Se ”, beh ¢ * rf ‘ ¥, ‘ rece ag bs ¥, = # Eo + > Se en fy" as ? $6 "y ee . srw Vast paren - wars Oh Si > + pete 4 hy vyete Share aye © ane ~ oo as oe mele ?) ¥ ay ?. w A. ve ve " Ne ~ iy +, 4 (" ‘ < ye Te a is ir - Wig a tee . > ww .s ~ ene a9 ¥9 ¥ ¥ « . * . 3 +9 es t's wade a -* +e ee . 1 eee a Fo 2g 6 0 ode , é Lond Pos . Oo ? ] « ee x ~ y % . oe hah r oy ay es TH nity re or mr i THE JOURNAL OF COMPARATIVE NEUROLOGY AND PSYCHOLOGY EDITORIAL BOARD Henry H. DoNALDSON ApoLF MEYER 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 Ropert M. YERKES, Harvard University C. JuDSoN HERRICK, University of Chicago Managing Editor VOLUME 20 1910 PHILADELPHIA, PA. THE WISTAR INSTITUTE OF ANATOMY AND BIOLOGY 11h a, Slaven, CONTENTS 1910 No. 1. FEBRUARY DonaLpson, Henry H. Further observations on the nervous system of the American leopard frog, Rana pipiens, compared with that of the Kuro- pean frogs, Rana esculenta and Rana temporaria. With two text figures. 1 Harat, SHINKISHI. On the length of the internodes in the sciatic nerve of Rana temporaria (fusca) and Rana pipiens:—being a reéxamination by biometric methods of the data studied by Boycott, 1904, and Takahashi, LOOSAE With threentextalteures:.. ++ 14... eee reesei ete ter cee ors 19 No. 2. APRIL CHARLES BRrooxover. The olfactory nerve, the nervus terminalis and the pre- optic sympathetic system in Amia calva, L. With thirty-five figures..... 49 Henry H. Donautpson. On the percentage of water in the brain and in the Sjonmaelll @oreol Gri avs aillovayoyrethns ATAYS WAV. ose oboe uneneaeene sb eveosotre 119 S.J. Hotmes. Pleasure, pain and the beginnings of intelligence ............ 145 No. 3. JUNE Orro C. Guaser. The formation of habits at highspeed, Withtwo figures... 165 I. W. BuackBurn. On the median anterior cerebral artery as found among the Uiass2 aise, © S\\YA oWisttp-ch 6 oq 1 =) aeRO eee Since CAI aI Ct ORE aian ahs Se Nucl cs. 185 Hansrorp MacCurpy. Degeneration in the ganglion cells of the crayfish, Cambarusibaromi ein, | Withnine fipuresia: 2 0). ...55.-- omens oe eee 195 ALBERT Kuntz. -The development of the sympathetic nervous system in anNTMIS, W/mK REINS obs cUaduasocdsoop He saa mee do duouLactoos 211 H.H.T. Jackson. The control of phototactic reactions in Hyalella by chemi- CEASE) sree as CAs Eo eae ie mst wai hey SMP se MC ee Cet ch] Es bok le le 259 ili gut { iv CONTENTS No. 4. AUGUST C. W. Prentiss. The development of the hypoglossal ganglia of pig embryos. Wight Metres 1. vias eo ok seen A. Ok ee eee ee... ee 265 AuBERT Kuntz. The development of the sympathetic nervous system in birds: “Den: fivures (49 S311. toa Gee ete te Ah ea 283 F. L. Lanpacre. The origin of the cranial ganglia in Ameiurus. Eighty- %07 Cle bt ROPE.” Sree task ot oe cae. se ge ee a) eee eee eee an, B90 No. 5. OCTOBER C. Jupson Herrick. The morphology of the forebrain in Amphibia and Rep- trlta 7) Bish Gy LOU CUES)... :\ 2 oy Dae Oo Sok pea 0s een 414 No. 6. DECEMBER Karu T. Wauenu. The réle of vision in the mental life of the mouse. From ibelort' College: Men meures.:) .Sewemeests ese cue eee eee 549 LawRENCE W. Cote. Reactions of frogs to chlorides of ammonium, potassium, Sod uuimaned ipa aria sy 2 pes aes, s-.07 Rest ahah. heels eke eee eee 601 SeERGius Moreuuis. The movements of the earthworm. Two figures......... 615 HDEPORVAL ANNOUNCEMENT, 2 .)2: cae as fe-8 ea 2 be ees ee 625 (a en : 'C.W. Prentiss. The development of the hypoglossal ganglia of pig embryos. CIP HCH UTES Hee Sites rye. cess Soclels 2) a ke shod Oe eae es 265 / Apert Kinrz. The development of the sympathetic system in birds. Ten AP ULES pee oe Aero cna ale x SS ass ee Ee es Re ne 283 F. L. Lanpacre. The origin of the cranial ganglia in Ameiurus. Eighty- eighictoures' seer pars: acess a), «. 2's oa ela ee Oa Ee ee eta ees 309 SUBJECT AND AUTHOR INDEX the origin of the cranial ganglia MEIURUS, in, 309. Amia calva, the olfactory nerve, the nervus termi- nalis and the pre-optic sympathetic system in, 49. Amphibia and Reptilia, the morphology of the fore” brain in, 414. Artery, on the median anterior cerebral, as found among the insane, 185. IOMETRIC methods of the data studied by Boy- B cott, 1904, and Takahashi, 1908, being a re- examination by; on the length of the internode in thesciatic nerve of Rana temporaria (fusca) and Rana pipiens, 19. Birds, the development of the sympathetic nervous system in, 2838. BiackBurn, I. W. On the median anterior cerebral artery as found among the insane, 185. BRooKOvER, CHARLES. The olfactory nerve, the nervus terminalis and the pre-optic sympa- thetic system in amia calva, L., 49. HEMIC ALS, the control of phototactic reaction in Hyalella by, 259. Coie, LAWRENCE W. Reactions of frogs to chlorides of ammonium, potassium, sodium, and lith- ium, 601. Control of phototactic reactions in Hyalella by chem- icals, 259. Cranial ganglia, the origin of the, in Ameiurus, 309. Crayfish, degeneration in the ganglion cells of the, 195. EGENERATION in the ganglion cells of the crayfish, Cambarus bartonii gir, 195. Development of the hypoglossal ganglia of pig em- bryos, 265. Development of the sympathetic nervous system in birds, 283. Development of the sympathetic nervous system in mammals, 211. Donautpson, 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, 1. Donaupson, Henry H. On the percentage of water in the brain and in thespinal cord of the albino rat, 119. | har the movements of the, 615. OREBRAIN, the morphology of the, in Amphibia and Reptilia, 414. Frogs, the reaction of, to chlorides of ammonium, potassium, sodium and lithium, 601. ANGLION cells, degeneration in the, of the cray” fish, 195. GLASER, Ortro C. The formation of habits at high speed, 165. I I ABITS, the formation of, at high speed, 165. Harat, SHINKISHI. On the length of the internodes in thesciatic nerve of Rana temporaria (fusea and Rana pipiens: being a reéxamination by biometric methods of the data studied by Boycott, 1904, and Takahashi, 1908, 19. Herrick C. Jupson. The morphology of the fore- brain in Amphibia and Reptilia, 414. Hotmes, 8. J. Pleasure, pain and the beginnings of intelligence, 145. Hyalella, the control of phototactic reactions in by chemicals, 259. ‘ Hypoglossal ganglia, the development of the, of pig embryos, 265. vl INDEX. NSANE, on the median anterior cerebral artery as found among the, 185. Intelligence, pleasure, pain and the beginnings of, 145. Tnternodes, length of the, in the sciatic nerve of Rana pipiens, being areéxamination by; biomet- ric methods of the data studied by; Boycott, 1904, and Takahashi, 1908, 19. acxson, H. H. T. The control of phototac- J tic reactions in Hyalella by chemicals, 259. untz, ALBERT. The development of the sympa- thetic nervous system in birds, 283. Kuntz, ALBERT. The development of the sympa- thetic nervous system in mammals, 211. ANDACRE, F’. L. The origin of the cranial gan- glia in Ameiurus, 309. acCurpy, HANsForD. Degeneration in the ganglion cells of the crayfish, Cambarus bar- tonii gir, 195. Mammals, development of the sympathetic nervous system in, 211. Mental life, the réle of vision in the, of the mouse, 549. Moreuiis, Seraius. The movements of the earth- worm, 615. Morphology of the forebrain in Amphibia and Rep- tilia, 414. Mouse, the réle of vision in the mental life of the, 549. Movements of the earthworm, 615. ERVE, olfactory, the nervus terminalis and the pre-optic sympathetic system in amia calva, 49, Nervous system of the American leopard frog, European frogs, 1. Nervus terminalis, the olfactory nerve, and the pre-optic sympathetic system in Amia ecalva, 49. Ca of the cranial ganglia in Ameiurus, 309. Pi pleasure, and the beginnings of intelligence, 145. Pig embryos, the development of the hypoglossal ganglia of, 265. Pleasure, pain and the beginnings of intelligence, 145. Prentiss, C. W. The development of the hypo- glossal ganglia of pig embryos, 265. ecay nervous system of, 1. Rana pipiens, on the length of the internodes in the sciatic nerve of, 19. Rana temporaria (fusca), on the length of the inter- nodes in the sciatic nerve of, 19 Rat, on the percentage of water in the brain and in the spinal cord of the albino, 119. Reactions of frogs to chlorides of ammonium, potas- sium, sodium and lithium, 601. Reactions, phototactic, control of, in Hyalella by chemicals, 259. Reptilia, the morphology of the forebrain in Am- phibia and, 414. of the, in birds, 283. Sympathetic nervous system, the development of the, in mammals, 211. See nervous system, the development Sympathetic system, the pre-optic, in Amia calva, 49. Wen role of, in the mentallife of the mouse, 549. ATER, on the percentage of, in the brain and in the spinal cord of the albino rat, 119. Wauau, Kart T. The role of vision in the mental life of the mouse, 549. FURTHER OBSERVATIONS ON THE NERVOUS SYS- TEM OF THE AMERICAN LEOPARD FROG (RANA EPPIEINS) “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 Kuropean 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 Physio- logical Institute at Zurich, where I had examined R. esculenta in 1904, and to Professor Sherrington for similar privileges at the Dy HENRY H. DONALDSON 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 noo monn TOs | WEIGHT IN GRAMS OF | Rega teeag OF WEIGHT SEX LENGTH IN |LENGTH IN | | IN GRMS. MM. MM. i | | | C.N.S. Brain Sp. C. | Brain Sp. C 11.6 M. 130 | .0918| .0666| .0252| 84.4 | 79.4 16.0 M. 150 11481 .0796| .0352 |- 85.2 | 80.7 1720 F. 159 1054.