cell body or the base of one of the thick dendrites and passing ventrally along the medial superficies of the hemisphere. I have impregnations of these cells in Golgi preparations of Rana pipiens which show thac in the rostral half of the hemisphere they extend from the dorsal angle ventrally along the entire medial surface of the primordium hippocampi instead of being limited to the dorsal angle as figured by Ramon y Cajal. Medullated fibers which I consider to belong

THE JOURNAL OF COMPARATIVE NEUROLOGY AND PSYCHOLOGY, VOL. 20, NO. 5.


446 C. JUDSON HERRICK

to the columna fornicis (see p. 423) pass downward into the septum from amongst these cells, but I have not been able to establish their connection with them. If this connection exists, then these cells constitute a true cortex hippocampi. The superficial layer of tangential cells can be easily distinguished from the deeper cells in sections stained with toluidin blue on account of their difference in form, which is shown diagrammatically in fig. 40. There is, however, no cell-free medullary or molecular laj-er between these cells and the so-called pyramidal cells which fill the remainder of the primordium.

In addition to the true columna fornicis fibers which pass down through the septum and lamina terminalis in the typical vertebrate way, as described above, there are other connections which are functionally of the same type. Unmedullated fibers from all parts of the primordium hippocampi and from the dorso-lateral part (the latter curving around the dorsal and mesial surface of the hemisphere) pass downward into the nucelus medianus septi. Some of these form one component of the complex tract known as the tractus olfactorius septi and may pass through this nucleus to enter the hypothalamus by the way of median forebrain tract, in which case they should be associated with the medullated tract already described as columna fornicis. Others certainly end within the nucleus medianus. Since the latter sends its fibers to the hypothalamus, this path with its interpolated synapse may also serve functionally the same purpose as the columna fornicis. No fibers pass from the true columna fornicis into the stria medullaris, as in reptiles and mammals, to form the tractus cortico-habenularis ; but there are other connections which functionally belong to the fornix system.

Some of the characteristic relations of the rostral part of the hemisphere are shown in the diagrammatic cross section, fig. 40.

In no Amphibia do the fibers of the commissura hippocampi system take the course characteristic of the reptiles and mammals, viz., above the interventricular foramen, to cross in the lamina terminalis in front of the foramen; on the other hand, they cross behind the foramen by two paths. One of these is ventral in the


MORPHOLOGY OF THE FOREBRATN 447

lamina terminalis, the other is dorsal in the commissura superior. The configuration is such that the primordium hippocampi touches the anterior border of the thalamus and, as in the larva, these two commissural tracts converge at the junction of the taenia fornicis with the taenia thalami.

The relations here are similar to those of urodeles and frog tadpoles (see pp. 426 and 439), except for the great increase in size of the supraforaminal part of the precommissural body, or pars fimbrialis septi, and the reduction of the eminentia thalami to form the small but dense nucleus of the hippocampal commissure, which lies at the junction of the fimbria fiber complex with the stria medullaris ventrally of the pars fimbrialis septi. The summary of fiber connections at this point given for urodeles on p. 427 applies also to the frog.

The posterior pole of the hemisphere is that portion of the two dorsal parts which projects caudad beyond the level of the lamina terminalis; it is the parathalamic brain. Here the dorso-lateral and dorso-medial parts come together as in urodeles and in some measure lose their distinctive characteristics, suggesting the undifferentiated condition of the so-called lobus hippocampi or nucleus pyriformis (Sheldon) of teleosts (see p. 451). It receives numerous secondaiy olfactory fibers from the olfactory bulb and gives rise to the tractus cortico-habenularis lateralis (tractus taeniae, Edinger) and the commissura pallii posterior, as has been already suggested by Elliot Smith ('03, p. 495).

The hemispheres of the adult frog are so much farther evaginated, as compared with the larvae and the urodeles, as to make correlation with reptiles and higher forms relatively easy. The dorsal parts are clearly separated from the ventral parts for the entire length of the hemisphere both by the external and internal zona limitans and by total fissures; and the ventral parts alone are in direct massive contact with the corresponding parts of the diencephalon. The dorso-median part is in contact with the ventral part of the thalamus, but not with any part of the thalamus dorsally of the sulcus diencephalicus medius. This contact is probably secondarily acquired in Amphibia (see p. 420) and it does not occur in Amniota.


448 C. JUDSON HERRICK

The relations of the ventro-median part remain practically unchanged except for further growth beyond the lamina terminalis.

The ventro-lateral part is as directly continuous with the pars ventralis (prominentia fascicularis) of the thalamus as in the larva. The elements of the primordial corpus striatum are more completely evaginated than in the larva; nevertheless the caudal end of this complex still remains behind the interventricular foramen in intimate relation with the lateral forebrain tract in the unevaginated wall in the same morphological level as the eminentia thalami of the larva.

The pars dorsalis thalami is greatly enlarged in the frog and its whole lateral surface is differentiated as corpus geniculatum laterale to receive collaterals and terminals of the optic tract. The fiber connections of the thalamus are essentially as already described for the urodeles (pp. 433, ff.), save for a much larger proportion of optic projection fibers from the corpus geniculatum laterale in the lateral forebrain tract, and also a larger component from the midbrain.

Gaupp ('99, p. 80) gives a very clear analysis of the walls of the diencephalon. He divides the central grey of the thalamus into habenular ganglion and superior, medial and inferior parts. The last belongs in the hypothalamus and the superior and medial parts correspond with the pars dorsalis and pars ventralis thalami of my description. The sulcus diencephalicus dorsalis is present and also the sulcus medius. Gaupp does not name these sulci, but gives to the cell-free zone bordering the sulcus medius the name zona limitans superior. The sulcus diencephalicus inferior is absent, but its position is marked by a similar " zona limitans inferior." Gaupp correctly describes the pars dorsalis (superior) thalami as ending in attenuated form rostrally of the habenula, and the pars ventralis (his medialis) as extending forward into the nucleus supracommissuralis, or nucleus of the hippocampal commissure.


MORPHOLOGY OF THE FOREBRAIN 449

COMPARISONS WITH FISHES

The brief account of the brain of Lepidosiren recently published by Elliot Smith ('08) shows striking resemblances with the Amphibia, with, however, divergent lines of specialization. The differentiation of a layer of cortical cells in the dorsal region is reptilian in type, though these cells are probably represented in the frog also (see fig. 40), and the ventral enlargement of the tuberculum olfactorium is unique. The similarity to the amphibian conditions is, in fact, so close as to justify us in reviewing Elliot Smith's conclusions in the light of the preceding discussion.

He describes the corpus paraterminale as comprising practically the whole median wall dorsally of the tuberculum olfactorium and further states ('08, p. 533) that it is divided into dorsal and ventral parts by an " indentation of epithelial cells" which the figures show to mark the level of the interventricular foramen. The ventral part is clearly the nucleus medianus septi. It is equally clear that the dorsal part is chiefly composed of the equivalent of the amphibian pars dorso-medialis, or primordium hippocampi. In the adult frog Elliot Smith has recognized ('03, p. 497) the distinction between the primordium hippocampi and that portion of the paraterminal body which lies dorsally of the interventricular foramen (pars fimbrialis septi of Kappers) ; but he has failed to make the corresponding analysis here. An examination of the photographs reproduced in his paper suggests that the resemblance to the Amphibia is even closer than Elliot Smith seems to have recognized.

The arcuate fissure is absent in Lepidosiren, as in the Amphibia. The fissure B of Elliot Smith's fig. 1 would appear at first to be the fissura limitans hippocampi, which marks the position of the zona limitans in the frog; but this I think is not the case. It is more nearly comparable with the ependymal groove which is seen in the frog larva (fig. 30), and marks the level of the dorsal boundary of the interventricular foramen. In both Lepidosiren and the frog tadpole the precommissural body (nucleus medianus septi) extends dorsally of this groove, while the fissura limitans hippocampi always lies wholly dorsal to the precommissural body. The


450 C. JUDSON HERRICK

photographs (figs. 2 and 14 of Elliot Smith's paper) show clearlyabove this fissure B a denser mass of cells separated from the overlying pars dorso-medialis by a zona limitans. These cells mark the most dorsal limit of the nucleus medianus septi or precommissural body, as I have defined it in this paper, and seem to be exactly comparable with the larger mass of cells seen in larval and adult frogs in the corresponding position. They may extend caudad for a short distance above the interventricular foramen, though the figures suggest that this is improbable.