| 0714.) ~.0340 | 84.07 | 8086 20.8 M. 170 “79324 oda!) 70888 | 8502 || ssi 22.5 M. 162 | .1165 | .0807| .0858| 84.5 | 80.4 26.4 M. 180 .1372 | .0946 | .0426| 84.4 | 78.4 27.6 F. 179 | [24164 ator | 0402-15 (S478) slut 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 Sono 43.9 M. 200 .1708 | .1210| .0498 | 85.8 80.7 47.0 M. 198 .1664 | .1140| .0524| 84.4 80.5 NERVOUS SYSTEM OF THE FROG TABLE 2 Data on R. esculenta, Zurich 1904. 11 specimens DODe ROA. RGD WEIGHT IN GRAMS OF ET eee oh WEIGHT | SEX LENGTH INLENGTH IN IN GRMS. MM. | MM. C. N.S: | Brain | Sp. €. Brain Sp. C. 12.40 Ii iil | .0818 | .0577 .0241 84.2 78.4 16.75 Jihe 144 | .0926 | .0634 .0292 83.4 79.1 18.43 F. 144 .0928 | .0650) .0278 83.2 (See 20.00 F. 161 S103 T POO Gn mr Os47 Sy) WOn2 22.00 164 | .1107| .0769| .0338| 84.0 | 79.0 24.10 M. 167 | e217 OO Saal .0376 83.4 78.4 33.85 M. | 7s | be'so7 0895 .0432 Soee UP SGqsOctiMss i Lah, | | 1478 |, GOO4s) 20474.) 8ae4 78.6 37.56 Vike SSH | .1490 .0993 .0497 82.9 78.5 37,.96.| F. "194 | 1427 | 20953,|) (04747|, "82.8: | 47.3 45.03 ae 196 | eb S37A8} .1078 .0500 | 83.9 | 78.4 TABLE 3 Data on R. esculenta, Zurich 1909. 11 specimens ODS men TOS | WEIGHT IN GRAMS OF | ed Sa ey WEIGHT SEX LENGTH IN LENGTH | IN GRMS MM. IN MM. | C.N.S. | Brain | Sp.C. | Bram | Sp. C. | 18.9 M. 143 57 3 | .1047 .0707 | .0340) 83.6 78.6 24.7 F. 167 65.0 .1065 ROM | O346) | Ni SorOmnoee 26.5 M. 167 63.4 a2 0) 0737 .0383 | 83.6 Ud tei 30.9 M. Mii | (OO .1198 .0830 .0368 | 83.9 78.5 By) 8} Ke 1830 u6SE0 OM .0873 | .0428 | 83.6 78.1 33.0 F. 184 | 70.5 lia | .0845 .0430 | 83.4 77.0 3 Bey | qelS88 |) (42:5 .1435 .0985 | .0450 | 83.8 | 78.2 47.4 1 | 204 | 80.0 . 1593 O83) {Oesi0) | feisil 7S) 48.4 ine | 193 79.3 .1545 | .1027 .0518 83.8 Utell ipa! We 20S 82.0 .1589 MOS) 204845) S3eOn 191 58.0 Eva ne DIG $7.0 .1858 .1278 .0580 | 83.6 | HAG HENRY H. DONALDSON TABLE 4 Data on R. temporaria, Liverpool 1904. 12 specimens BODY WEIGHT IN GRAMS OF PERCENTAGE OF BODY TOTAL WATER WEIGHT SEX LENGTH IN | LENGTH IN IN GRMS. | MM. MM. | C.NJS: |) “Brain Sp. C. Brain Sp. C. 14.05 iB: 144 .O881 .0596 | .0285 | 82.3 78.2 HG ELON ey 151 .0991 .0690 -0301 | 82.7 79.0 17.65) M. 154 .0916 | .0618 | .0298| 83.0 | 78.5 21.75 M. lal 1045 .0671 | .0374 | 82.8 78.2 23.45 | M. 162 .0947 40628 | .0319)) 82.1 77.0 24.17 Be 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 | ahaa) 20813 | “Oolt |) (sie3 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 Wy elilan .0766 | .0396 | 82.7 78.0 TABLE 5 Data on R. temporaria, Liverpool 1909. 16 specimens sGiDs | Toran ODT WEIGHT IN GRAMS OF ui eae or WEIGHT SEX LENGTH IN | LENGTH IN IN GRMS. MM. | MM. | CINE? | Brain Sp. C Brain Sp. C. 14.5 F. 148 | 53.3 .0772 | s0522) | 4 02507 0n S222 76.8 17.2 M. 154 56.0 .0903 | .0598 .0305 | 83.6 79.6 1931 iE. 155 58.5 .0907 | .0621 .0286 | 82.4 76.3 21.0 Meo 7 062 60.3 .1014 | 10663°| 70351) 83.8 79.5 24.2 Mi jp 276 64.5 -1099 | .0736 0363 | 83.9 78.5 25.4 Me |, eae 60.0 .0994 | .0672| .0322 | 84.5 78.2 26.0 M. 162 | 60.8 | .1066 | .0702 0364 | 84.0 79.4 26.1 F. Re 6528 | SIT | .0787 0404 | 84.0 79.4 26.9 F. 163° || 63:2 10929) 0787 |” 30855.) 9 18424 76.3 27.9 By | 175 | 66.7 -1149 .0786 | .0363 | 83.8 79.5 29.2 Nestea ellie! 66.5 .1114 | .0744 | .0370| 83.8 79.5 29.4 M. 170 66.0 .1356 | .0887 | .0469 | 83.8 78.9 29.8 MM. } ~=6168 62.5 PLLC Te een Omoll .0416 | 83.8 78.9 32.1 F. 184 73.2 .1314 0864 .0450 | 84.3 79.2 33.3 M. 167 64.0 .1373 0900 | .0473 | 84.3 79.2 39.1 fe 196 76.8 . 1452 | 0964 | .0488 | 82.3 77.8 NERVOUS SYSTEM OF THE FROG & 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. esculenta 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 ease this departure from the rule is indicated in the condensed tables (6, 10, 12, 18,) 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 vear, 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 99.80 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 BODY WEIGHT BODY WEIGHT PER IN GRAMS MILLIMETER, IN GRAMS 14.9 .102 Rot 2302 .135 R SS Ae cle ca RES Cae EE 2. ony OER ae 30.8 168 ND ppailtss 15.9 .114 99 Rerescullentar LOO ey siete «kos ac i: oS bP ek REY yo eae : 22.0 134 35.0 [2]* .199 40.2 .208 23.4 .146 FR -esculemta all9 OO ae sts 5 iiss olan o Die dae te oe he eee 32.1 W7 43.8 .225 | 55.2 [2] 262 (15.9 .107 Restemporarianl 904 ees tani ace Anta Reo ee 23.1 137 28.0 62 Bil os! sheer 18.0 [4] .116 . DAS) 2 151 Ri cemporaria LOO, . ; copes cee. eis. eee eae a ee 26.9 . 159 2985 .173 34.8 .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 inthe years NERVOUS SYSTEM OF THE FROG i 60 é oO | fe) 6 18 Li SMS 3 (a Ss é SS 2B = W : 2 4s Fy=z 5 o iva) alls rs) = |Z 4 PS Sees 7e si (oc ch [oGe sik eO#o 2 @) alr r F Sia |, is) a a 8 a O oO ie O = ° 48 O ie) Be 0 o re: | [ag © © < “ x N nN Q. 4 zg 2 se cu 40 Cuart 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 PER CENT [Ree | CUOMO, cisteteG ech Sen Oe Recto 75 ‘cd Slog ota oeMeue aye 9 66.6 IRMeSGulentalaeentecn Satyr, «oak are ee ere eee kes Ss ee eles 11 65.1 | ESP LEY 001) OO) GEN GILT ain ctr cee EG 6 hos oed GRSeSONCE aA 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 treat- ment the amount of flexion was reduced, and the total length thereby slightly increased. This naturally reduced the percent- age value of the sum of the lengths of the leg bones, the measure- ments 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 g 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 | | FOOT | NO.OF | FEMUR TIBIA | (TARSUS | SPECIMENS | AND PES) — | per cent | per cent per cent iR. joie MOM eens ee 12 De OO GS IR joryercals TOUS Steak VG ok | 19 26.0 | 29.3 44.7 IRi; @aoulkerontay OO, no 5 Sota c cod ote OO BIee o 6 5 26.3 28.2 45.5 Joe Gs Ouillenninen NOOO 6 cts oc oo. 5 010 Oe ee 12 26.8 28 .2 45.0 FuRcemlpOrania 904% Weems cieke snc. - ach cus ic = 6 AG) 28.7 45.2 ReRGeTMORATanl OOO meen ewe miei, = oie) a. c/a 3) « 16 25.7 28.3 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 CEN- TRAL 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 value of the length of the entire central nervous system—the total length of the frog being taken as the standard PERCENTAGE VALUE OF THE LENGTH OF AVERAGE THE ENTIRE CENTRAL NERVOUS SYSTEM NO. OF SPECIMENS | janes MM. as | Rana Rana Rana pipiens | esculenta temporaria CTPA CLROY OY: Boe th Se Gy Oro da ARS Bacio 152 aD 1 desi eea acs RRR AOS oe, BU ENOL | 155 17/510) Paha eh ae vind much ce rit Ova a oh By OSPR CR NCS 159 18.4 Ser cone Seance ates Gestohs aie Steve Mtn OR ene 167 17.6 Ojala Sear crn or ny ture oe eM a CIR RENE ill We 8 QR GUGOA Wire tae seen ce cs naimtee eh sie. 5. yjekevors 176 16.7 Bis ASS RAED EAD AIS O70 Coe a ee 181 16.9 | DEES A ERE ire a cr tineneh one, & Gla oeaige 182 | 1a 2 Mee: Une. Ap: cedar | 195 pe alee AM CLOOA) Rc Maree tena aais ss ovale 196 16.3 9 DEMIS ONSER er Pn)! oe | 210 16. (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 con- sequence 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 Euro- pean forms. The fact that the values for the weight of the cen- tral 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, -09 Gms. NERVOUS SYSTEM OF THE FROG AR. pipiens WEIGHT OF C.N.S. e . aX e O a rx O — id @ R.escutr. 1904 oO ” “ 1909 = R Temp. 1904 i = oOo " 1909 5 O O 20 CHART 2. BODY WEIGHT Gma 25 30 35 40 45 50 Showing the weight of the entire central nervous system. 55 fa 6a 1 ys 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 23.2 . 1256 Re DIPlens ex othe eee eet. aaa ke ee ee 30.8 1463 43.2 . 