Elliot Smith explains briefly the morphology of the rudimentary pallial formation of Lepidosiren and I shall return to the consideration of his valuable suggestions in the final discussion (p. 490). None of the available descriptions of the brain of Ceratodus are well adapted for the comparisons here instituted. Bing and Burckhardt ('05, fig. 15, p. 550) give a figure of the embryo which shows that at this stage the cross section of the forebrain at the level of the interventricular foramen is almost identical with that of the just hatched Necturus (cf. Kingsbury, '95, pi. ix, fig. 8). Protopterus, as figured by Burckhardt ('92) is much more sim ilar to Lepidosiren than is Ceratodus.

The fiber tracts of the Dipnoi are not sufficiently well known to permit us to control all of the homologies here suggested; but so far as known they support them. Burckhardt 's meager ac count of the forebrain tracts of Protopterus shows that the commissura pallii anterior is present (termed corpus callosum), but it is not clear whether it enters the lamina terminalis behind the interventricular foramina as in Amphibia or in front of them as in elasmobranchs. His figures of cross-sections show that the diencephalon and telencephalon are divided longitudinally into parts which correspond closely with those of Lepidosiren and Amphibia. He discusses (p. 25) these longitudinal zones in comparison with those of urodeles, but since he incorrectly homologized all of the parts of the hemisphere (see Elliot Smith, '08), his analysis is not fruitful.

The cerebral hemispheres of the higher ganoids and toleosts have developed in a different direction from that taken by the Amphibia. In the olfactory bulb alone is the massive wall of


MORPHOLOGY OF THE FOREBRAIN 451

the telencephalon fully evaginated. The telencephalon medium is elongated and in its massive floor and side walls are differentiated structures corresponding with the four parts into which we have divided the amphibian hemispheres, the membranous roof of the ventriculus impar being the equivalent of the median and lateral plexuses of the urodele hemispheres. In the higher fishes the enlargement of the massive parts has led to their lateral eversion, not to an evagination of the whole wall as in Amphibia. The subdivision of the teleostean "hemisphere," therefore, resembles more closely the plan of the diencephalon than that of the telencephalon of Amphibia (see p. 477 and fig. 83), but is in iirportant respects unlike both of them.

This fundamental difference between the teleostean and amphibian types of cerebral hemispheres was first pointed out by Mrs. S. P. Gage ('93) and the idea has since been elaborated by Kappers and Edinger. An analysis of the teleostean forebrain will shortly be published by Dr. R. E. Sheldon, in connection with which this problem is fully discussed. Accordingly, no further reference will be made to it here save to add that, though very different morphogenic factors have operated in the evolution of the teleostean brain as compared with elasmobranchs and amphibians, nevertheless the homologous parts are recognizable. Our conclusions follow, with some important modification in details, the interpretation of Johnston ('10, p. 155).

The selachians have been recently re-examined by Sterzi and from his descriptions it is evident that in these brains the telencephalon medium is extensive. In the lower elasmobranchs the evaginated part contains little except the olfactory bulbs, and even in the higher sharks the remaining parts of the telencephalon are differentiated for the most part as local thickenings of the wall, with a relatively small amount of evagination as compared with the Dipnoi and Amphibia. In the diencephalon of Acanthias one of Sterzi's figures ('09, p. 577, fig. 232) shows the sulci arranged in much the same way as in urodeles. See also Burckhardt's figure of Scymnus ('07, p. 354, fig. 21). These probably separate homologous parts in amphibians and elasmobranchs, but the determination of this point must await a fuller study of their i


4")L > C. JUDSON HERRICK

fibrous connections. Further consideration of these brains will be reserved pending the appearance of a paper on the subject now in press by Professor Johnston.

Some considerations regarding the brains of cyclostomes will be found in the discussion on page 470.