1674 (15.9 .0890 22.0 .1142 Re escmlentaell9 Q4 gers ar acceso -c SRa ROR ee oLee 35.0 [2] 1402 (40.2 . 1498 23.4 .1077 32.1 . 1258 Rvesculem tal OOM es scerersars, sicic apen ee tie cele, < care 43.8 1524 (55.2 [2] .1724 (15.9 .0929 ‘ 23.1 1108 RetemporaniarlO04 ww... cin sad. See oe ae wae 28.0 1213 | 31.3 1323 (18.0 [4] 0899 | 25.2 1053 RatemporariaelG 09 Ree s.r eS 352 «5c ee Oe oe 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 conveni- ently shown in still another way. I have been able to point out (Donalsdon ’02) that a fairly accu- ate determination of the weight of the central nervous system in frogs can be made by the formula C. N.S. = (Log. Bd. W. 4/1) where C. N.S. is the weight of the central nervous system, Bd.W. the body weight in grams, / 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 ner- vous 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 VALUE OF C. BODY WEIGHT R. pipiens 1904 \ 28 .0 ZGe2 Average of 12 records if ‘ R. esculenta 1904 Average of 8 records TEP Tes coo Sow bcd nla 32.4 24.6 First “weight group”’ omitted R. esculenta 1909 AW CRA CCLOMORECOLUSM NOL ye eh fscnc.s cals SORA eee ces oe 33). dl 23.0 Last “ weight group”’ omitted Difference 1.6 R. temporaria 1904 | ; 24.6 22.8 Average of 12 records | 4 R. temporaria 1909 | Average of 16 records { SL iste Mobic ails) ‘s; elel wishes. 0: eli eta iaiterapepeielciteiejiel ef «(si = 26.8 21.9 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 pre- viously 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 com- munication (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 aia WEIGHT | aVERAGE VALUE FOR C. | GMS. BORE its Bs - 1902 2S 28.6 AEE 3 Pe tl ee 1909 PART 5) NERVOUS SYSTEM OF THE FROG ts 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 speci- mens 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 rela- tion 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. Finally 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 14.9 Be ae yah 2 Pepys Mate] OPH OIC AST, Peet, cde ico Ser RM EOEMERCKE One ore, Sh sotinde toch SSS Camels ee 30.8 9 39 | 43.2 2.28 15.9 2.29 Pepa) Zee esculenta (AM aret ar ci.-.s 3 < wie CUO OREC IER Eire cts at ene ee | 35.0 [2] 2.09 | 40.2 2.05 23.4 74,0733 R 1 1909 Del 2.09 ESCM MCA OOO se ret Lehn c cis. 5. ss MIRE eet Reale ce ee eee 43.8 2.05 soe e 2 2.24 15.9 215 ' eal 1.86 VE temporaria GOs scat gees) cre etc Pe eas ae ne ee tee 198 0 1.87 [31.3 iss 18.0 [4] 2.02 Wy 2502 lReprllek at oY oye?) alte atl |(0)0)2 a: Ue te. ee me em CER OL ame ee OC ak Mn 4 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 ily 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 the spinal cord. Averages from groups of three BODY PERCENTAGE OF WATER IN WEIGHT Brain Spinal Cord 14.9 84.5 "80.2 R. pipiens 23.2 84.7 80.1 ‘ SR eh RE RRRETS oxs.8o3. 3, va, cue Oe ale aan Sie ANS3 Pe 85.7 81.2 (15.9 83.6 78.6 Reesculembapl| OG OQAR pe eee ees ss Shs econ crn pre ee 2210 83.3 18.9 BORO 2] Sono 78.4 40.2 83.2 fishes 20.4 83.6 78.6 9) € ye Resculentarl GO Oneee teers: eh in TRO) Os ws ge hes Os lia eI 4.0 425 578 Srey ie | S80 714 Gai a ads 805 TT a ue crs a a hee eR 5.3 623 834 6.3 645 917 73 805 1039 capiC ico ene eee... ss. ee Hee sone S71 837 6.3 | 828 963 Tesh SSS 1001 ACen er ear’... (Lae 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 differ- ent 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 diam- eter 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.” 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 per- mit 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 diam- eter increases towards the periphery, but the three lines repre- ? Takahashi has pointed out that the internodes on the fibers from the American frog, R. pipiens, are shorter than those found by Boyeott on the fibers of R. tem- poraria (fusea). 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. Thistrelation agrees with that found to exist between the two characters in the frogs of different sizes. We feel justified there- fore 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 follow- ing final conclusion :—The exponential equation y Be A eh 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 fhe 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 thefactors, Takahashi counts the segmental influence, by which I understand the elonga- tion 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 con- cluding that the growth of the segments in Rana pipiens follows also an exponential formula. Although unfortunately I have not- sufficient 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 incre- ment 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 seg- ments 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 elonga- tion of the segments, we shall find a progressive increase in the length of the internodes towards the periphery, and the exponen- tial 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 differ- ent 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 Biren, 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. Boycort, 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. 370-380. Davenport, C. B. Statistical methods. Second edition, New York. 1904 Kewuicotr, Wm. E. Correlation and variation in internal and external characters 1907 in the common toad (Bufo lentiginosus americanus, Le C.) J. of Exper. Zoél., vol. 4, no. 4, pp. 575-614. TakauHasuHI, K. Some conditions which determine the length of the internodes 1908 found on the nerve fibers of the Leopard frog, Rana pipiens. J. of Comp. Neurol. and Psychol., vol. 18, no. 2, pp. 167-197. Showing correlation of length LENGTH OF THE INTERNODES CORRELATION TABLE 10 TABLES. and diameter of internodes in micra 45 GROUP I Bites |< -| a | eee eee Bele | = SRS ee! Se | |) en ce EES Nias Sh eee | 2 | 2 | amma eh dl ee Diameter | Ss | a ee Poa 3 S Re HD Ss | = 2 | A on] Crh ae | hot | 1 Dae, SELLS. } 2 | | | | 2 eee Pe ee | | ces 7 9 3 1 | 23 Od emer a e Oatets wa | 2" Seana Ee | te ye as Te eee Raed} 08. 23 TMS eS aul yo | 48 Gua Menara cet 1 | 3 | <3 | LORS a2 Ig 23 Doe aa eee ae | | 6 1 | 7 I ORD oe a ap ee re | 1 | 2 Leet 8 Motels, tees eo. 18- | 38 | commoneing acu = Raley | TABLE 11 Showing correlation of length and diameter of internodes in micra GROUP II 3 | | | | SS | S) =) =) i) ro) Sy sis 2/2/28) 2/8/8/8/2/2/8/2/8/ Sle = Hig/e/¢/2/8is/el/el/e|4/4/4/sleigi2 Pramae tere SE eS Sree ihe rch” | 2 cee imeeenaae tgm CN SS BES ee Sed 4yu Po Pirie alle aoa | | 7 5 AG \ 20/08 (15) Z| “3; \a igen | i Gs an! | Zui 30/\) 5001/32 | 15 | 9) ea | | 1] 154 if | AaletSul gl 85-53: | S3a) Tale een leo | 289 ee Sy ielse| 68, (L06, | 57 |.30 (e220 iea 61\.-3 318 9 eal B39) 30 | 37 | 18t aero el 152 HONE), ea 529) | 37. |. 22.| 6 ope 1 106 1 4 | 1 dike 3 Totals| 2 | 16 | 95 [286 [306 [191 /103 | 60 | 21/13) 3| 0 | 0 | 0/1 1097 46 TABLE 12 SHINKISHI HATAT Showing correlation of length and diameter of internodes in micra GROUP III © ] | } St S i) elelelel/elele a8/8/8/818/2/8/8/8/2/81/8/8/8/8/8/2/8/8) 3 Diameter i yi | “Peale fala fal ap al al A A & -|—— Sy 1 ek) | | 1 4 3 Brot | 7 5 | 4 14 23 8 5 2 a) | 57 6 1] 6} 19] 29) 22! 17 eae 106 a 20| 81) 77| 29) 23) 12) 3} 2 247 8 2| 16 57| 80) 73| 37| 24) 16) 4) 4 315 9 | | JI 8 44! 88} 87| 40 18] 6 2 ia fe 295 10 | | | 4| 551111134] 80 41] 200 2 | 3 1) 1) 452 11 ctl PS) ab SS) Pel 20 12 1 gets esl Pa | | 14 13 1 Ht | 3 — a = —— Totals | 1) 17| 83291/400/354)203'101| 45/ 10, 4) 3) 0} 1) 0} 0] ©} 1) -1| 1515 TABLE 13 Showing correlation of length and diameter of internodes in micra GROUP IV 2\8lclelele|2 Slzlzigislsieisisisis SSSR SF FIRS See tial a F181 2 Diameter : : E : : E z : | 5 : 5 = : : z E 2 | 31 1| 2 3 4 0 5 1 | 1 6 1 40 3 4 8 | 15 th Ges 7} 7 6h aie 36 8 2 a. 5) 2) Aesae 24 9 I] 4| 6 17 24/13) 7] 1) 2 75 10 5) 221 33| 50 33) 211 14) 17) 6 2 2 2 1208 11 111012) 10 15) 14 10 3) 5) 3} 3} 1 | 87 12 4| 6 17| 19) 14/10/11) 3| 4| a 1 90 13 Oe Al Tl. (G)eSie Oe araraee BL 1 | 46 14 } 5) 6 2 2 1 17 15 1, ite 3 Totals scree 88| 131} 106 74] 41) 36 19 10, 8 4 1) O| 11605 LENGTH OF THE INTERNODES 47 TABLE 14 Showing correlation of length and diameter of internodes in micra GROUP V Nl a | | | i Daly } | 3 | elolel|olo|/el/elelelelislelelelelicloloelelo 5 Sigieieiegigjeie/sisisis eSlsilelsgisiclicigisisisisicisis OSS rove i=) baal N oD sH Yen) ico) MR OID) O|) SAIN! D/H) 19 |) 1/0) dD oe | ACCT a a [ef ee re ea a i |S at SA a Pa BRS oleiliololollo oo | | o |S S/o] lSiele/Ssiel/elelisieisicic| B = \ 5 | | | | | | ae aa! | | | | 4 | 1 | | ie as | | AL Sepa ge ile i Ta | hae Psa 6 1| 1) 5) 3) 4) 2) 1 1) 2 edie || | 21 7 | 1) 3/11) 16 20, 12; 11) 7 4 a | | 96 Sevetles! slo) LE 20N 25) 24 W412) 7] easement | | 150 9 1) 1/10) 12) 17) 25) 25; 13] 19) 11) 4) 8} 5) 4) 2 1 | | 158 10 2/15] 32] 46) 40| 72, 56) 69) 46) 41) 37/18|11| 9} 3) 2! 3 1) | 1) 3} | | 507 11 2} | 5| 12] 15| 21) 26 34] 42) 19] 24) 19|14/10) 9) 6| 4] 2 | | 264 128 | | 2} 6} 28] 58) 70) 44| 44 40 33] 28/30/1614) 6| | 6 4| 1] 1| 1] | 1) 433 13 1| 7 16) 12] 20 23) 45| 23) 18] 12/13) 8] 5| 8 3) 3/2} | | a) | | 220 14 | Ze sOn 2) A 2! 13) 6) Ia ost | I | Lee etal eee ie | | el 1) 3] 1 Warley eel lel | SSS Pee — bat Totals| 1| 5 5 1255100175200251 191280157 HO ISO5648281115 9 1) 3) 6| 0} 2/1916 if . : i mi lag ae 7 - > TH - Z - a aad Mk oy iia > 7 : i _ a ? . >» s . ( 7 ’ i : i=, ieniel bo +e, hs anes “9.5 & - ot | POSITION OF MODE POSITION OF MODE aN ae al I T ies 350F . \. 300 300 INTERNODES 250 DIAMETERS 250 200 Fig. 2 150 100 50 13 15 MICRA 0) 50 250 450 650 850 1050 1250 1450 1650 1850 2050 2250 2450 2650 2850 MICRA Cuartl. 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 AMIA CALVA, L. CHARLES BROOKOVER From the Anatomical Laboratory of the University of Chicago WITH THIRTY-FIVE FIGURES Ia bs ieyaripr hanya LAW ey ela Vo 70 Let hi eR Pe oe A a 50 ES tonicaliS Ke Gch r peerrrere teres overs fis chile hin), GR RoR ee Ee ce eee pene 52 Embryological Development of the Olfactory Organin Amia................. 57 AGult structuresiof the Olfactory Organ in Amia..&. 22... a.6o- «0.8 se vee lee UU Central Relations of the Nervus Terminalisin Amia......................00 83 Intra-cranial Sympathetic System Posterior to the Nervus Terminalis in ATTA ERE ant ea reeds stole Scene «dake SAE BES SR SEN eed Oe ee Pete) Ae 92 The Nervus Terminalis of Lepidosteus and Teleosts ......................... 106 DiscussionandeltheoreticaliConchusions.. 44. -ceeee teen eene eee eerie one de 108 STAMINA FOL Ges UNUS mae sare eae co sicnctee 2 sn he ee re es EE ee ea lee rar 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 inter- esting 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 appara- tus. 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 cyto- logical results. Most of these sections were stained with Heiden- hain’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 forma- lin. 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 forma- lin 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 clear- ing 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 com- plete 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 medul- lation. 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 chrom- oxalic mixture cited by Houser (’01) gave better results than a2 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 com- pound microscope to determine when impregnation had taken place. The pieces were kept in the ice box while in the ammon- ium 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 sub- stantial 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 con- tributions 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 ol- factory epithelium. Since that time his results have been con- firmed by various workers repeatedly. Bedford (’04) gives a brief summary of some of the more important papers,—espe- cially those of an embryological nature. His résumé of the liter- ature 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 D3 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 em- bryology of the olfactory nerve is simple, but Milnes 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 after- ward 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 Loey (’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 develop- ment in Squalus acanthias (99), but later (05) he came to con- sider 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 vomeronasale (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 vomero- nasale does not exist as such among a largenumber 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 Jacob- son’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 olfac- tory nerve of the carp, and Sheldon (’09) found its central con- nection 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 ao 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 can- not say anything definite as to its function at present. John- ston (’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 cerebro- spinal 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 Ganbili2): 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 how it 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 neurol- ogists now call the olfactory bulbs in which are located the mitral cells and their glomeruli. Olfactory lobes are now commonly used to indicate olfactory centers in the forebrain proper. Con- sequently, 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 macro- scopically I could never be absolutely certain that I saw any- thing different from olfactory fibers, which often break up into various small bundles before joining the olfactory bulbs. Any- thing 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 waseasy for me to confirm his find- ings in themain. Thisledmeto 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 Fi 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 IN 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 acid fuchsin. 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 but- ton appears as a cord of cells tapering downward and forward neural tube 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 hematoxylin 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 mono- rhinism 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 ver- tebrates. 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 olfac- tory 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 can- not 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 produc- ing a red line, between the mesodermal elements and the caudo- dorsal 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 olfac- tory 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 sae will appear later.’ The embryos showed slight motions at this age. About twelve hours later than the last stage, or some twenty- two 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 elonga- tion 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 con- nection 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 de- termined, as the method of Bielschowsky which was tried, did not differentiate any fibers in young or embryonic Amia. The con- nection 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 shght 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. XX 450. Six micra thick. Some cells which have the characters of neuro- blasts 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 dis- tance 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 anter- ior border of the placode. Some three or four cells are seen among the olfactory fibers just as they leave the nasal capsule. Mesen- chyme 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 Fic. 38. 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 peri- pherally, ramifying as small bundles among the cells of the cap- sule. The thickness of the capsular epithelium is not so great in the middle of the mesal side of the capsule. This is opposite &. ts “owt eyed sl Ep oe MR AREY Katharine Witt. ob. Fic. 4. Horizontal section of Amia 6 hours after hatching. Shows two rami of the olfactory nerve asit 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 susten- tacular cells of the nasal epithelium. Later the larger part of these nuclei along the olfactory nerve produce sheath cells for Fic. 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. Stainedin thionin and acidfuchsin. 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, can-- not 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 olfac- tory 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. Consequently, 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, @). NERVUS ' TERMINALIS ‘IN ‘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 ap- pearance. 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- iE +f Lie Diy Pe 7 sips a 6, : BUS eRe y SOREN ea geese a Ls Fic. 