REPTILIA

The reptilian cerebral hemisphere is characterized chiefly by the highly developed corpus striatum complex and by the presence of true cortex in the pallium. This cortex is arranged in three distinct laminae (cf. fig. 61), the lateral cortex (paleocortex of Kappers, 'OS), the dorsal cortex (cortex ammonis, Kappers) and the dorso-medial cortex (fascia dentata, Kappers). The dorsal and dorso-medial cortex both extend to the extreme rostral end of the hemisphere, the latter here resting directly upon the underlying secondary olfactory center of the olfactory crus. This condition is repeated exactly in the lower mammals in relations which permit us to homologize the dorso-medial cortex of the lizard without question as cortex hippocampi. Comparison with the adult frog shows that there too the dorso-medial part of the hemisphere extends forward in the same way to the olfactory bulb over the pars ventro-medialis, a fact which justifies us in calling the pars dorso-medialis of the frog primordium hippocampi for its entire length.

The reptilian septum is further differentiated as compared with the Amphibia, and is in fact similar to that of lower mammals. The unitary collection of cells in its dorsal part which in Amphibia was termed nucleus medianus septi, or precommissural body, is here subdivided into several nuclei. The term precommissural body may be retained for all of these nuclei related to the amphibian body of the same name, including the reptilian bed nuclei for the commissures in the lamina terminalis (and the nucleus of the commissura pallii posterior of lizards), though it must be borne in mind that the term is used here in a more restricted sense than by Elliot Smith. That author's pre-commissural or paratenuinal body of reptiles includes also the vestigial primor


MORPHOLOGY OF THE FOREBRAIN 453

dium hippocampi, which, as will appear immediately, I exclude from it. The nuclei of the hippocampal commissure and tractus cortico-habenularis (see p. 461) should also be excluded.

The most highly differentiated cortex of reptiles is the dorsomedial. Its relations to the ventro-medial part of the hemisphere have been investigated in adult brains of several species of reptiles and in the extensive series of embryos in the Harvard Embryological Collection.

Chelonia. We shall consider first the turtles. In transverse sections of an embryo of Chrysemys marginata of 16.7 mm. (Harvard Embryological Collection, no. 1092) the ventro-medial part, or septum, has already attained great size; and, although the process of proliferation of neuroblasts from the central grey is still active, the principal regions of the adult hemisphere can be recognized. The dorso-medial cortex is well defined; the lateral cortex less clearly. A section through the middle of the septum (fig. 43) shows that the medial wall of the hemisphere is subdivided, as in Amphibia, by a definite zona limitans into dorsal (hippocampal) and ventral (septal) parts; but the lateral wall does not conform to the amphibian type. The tuberculum olfactorium, precommissural body (nucleus medianus septi) and nucleus accumbens septi can be recognized. The locus of the nucleus lateralis septi lies between the two last. Fibers of the fornix and commissura hippocampi systems are seen crossing the zona limitans. 4 section taken 150 micra farther caudad (fig. 44) shows dorsal to the zona limitans a thin portion of the wall which contains fimbria fibers and a few nuclei. In a section taken 50 micra farther back and immediately in front of the interventricular foramen (fig. 45) an extensive septum ependymale appears and above it a slight thickening containing the fimbria fibers, the limbus medullaris of His. This belongs to the dorsomedian part of the hemisphere and represents a small residue of the unspecialized amphibian primordium hippocampi, the remainder of the primordium having been transformed into cortex hippocampi (dorso-medial cortex). The commissura hippocampi crosses in tne lamina terminalis in the plane of this section, its fibers having curved downward rostral to the interventricular foramina and septum ependymale.


454 C. JUDSON HERRICK

In Chrysemys marginata of 27 mm., cross sections show that the medial and lateral nuclei of the septum are well differentiated and there is a membranous sac or pocket which projects forward on each side above the lamina terminalis and its commissures (fig. 46). This is the recessus superior which Elliot Smith ('03) has recognized in monotremes and other vertebrates and which Mrs. Gage has recently demonstrated in a five weeks human embryo (at the Boston meeting of the Association of American Anatomists, Dec. 28, 1909). The paraphyseal tubules spring from its dorsal surface and the lateral plexuses arise immediately behind it. The recessus superior has been described in the gecko embryo by Tandler and Cantor ('07).