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 ain 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 distin- guished 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 in- different 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 ata 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 Katharine Hill. 14°4 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 ¢ 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 recog- nition 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 sec- tions 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 trans- versely. 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 ab- sorbed 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 mi- nalis. The nasal capsule was cut away from one side of the young, one or two days after hatching. The object was to re- move 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 amore 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 olfac- tory 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 termi- nalis somewhat proportional to the size of the olfactory nerve as compared with the ganglion and olfactory nerve on the unin- jured 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 develop- ment 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 ganglon 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 imme- diately establish fibrous connection with theneural tube. Later, nuclei migrate out from the nasal sacs along the olfactory fibers as Disse (’96), and others have found for other forms of verte- brates. Still later some of the nuclei which have wandered out along the nerve form a ganglion, as Carpenter (’06) 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 AMIA ral 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 surround- ing 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 olfactory nerve in the adult is about one and one half times as long as the brain. From the first ap- pearance 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- 12 CHARLES 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 Schneider- ian 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 I olf fibers Ce) Fia. 8. Golgi impregnation of part of the ganglion cells of the nervus termi- nalis 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 ie 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 ace 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 epi- thelium in the section (fig. 9). As the figure shows, the cells are practically all located between the anterior and the posterior ome olf. bulb 4 nerve Fic. 9. Diagram from camera lucida outline to show the distribution of the ganglion cells of the nervus terminalis in young Amia50 mm. long. Delafield’s hematoxylin. X 20. limits of the nasal capsule. One or two cells were found prox- imally 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 bul- bus 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 olfac- tory 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 ventro- medianly 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 ol- factory 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 Ke) ge lateral ramus ~ nervus terminalis y Fic. 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 olfac- tory 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 sub- stance. It appears similar to the olfactory nerve proper and can- not 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 olfac- tory nerve. NERVUS TERMINALIS IN. AMIA 1D 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 differ- ent cases as is well shown on the opposite, or left, side of this same fish where it is never more distinct within the cranial cavity 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 Fic. 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 1t 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. . > ! We | !"y nervus terminalis Fig. 12. From same series as figs. 10 and 11, but farther caudad within the cran- ial 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 cor- puscles 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 cran- ial 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 IN AMIA rer beyond the cranial cavity to reach the nasal capsule, but this last point could not be determined with certainty from the prepa- ration. However, the main blood supply to the nasal capsules is de- rived from the external carotid artery. The branch which sup- plies 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 ap- pearance 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 for- ward 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 arteri- oles 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 aqua- rium containing a living adult, the nasal tube (fig. 13) inhaled 75 CHARLES BROOKOVER the juice. Later the colored fluid appeared at the exhalent ap- erture. 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 exhala- tions. 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 --mid-rib exhalent aperture Fic. 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. XX 3. the blood can be seen circulating laterally each way from the mid- rib 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 acurrent 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 prepara- tions show that there are in Amia three main types of cells (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 cihated cells are most numerous, while the olfactory cells come next in point of numbers. =< = ra 5 => = es seen Og a sa 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 895 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 clear- ness until a large number of Golgi preparations were made of young Amia about 75 mm. long. One hundred of these small olf - k ulk mervus levminalis ee / Noa a oy EE ze Ba A BES at.carotia artery suleus olfactorius Fie. 19. Reconstruction from four sagittal Golgi sections 60 micra thick, of young Amia head, total length of fish 75 mm. AIl non-essential details omitted, to show the course of the nervus terminalis as seen from the median plane pro- jected upon the outline of the most lateral of the four sections. The numbers in- dicate the sections from which a given part was taken. X 50. fishes’ heads were treated by the Golgi rapid process and trans- ferred 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 im- pregnated, 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 vary- ing 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 prosen- cephalon 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 sec- tions 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 ven- tral sections of the olfactory bulbs being smaller in area and the fact that the mass of the bulbs was slightly shrunken and con- tracted 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 8 olfactory newe eruus terminali [ olgaclory bulb / 7 Z sulcus --7 ie olgactorius Fria. 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- 8S CHARLES BROOKOVER five fibers in the nervus terminalis at this age at the point where it joins the olfactory 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 Katharine Hill \ Fie. 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 b. Supporting elements on the median surface of the bulbs atc. 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 dissec- tions. However, I have frequently found a bundle of non-medul- Jated 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 com- pact bundle of five or six fibers accompanies the main artery. From near it two fibers run dorsad along a blood vessel. The we ZA Npactory bulk eas Fia. 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. 225. main bundle, which can be followed from posterior to the suleus 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 accompany- ing 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 rela- tion to the internal carotid artery as shown in the Golgi prepara- tions just mentioned (see also fig. 32). In one Cajal prepara- tion 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 olf, nerve algact ore \ ull Fia. 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. XX 225. the bundle of the olfactory nerve containing nervus terminalis fibers. What was thought to be a nerve cell withits 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 9] undoubted nerve cell shown in fig. 12 near the nervus terminalis. Golgi preparations of adult Amia show that there are branching nerve 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. A pr? olf. nerve Fic. 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 terminatis running far caudad ventral of the brain in Protopterus, it would not be unnatural to infer that the nervus terminalis in Amia extends ecaudad 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 traced 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 an- terior 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 xxl). 