I have preparations of adult brains of Chrysemys marginata by the Weigert method which show that here also there is an undifferentiated region between the precommissural body and the cortex hippocampi which contains fimbria fibers and scattered small cells. This is a vestige of the unspecialized primordium hippocampi which has not assumed cortical characters. In the middle region of the septum the nucleus lateralis septi has grown up dorsally into contact with the cortex hippocampi on its ventricular side very much as shown for the alligator in fig. 66; but both rostral and caudal to this level the primordium hippocampi occupies the whole thickness of the wall, where it is bounded below by the fissura limitans hippocampi. It is everywhere separated from the cortex hippocampi by the fissura arcuata.

In the adult box tortoise, Cistudo Carolina, the same conditions prevail save that the lateral nucleus of the septum is smaller here. Fig. 47 illustrates the relations immediately rostral to the interventricular foramen in Cistudo.

In none of the turtles which I have examined, either embryonic or adult, do any cells of the precommissural body extend backward above the interventricular foramina, as in the frog and lizard.

The relations of the telencephalon to the diencephalon are clearly illustrated by figs. 48 and 49, where the resemblance to urodele larvae is very striking. These sections of a 12.3 mm. embryo of Chrysemys marginata are cut transversely to the sagittal plane and so inclined that the dorsal surface is much farther


MORPHOLOGY OF THE FOREBRAIN 455

caudad than the ventral. Fig. 48 passes through the caudal border of the interventricular foramen on the left side and immediately behind it on the right. It passes through the anterior commissure below and the dorsal sac rostral to the habenulae above; cf. fig. 50, a sagittal section of an older embryo. From the interventricular foramen, which is very wide immediately rostral and dorsal to this level, two ependymal grooves extend backward in the same relations as in urodcles, the sulcus diencephalicus medius above and the sulcus diencephalicus ventral is below. Between these is the pars ventralis thalami which is seen to connect broadly with the lateral wall of the hemisphere. In fact it is the precise equivalent of the eminentia thalami of amphibian larvae. It is completely separated from the dorsomedian part of the hemisphere by the membranous posterior chorioidal fold and incompletely separated from the ventromedian part by the sulcus ventralis. In the adult its rostral end is carried forward, as in the frog, to form the nuclei of the hippocampal commissure and tractus cortico-habenularis (see p. 461).

Dorsal to the sulcus medius is the pars dorsalis thalami, and above this the sulcus dorsalis and habenula. These relations come out still more clearly in a section taken a little farther back (fig. 49).

Fig. 50 is a sagittal section through the brain of an older embryo of Chrysemys marginata 26.7 mm. long and fig. 51 a parasagittal section from the same embryo illustrating the relations of the diencephalic sulci. The pars dorsalis thalami is here, as in reptiles and mammals generally, divided into a central major part and a nucleus dorsalis containing more densely crowded cells. Both embryological and phylogenetic development show that these two parts have a common origin.

The subdivisions of the diencephalon are shown with great clearness in the models of the brains of gecko embryos figured by Tandler and Cantor ('07, see especially figs. 10, 14, and 17). The boundary between the pars ventralis thalami and the hypothalamus in these embryos is less distinct than in the Amphibia and the chelonian embryos here described. Tandler and Cantor's


456 C. JUDSON HERRICK

scheme of subdivision of the reptilian prosecephalon seems to me faulty, as will appear beyond. This is probably in part due to the fact that they did not study sufficiently early stages but chiefly because of imperfect knowledge of the functional factors involved in development such as would be brought out only by a study of the adult fiber tracts.

Lacertilia. — In the lizards the precommissural body is much more highly developed than in the turtles and this leads to important differences in the development of the septal region. In the adult the nuclei of the septum extend much farther dorsad and caudad than in turtles in relations similar to those of the pars fimbrialis septi of the frog.

The Harvard series 1601 comprises frontal sections of an embryo of Lacerta of 7.6 mm. (measured as coiled, total length unknown) cut in approximately the same plane as the older embryo shown in figs 53 to 57, though somewhat more nearly transverse to the long axis of the diencephalon. At this age the commissura pallii anterior is present, though small, but the commissura pallii posterior has not yet appeared. Tandler and Cantor ('07) find that in the gecko also the commissura pallii anterior develops earlier than the commissura pallii posterior.