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 hypo- physis, 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 6 in fig. 25). The latter ganglion is situated in close proximity to the base of the Peale, eo ---Pinkus’ nerve -brain case cartilage +--ciltary ganglion --pineal stalk -profundus ganglion --paraphysvs --profundu S nerve 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 in- nervates the meninges of the forebrain. Also, it shows the relation of the pre- optic sympathetic system to the pineal nerve and to the nervus terminalis. X 4. dorsal sae (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 sur- rounding 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 pos- terior part of the sympathetic system and the ciliary ganglion, it may well establish a connection from the sympathetic system Fic. 26. A portion of the ganglion cells from point 6 in fig. 25. Camera drawing from the most crowded portion of the ganglion, the meninges being mounted with- out 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 prob- able 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 theevanescent thalamic nerve described by Miss Platt (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 evi- dence 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 (6 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 para- physis 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 dien- cephalic 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 BROOKOVER of the pineal stalk (fig. 30). Kingsbury (97, a, fig. 4) shows the opening of to the duct of the larval paraphysis. There are many instances in Golgi impregnations that seem tome toshow that there is an intrinsic nerve supply to itstubes (fig. 27). Fibers often le in the closest proximity to the blood vessels on one side and to para- physis 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 SSS ety epithelium on ART, = Hs Th ee a | Fic. 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 theforebrain. X 200. blood vessels on the other. The richest supply of nerves is found among the paraphysis tubes to which the majority of the medul- lated 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 examina- tion into its structure and function. It may be said at the out- set 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 jn producing motion in the encephalic fluid. Strips of the epithe- Fie. 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 fore- brain 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 rhinoceels 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 Niss] 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 epithe- lial 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 John- ston (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 asingle layer of cubical cells in theepithellium. These areciliated 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 epithe- lium sometimes has intra-cellular fat globules. There is some- times 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 encephalie fluid in Amia where there are nerve fibers in the basement membrane, we might infer that cilia are con- NERVUS TERMINALIS IN AMIA 99 trolled by nerves. However, Pitter (03) states that it has not been proven that ciliary action is under the control of the nervous svstem except perhaps in a few molluses and annelids. The pineal stalk of Amia has an innervation that showssome 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 habenule, 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 tneal stalk ciliated epithelium We t 1 fore t Fig. 29. From the same preparation as fig. 27, showing the relation of nerves of the meninges to the pineal stalk. 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 cen- trally to be lost between the habenula and the superior commis- sure. 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 shghtly 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 -m. |. dorsal sac . habenula epiphysis Paraphysis tubes paraphysis duct velum transversum pallial fold rhinocoele ant. com. olfactory bulb vasculosis Fia. 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 para- physis, 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; 7. l., inferior lobe; mam., mammillary body; m. l. dorsal sac, median and lateral portions of dorsal sac; op. n., optic nerve. vitam. A number of fine impregnations were procured by inject- ing methylen blue through the ventral aorta just anterior to the heart. Fig. 33 shows a drawing made from such a prep- aration 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 ditferentiation into axones and dendrites (figs. 34 and 35). All the cell processes look quite similar, as is the case in an intes- tinal 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 habenula Drentvicle Fia. 31. Approximately transverse section of a Golgi preparation of young Amia about 12 mm. long, to show the details of a neurone of the pinealstalk. 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 inner- vation 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 de- seribed 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 sympa- thetic 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 sym- pathetic 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 for- ward 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 ; Sulcus olfactorius i ; forebrain ventricle nh beer Kia. 32. Reconstruction from ten camera lucida drawings of consecutive sec- tions 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 termin- alis. 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 com- pact bundle of six or eight fibers at the level of the posterior ven- tral margin of the olfactory bulbs until they scattered beneath the paraphysis (fig. 80). Finally in three out of four Golgi prepa- rations 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 maxi- mum number does not exceed twelve in adult Amia. Almost invariably when the conditions have been favorable for impregna- tion 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 intra-cran- ial 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 satis- fied 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). Thenature 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 (hes 10> rand 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 verte- brates 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 com- bined 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 maxil- lary 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 Ameturus 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 LO7 sequently, only a few points will be mentioned here. In a speci- men 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 Lepi- dosteus. In the carp (Cyprinus carpio) there were found about three hundred ganglion cells scattered along a more or less dis- tinct 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 prepara- tion 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. 8. Jackson, the develop- ment of a ganglionated nerve in two species of Ameiurus has been worked out in detail. The paper will soon be ready for publica- tion and will show that there is the closest similarity in the devel- opment of the olfactory nerve and the nervus terminalis in Ameiu- rus 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 con- dition 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 gang- lion 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 through- out the fishes. We have quoted Locy (’03) and Pinkus (705) 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 vomero- nasale) 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 physiologic- ally, 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 ventro-median, 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 Amia 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 de- scribed by Blaue (’84) in fishes and amphibians have been shown by subsequent authors to be collections of ordinary olfactory epithe- lium. 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 unspe- cialized sensibility similar to that which Parker (’08) or Sheldon NERVUS TERMINALIS IN AMIA at (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 develop as an invagination from the ectoderm. The last suggestions receive but little support from the embryo- logical development of the nervus terminalis. Both general cutan- eous and some unspecialized visceral ganglia are developed from neural crest (Landacre, ’08 and later unpublished observations), while the epibranchial and dorso-lateral placodes afte 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 placcdes 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 already been referred to; viz: that its fibers are of sympathetic (visceral) type, probably vaso-motor. The embryological evi- dence 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 sympa- thetic neurones originate from neural crest (Jones, 705). We have mentioned that Carpenter (’06) found an apparent excep- 12 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 men- tions the blood vessels. The cells in the periphery are many times more numerous than the fibers that were found entering the pros- encephaion. 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 ner- vus 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 fore- brain 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 through- out the vertebrate series. Aichel (95) cites a number of authors who have found fibers in the nasal capsules differing from olfac- tory 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 ter- minalis, 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 ganglion- ated 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 terminalis in Amia originates in common with the olfactory nerve, from an ectodermal placode. 3. In early stages the cells of the ganglion cannot be distin- guished from the undifferentiated mass of cells which produce sheath cells of the fila olfactoria. 4. Incidentally, we have confirmed the results of recent inves- tigators who find the olfactory neurones of the first order arising in the ectoderm and remaining there in the adult. 2 114 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 lhe in the dorsal suleus between the two rami of the olfactory nerve in the adult nasal cap- sules. 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 prosen- cephalon 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 anintra-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 the cranial cavity dorsally opposite the anterior commissure. Some of these fibers are medullated and supply the paraphysis tubes. NERVUS TERMINALIS IN AMIA 5 15. Ganglion cells are found along this ramusof 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 habenulz, 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, Orro. Kurtz Mittheilung iiber den histolog. Bauder Riechschleimhaut 1895 der embryonalen Teleostier. Sitz. der Gesell. f. Morph. u. Physiol. zu Miinchen. p. 73, Bd. 11. Auuis, E. P. The cranial muscles and cranial and first spinal nerves in Amia 1897 calva. Jour. Morph., vol. 12, no. 3. Bearp, 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.« BeprorpD, EK. A. The early history of the olfactory nerve in swine. Jour. Comp. 1904 Neur. and Psych., vol. 14, p. 390. 116 CHARLES BROOKOVER Brine AND BurckHARDT. Das Zentralnervensystem von Ceratodus forsteri. Anat. 1904 Anz., Bd. 25, no. 23, p. 588. BuavB, Jutius. Untersuchungen iiber den Bau der Nasenschleimhaut bei Fischen 1884 und Amphibien, namentlich itiber Endknospen als Endapparate des Nervus olfactorius. Archivf. Anat.[u. Physiol.] Hft. 3 u.4, p. 231. Brookover, Cuas. 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 dela Belgique, Tome 36. Casa, 8. R. y Une Methode simple pour la Coloration Elective du Réticulum 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. Zodl., Harvard, vol. 48, no. 2. Dean, BasHForD. On the larval development of Amia calva. Zodél. Jahrb. Abt. 1897 f. Syst., Bd. 9, p. 639. DeVries, Ernst. Over het Ganglion Vomeronasale. Koninklij7. Akad. Weten- 1905 schappen te Amsterdam, 11 May. Dissz, J. Ueber die erste Entwickelung des Riechnerven. Gesell. z. Beford. 1896 d. ges. Naturw. Marburger, Sitzungb., Oct. Herrick, C. Jupson. The nervus terminalis (nerve of Pinkus) in the frog. 1909 Jour. Comp. Neur. and Psych., vol. 19, no. 2, p. 175. His, W., Jr. Uber die Entwickelung der Riechlappens und des Riechganglions 1889 und iiber diejenige der verlingerten Marks. Verhandl. d. Anat. Gesell. HovusmrR, GitperRT L. The neurones and supporting elements of the brain of a 1901 selachian. Jour. Comp. Neur., vol. 10, p. 65. Huser, G.Caru. Observations on the innervation of the intra-cranial vessels. 1899 Jour. Comp. Neur., vol. 9. p. 1. Jounston, J. B. The brain of Acipenser. Zodél. 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 nervoussystem 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. Jonrs, WALTER C. Notes on the development of the sympathetic nervous sys- 1905 tem in the common toad. Jowr. Comp. Neur. and Psych., vol. 15. pe Lis. Kapprrs, 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. Kinessury, B. F. The structure and morphology of the medulla oblongata in 1897 Fishes. Jour. Comp. Neur., vol. 7, p. 1. 1897a The encephalic evaginations in ganoids. (Jour. Comp. Neur., Vol: 7; Dp. 3%. ~ NERVUS TERMINALIS IN AMIA 117 LanpacreE, F. L. The epibranchial placodes of Ameiurus. Ohio Naturalist, 1908 vol. 8, p. 251. Leypic, F. Zirbel und Jacobson’sche Organ einiger Reptilien. Arch. f. Mikr. 1897 Anat., Bd. 50, p. 402. Locy, W. A. New facts regarding the development of the olfactory nerve. Anat. 1899 Anz., Bd. 16, p. 273. 1903 A new cranial nerve in fishes. Mark Anniversary Volume. 1905 Onanewly recognized nerve connected with the forebrain of selachians Anat. Anz., Bd. 26. Meek, WatTer J. Astudy 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. Thereactions of the vertebrate embryo to stimulation and the 1907 associated changes in the nervous system. Mitt. Zool. Stat. zu Neapel, Bd, 18, p, 535. Pinxus, Feuix. 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. Puart, Jutia B. A contribution to the morphology of the vertebrate head based 1891 on astudy of Acanthias vulgaris. Jour. Morph., vol. 5, p. 79. Pirrer, A. Die Flimmerbewegung. Frgeb. d. Physiol., vol. 2, p. 66. 1903 Reap, Erriz A. A contribution to the knowledge of the olfactory apparatus in 1908 dog, cat and man. Amer. Jour. Anat., vol. 7, May. RerGHarpD, Jacos. The natural history of Amia calva Linneus. Mark Anniver- 1903 sary Volume, p. 59. RusascuKin, W. Ueber die Beziehung des Nervus Trigemimus zur Riech- 1902 schleimhaut. Anat. Anz., Bd. 22, p. 407. 1903 Zur Morphologie des Vorderhirnes der Amphibien. Arch. f. Mikros. Anat., Bd. 62, p. 207. Scuuttzre, Max. Untersuchungen iiber den Bau der Nasenschleimhaut, nam- 1862 entlich die Structur und Endigungsweise der Geruchsnerven bei dem Menschen und den Wirbelthieren. Abhandl. d. Naturf. Gesell. zu Halle, Bd, 7. Sewertzorr, 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) ps 19i. 1909a The reactions of the dogfish to chemical stimuli. Jour. Comp. Neur. and Psych., vol. 19, p. 273. SHELDON, R. E. anp Brooxover, Cuas. The nervus terminalis in teleosts. 1909 Anat. Record, vol. 2, no. 4. 118 CHARLES BROOKOVER Von Kuprrer, C. Studien zur vergleichenden Entwickelungengeschichte des 1893 Kopfes der Kranioten. Heft 1, Munich and Leipzig. Witson, J. G. The relation of the motor endings on the muscle of the frog to 1904 neighboring structures. Jour. Comp. Neur. and Psych., vol. 14, Deas DESCRIPTION OF FIGURES Fia. 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. Fic. 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. Fra. 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 TERMINALIS IN AMIA CHARLES BROOKOVER PLATE | 33 —_——— ee &F—w oe wo” THE JOURNAL OF COMPARATIVE NEUROLOGY AND PsycHOLOGY—VOL. 20, No. 2. 