A section through the mid-region of the septum, corresponding approximately to the plane of fig. 44 of Chrysemys, shows that the nucleus lateralis septi has grown dorsad so as to invade the ependymal border of the primordium hippocampi (fig. 42), a condition which is found in this region of adult Chrysemys but not in the embryo of the age shown in fig. 44. This position of the lateral nucleus is characteristic of the entire length of the septum of Lacerta at this and all later ages. The ventral limit of the original primordium hippocampi can, however, generally be recognized on the outer surface of the hemisphere of adult lizards. The region corresponding with the membranous septum ependymale of of the Chrysemys embryo of 16.7 mm. (fig. 45) is massive in the Lacerta embryo of 7.6 mm., containing fimbria fibers and cells of the medial and lateral nuclei of the septum.

Passing caudad in this series of sections of Lacerta, at the level of the commissura pallii anterior a portion of the nucleus median us


MORPHOLOGY OF THE FOREBRAIN 457

septi extends dorsally of the commissure, giving relations very similar to those of the older embryo of fig. 54. The nucleus lateralis septi also extends for a very short distance dorsally of the interventricular foramen, but not so far as in the older embryo, figs. 55 and 56. No cells from the septum extend farther backward than the dorsal border of the foramen, the nucleus of the commissura pallii posterior being as yet undeveloped, in correlation with the absence of the commissure itself.

Figs. 52 to 57 illustrate sections through the brains of older embryos of Lacerta (about 36 mm. long when uncoiled, head 5 mm. long) from the Harvard Collection. Fig. 52 is a parasagittal section through the brain of one of these specimens and upon it is indicated the approximate plane of section of each of five sections through another specimen. In this figure the precommissural body is seen to rise up in front of the interventricular foramen and extend backward for a short distance above it. The nucleus lateralis septi, separated from the nucleus medialis by fimbria fibers, extends far dorsal and caudad into contact with the margin of the cortex. The thalamus is related to the telencephalon as in the turtles already described.

Figs. 53 to 57 illustrate a series of sections taken in the planes indicated on fig. 52. The primordium hippocampi is clearly defined on the medial surface of the brain, being bounded by definite sulci at the level of fig. 54. The fissura limitans hippocampi separates the precommissural body from the primordium hippocampi, and the latter is separated from the dorso-medial cortex (cortex hippocampi) by the fissura arcuata. Attention is especially called to the fact that the fissura arcuata lies within the hippocampal formation, not below or above it, and to the importance of this fact in determining its homology in vertebrates.

The fibers of the commissura hippocampi are here divided, one part crossing in the lamina terminalis as in turtles, here termed the commissura pallii anterior (figs. 53 and 54), and another part derived from the posterior pole crossing in the velum transversum at the rostral border of the epithalamus (figs. 56 and 57).

A special collection of cells belonging to the supra-foraminal extension of the nucleus lateralis septi passes backward and far


458 C. JUDSON HERRICK

lateralward along those fibers of the fimbria which enter the commissura pallii posterior (figs. 56 and 57). This nucleus I term the nucleus of the commissura pallii posterior. It clearly is a part of the precommissural body. Thus the commissura pallii posterior, like all of the commissures in the lamina terminalis, is embedded in a matrix of cells belonging to the precommissural body, though in this case these cells lie far laterally and not in the velum trans versum itself.

The relations of the commissura pallii posterior to the precommissural body were first described by Elliot Smith ('03) in Sphenodon and lizards. Two embryos of Sphenodon in the Harvard Collection permit a re-examination and full confirmation of his description. Figs. 58 and 59 pass horizontally through the crossing of the commissura pallii posterior in the velum transversum in an embryo of 25.2 mm. Fig. 58 shows the nucleus of the commissure at its greatest extent and lying much nearer the midline than in Lacerta. Fig. 59 is 70 micra farther dorsal nnd illustrates its farthest extension into the fimbria. The relation of the nucleus to the precommissural body is broader and more evident than in Lacerta and similar in plan.

Kupffer ('06) has described the brains of advanced embryos of Anguis fragilis in which the structure of the septum is clearly brought out. In fig. 60 I present a copy of his drawing of a section through the lamina terminalis. The precommissural body is very large and is termed eminentia medialis. At the level figured it is divided into a ventral part forming the commissure bed (tr.), a lateral part (em.), the nucleus lateralis septi, and a dorsal mass (m'). These relations conform very closely to those of the lizard embryo (fig. 54). Kupffer's nucleus, m, which he regards as part of the septum and identifies with Gaupp's ganglion septi of the frog, is the primordium hippocampi and his fissure (J.c.h.), is the fissura limitans hippocampi of my description. A section taken through the commissura pallii posterior shows an "unbekannter Kern" (Kupffer's fig. 259, k) which is clearly the nucleus of this commissure, as described above for Lacerta.

Figs. 61 and 62 illustrate transverse sections through the brain of an adult lizard, Plnynosoma, the so-called horned toad of the


MORPHOLOGY OF THE FOREBRAIN 459

western United States. In this animal the nucleus lateralis septi is smaller than in many other lizards.

In the lamina terminalis (fig. 61) the precommissural body has risen above the preoptic recess and forms the "bed" of the commissures. It is broken up by the fiber tracts of the anterior commissure, commissura pallii anterior and fornix into detached clusters of cells, one which (n.c.p.), extends caudad above the interventricular foramen as in larval Lacerta. The section figured is slightly oblique, the left side being farther caudad and showing the relations immediately rostral to the foramen. The relations of this nucleus to the commissura pallii posterior caudad of the foramen are seen in fig. 62. Dorsally of these parts of the precommissural body is a small residue of the primordium hippocampi, filled with fibers of the fimbria. The fissura arcuata is incomplete, not being marked on the ventricular surface, this region being invaded by the nucleus lateralis septi. The cortex of the dorso-medial wall (cortex hippocampi) is much more highly developed, with more numerous and more densely crowded cells, than is any other part of the cortex. Fig. 62 is taken through the foramen interventriculare on the right side and the commissura pallii posterior just behind the foramen on the left. On each side is seen a small remnant of both the precommissural body (n.c.p.) and the primordium hippocampi.

Figs. 63, 64 and 65 illustrate horizontal sections through the brain of another lizard, Sceloporus. The sections are slightly oblique, the left side being a little farther dorsal than the right. The left side of fig. 63 shows the dorso-median cortex at its greatest extent, a condition which prevails at all levels dorsally of the one figured. Immediately ventrally of this level, as shown by the right side of fig. 63, the caudal wall of the hemisphere becomes membranous and the thick primordium hippocampi appears in the median wall. Fig. 64 passes through the interventricular foramen on the right side and shows more densely crowded masses of cells surrounding the foramen ; these belong to the precommissural body. The left side shows that this nucleus extends above the foramen also as nucleus of the commissura pallii posterior (n.c.p.). Rostrally of the foramen is the thick septum, which is


460 C. JUDSON HERRICK

bounded farther forward by the perfectly differentiated dorsomedial cortex. A remnant of this cortex is also seen at the caudolateral angle of the hemisphere. Similar conditions prevail ventrally of the foramen (fig. 65), save that here the precommissural body occupies a much larger part of the "septum," as well as the "bed" of the commissura pallii anterior.

Laterally of the median plane the commissura pallii posterior lies in contact with the stria medullaris for a short distance at the rostral border of the pars dorsalis thalami and habenula (figs. 52, 64), and some of my preparations of Phrynosoma and Sceloporus seem to show a fibrous interchange between these tracts. But my specimens of these lizards are not well adapted for the study of fiber tracts and the nature of these fibers must be left undecided. Most of the fibers of the stria medullaris pass farther caudad to reach the habenula and superior commissure and do not share in this contact with the commissural fibers.

Many years ago Kupffer ('93, p. 57) commented upon the fact that in the Amphibia the superior commissure lies farther rostral as compared with the other elements in the diencephalic roof than in most other vertebrates, and in his last work ('06) he figures the relations of these elements in a large series of vertebrate brains. In Amphibia the relations of the fibers of the commissura pallii posterior have perhaps determined the position of the superior commissure and in lizards we seem to have an intermediate condition. In all higher forms there is no direct connection of any sort between the posterior poles of the hemispheres and the diencephalon, probably by reason of an increase in the extent of the posterior chorioidal fold and fissura chorioidea.

Crocodilia. — The brain is more highly differentiated in the Crocodilia than in other reptiles; and T shall describe briefly some of the features of the brain of a young specimen of Alligator mississippiensis 25 cm. long, stained by the silver reduction method of Cajal.

In the mid-region of the septum (fig. 66) the nucleus lateralis is seen to receive fibers from the columna fornicis and commissura hippocampi as described by Cajal ('04, p. 1055) in the mouse. The wide fissura arcuata extends far rostrad, as in lower mammals,


MORPHOLOGY OF THE FOREBRAIN 461

and its floor is occupied by an undifferentiated primordium hippocampi, filled with fimbria fibers and containing scattered large cells. Comparison with embryonic and adult turtles indicates that the position of the nucleus lateralis septi laterally of the primordium hippocampi is secondarily acquired. The nucleus lateralis arises wholly ventral to the primordium hippocampi and grows upward into its present position late in the ontogeny.

The columna fornicis passes downward rostral to the anterior commissure to enter the medial forebrain tract for the most part farther forward than the plane of this section. The tract designated fornix (c.f.) here contains also external arcuate fibers, i.e., axons of the cells of the nucleus lateralis septi.

Fig. 67 passes through the rostral border of the anterior and hippocampal commissures and the caudal end of the nucleus lateralis septi, above which lies the primordium hippocampi containing scattered cells and fimbria fibers. Associated with the commissura hippocampi is a nucleus of densely crowded cells, which is clearly homologous with the nucleus of the commissura hippocampi described for Amphibia. Here it lies ventrally of the commissural fibers and just caudad of the level figured its cells cross the median plane. Cells of the precommissural body surround this commissure, the anterior commissure and the fornix and are mingled amongst their fibers, but this nucleus is sharply circumscribed and easily distinguishable from the precommissural body.

At this level the precommissural body is bounded above by the recessus superior and below by the preoptic nucleus. It disappears immediately caudad and there is no pars fimbrialis extending backward above the interventricular foramina. The tract marked fornix (c.f.) in this figure contains external arcuate fibers, columna fornicis and tractus cortico-habenularis.

At the level of the foramina (fig. 68) the nucleus preoptic us is greatly enlarged and large fiber tracts pass between it and the corpus striatum complex.

The nucleus of the hippocampal commissure is in contact laterally with a similar nucleus among the fibers of the tractus cortico r habenularis as they pass upward and laterally from the fornix

THE JOURNAL OF COMPARATIVE NEUROLOGY AND PSYCHOLOGY, VOL. 20, NO. 5.


462 C. JUDSON HERRICK

toward the stria medullaris. I shall term the latter the nucleus of the tractus cortico-habenularis. These two nuclei have been differentiated from the single commissural nucleus of Amphibia, and they represent here also the rostral end of the pars ventralis thalami (cf . fig. 69) . Fibers (probably collaterals) from the commissure hippocampi and tractus cortico-habenularis end freely among their cells. Other fibers from the same sources pass through the nuclei to enter the stria medullaris and still others pass between the nucleus of the tractus cortico-habenularis and the corpus striatum complex. The significance of the latter fibers is uncertain. Both nuclei, like their precursor in the Amphibia, clearly represent a correlation center of some sort interpolated in the cortico-habenular path, probably for efferent discharge into the somatic motor centers.

The nuclei of the septum, including the bed nuclei of the anterior commissure and commissura pallii anterior and posterior, belong morphologically in the pars ventro-medialis of the hemisphere and constitute a correlation center interpolated between the olfactory bulb and the cortex hippocampi on the one hand and the hypothalamus on the other hand. In amphibians, reptiles and mammals columna fornicis fibers (or collaterals) end in this nucleus, which is also related by association tracts with the corpus striatum complex