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 discus- sed 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 spe- cial device for preventing evaporation during dissection was used. The percentage of water applies therefore to the nerve struct- ures 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 Donald- son ’08, p. 346). Each brain or cord was placed in a small glass- stoppered 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 sulphurie 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 correla- tion 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 ti ata PERCENTAGE OF WATER 20 BRAIN AND SPINAL CORD ALBINO RAT ACCORDING TO BODY WEIGHT MALE FEMALE - — -- TT 5. SS : FS SSS SS Ee SPINAL CORD \ Sr ee z —. aL 10 20 30 40 50 60 70 80 90 100 10 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 320 330 BODY WEIGHT Gms. Cart 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. PERCENTAGE OF WATER ia 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 —_ tnerements. PERCENTAGE OF WATER BODY WEIGHT BRAIN SPINAL CORD | i eee it | 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 FEO | 74.4 73.2 65 | 78.9 428 73.2 72.8 75 78.9 78.6 73.2 732 85 | (ee 78.5 72.5 72.5 95 | 78.8 78.5 721 72.0 105 78.4 7302 ile, 71.5 115 rae | 783 hile ve | 71.5 125 78.6 | 78.3 71.4 | vile 135 | 78.4 | 73.2 m8 70.6 145 | 78.2 | 78.0 71.3 70.4 155 | 78.5 78.3 alee 7120 165 78.4 | 78.2 70.6 71.0 175 | 77.8 | 77.5 70.0 71.0 185 78 77.8 HOHE 69.8 195 78.0 Wee | 70.7 68.6 205 78.0 78.0 | 69.6 69.3 215 77.5 ihe | 69.5 68.7 225 78.0 | 69.8 235 78.0 77.0 69.5 | 68.3 245 78.0 | Ties 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’o8, pp. 156-157)» the individual records are not printed. These however are on file and copies of them may be had by application to the Director of the Wistar Institute. 22 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 con- trasted, 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 per- centage 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 asin the brain. The observations are given in table 1 and in chart 1. Percentage of water in relation to age. To support the sugges- tion that the males in the foregoing tables show a greater per- centage 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 per- centage of water in the different age groups as follows: AGE PERCENTAGES OVO Gaye ce. ccolina sce .. total range SUMS eeene. cr Selo a Soe ge 86-89 LO DUGAY Brace. ene 6 aes total range >, UMS eae ot I tek Ee ee 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 :— PER CENT PER CENT OF WATER OF WATER eawy. 4 Drainsy cn dey eens a 78.651 M2 TASOR ec hui. danos Lea COrds 1 Bed re Of Wc ty Peon a eae 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 individ- ual 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 0.86 percent. 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.? PERCENTAGE OF WATER ] AGE IN DAYS BRAIN SPINAL CORD Male Female | Male Female Or) 8 8s 4 SON es eee We Rs 10-20 Roar 83.4 80.5 | 80.3 20-30 | 81.3 81.6 | V7 2 | rye 30-40 | 79.4 80.0 | 74.3 74.8 40-50 | 79.2 79.0 | 73.9 rer 50-60 79.0 79.3 72.9 74.2 60-70 | 79.3 78.8 74.5 Toe. 70-80 | 78.9 | 78.8 72.9 1208 80-90 Hoes | Toe 72.8 73.8 90-100 78a | 79.0 73.0 74.1 100-110 | 73.3 | 78.0 | 70.0 70.8 110-120 | 78.6 | Weer | WA 1225 120-130 | 78.3 | FD 1.6 wet 130-140 | TB oe 78.0 70.0 71.0 140-150 | 78.0 72.0 150-160 | 7k 78.0 70.6 70.8 160-170 78.2 78.3 71.0 lees 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 (Re Wa) 78.3 71.0 ARI 220-230 | 7307 W303 (ee 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. 7 o as [= of ad 2 _ = a as ~~ —" — = i. ’ = om a - 7 S = ~ oa 7 art - - 7 = Sa ' 7 : = ® yO - = a : A . - G yy 0} = = a % ~ 3 a 4 7 i is ‘ j ‘ av mi ‘ Si se = we. = ‘ eee ae oF CG r - wee an \ 1 rca i an : mn 90; PERCENTAGE OF WATER PERCENTAGE OF WATER 90 és BRAIN AND SPINAL CORD ALBINO RAT pare ACCORDING TO AGE MALE ——— FEMALES ° 10 20 30 40 50 60 70 80 90 100 10 720 130 140 150 160 170 180 190 200 210 220 230 240 DAYS AGE Cuart 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. 163- 164). The effect of this lengthening would be to diminish the percentage of water. The influence of passive lengthening is dis- cussed 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 rela- tions 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 Vans —— 12-6 low (a -- S-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 e949) — 12.8) lox (@) [3] and from a body weight of 35 grams on Sor = o-0 love @) [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, 709; 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 there- fore 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 | 85.7 10 | $5.26 82.10 15 | $3.50 84.8 79.80 81.4 20 82.24 78.26 | 25 | Seay 81.2 76.98 | rama) 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 esi 75 78.99 | 78.9 73.01 13.2 85 78.87 78.8 72.66 (2.9 95 78.76 78.8 712.34 Toll 105 78.66 78.4 72.06 Gle5 115 78.58 sh 7 71.81 71.7 125 78.50 78.6 | (Ol ay/ 71.4 135 78.43 78.4 | 71.35 | mes 145 78.36 78.2 NS (Ale3} 155 78.30 78.5 70.96 ((Ne?- 165 78.24 78.4 70.79 | 70.6 175 78.19 77.8 70.62 70.0 185 78.13 me 70.46 70.7 195 78.09 78.0 70.31 70.7 205 78 .04 78.0 CORT 69.6 215 78.00 es 70.04 69.5 225 77.96 78.0 69.91 69.8 235 17.92 78.0 | 69.79. 69.5 245 77.88 78.0 69.67 70.0 255 77.84 78.0 69.56 70.0 265 77.81 Tht ee 69.45 69.2 275 Cate 78.3 69.34 69.5 285 U1 .04 77.0 69.24 68.3 295 etal 78.0 69.15 70.0 305 77.68 ies 69.05 69.0 315 77.65 Tf 3) 68.96 68.5 320 77.62 78.0 68.87 68.0 470 : “I “I “ oO | & fy) oo Oo 90 PERCENTAGE OF WATER PERCENTAGE OF WATER 90 = BRAIN AND SPINAL CORD ALBINO RAT 86 ACCORDING TO BODY WEIGHT ma % MALES ONLY. 82 5 Ni ge e Oe See Tae ry -—_e—,—_ + _o_#__, L 78 ——————— ae e e 76 76 e THEORETIC CURVES es "2 la rae eee SPINAL CORD =. e e a Nl 1 je eee (Se ee o 10 20 30 40 50 60 70 80 90 100 0 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 320 330 Birth BODY WEIGHT Gms. CuartT 3 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 Lag 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 animais had both heavier brains and heavier cords. He noted also that the mated rats, as compared with the unmated of like age, had the follow- ing percentages of water: NO. OF PERCENTAGE OF WATER CASES Brain Cord Hemalemnn ated cnn seese 6 at ater : era cl ae) (8) hse 68.51 Remalessunmiabe dy ir eee. srcteies : hse tae keen (10) 0.30 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 ner- vous system of the mated rats must be regarded as physiologi- cally younger. Such slight differences would, of course, not be worthy of remark if they had been obtained merely by the averag- ing 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 com- parisons, 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. ENCEPHALON SPINAL CORD PER CENT PER CEN1 Maletcontrols.. eter ns. Soo. s.. wee ke ne eee ae ties 2 ete 77.50 69.71 Malevexperimentéeds 305.5555 isonet ies eee dies 77.75 70.05 Hemale COmErOI sco ae csion eke NA chee 5 eee TROT 77.50 69.40 Bemale experimented |) crisuy eset cede sioe oil tee ete UU sth 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 ae 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 AGE NO. OF CASES Brain Cord IVCER Si doutecsctomec 7 stunted 78.618 72.613 IVAUCEIN IS aes cee woods 7 controls 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 mo- . ment 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 deter- mined that in rats killed at the end of three weeks of underfeeding the experimented rats, though initially heavier, had on the aver- age 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 experi- ment, 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 corre- lated 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 experi- mented 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 preceding descriptions 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 (3f 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 abso- lute 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 AMOUNT or WATER BRAIN BRAIN Seb WR) SPINAL CORD WEIGHT | y WEIGHT Male Female Male Female grams | grams grams grams grams grams 0.25 0. 208* Onliia- 0.03 0.028 0.027 0.35 | 0.325 0.290 0.07 0.067 0.062 0.45 0.350 0.400 Weil 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 OFS5 7 | 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 0.275 120) | 1.025 1.012 0.43 0.308 0.308 ikers) 1.088 1.067 0.47 0.340 0.318 Abe | 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 1275 1.359 1.355 0.63 0.434 0.430 1.85 1.450 1.444 0.67 0.465 0.450 IOS) 1.530 1.520 OR 0.473 0.470 2.05 | 1.636 1.550 0.75 0.520 | Zealot 1.650 sere centages still more irregular. However, a second series of calcu- lations 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. a2 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 incre- ments of weight afterabrain weight of0.35grams. 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 grams. 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 iss increasing brain weight; a fact already demonstrated by the previous tables and charts. WEIGHT OF WATER CONSTANT PERCENTAGE ‘7 GMs. BRAIN es is WEIGHT OF WATER a . 13 BRAIN : Le ALBINO RAT a iva 4 ACCORDING TO BRAIN WEIGHT a Mugo 1.3 MALE FEMALE ———— ues ue BRAIN WEIGHT GMS. | | | BODY WEIGHT GMS. 7 10 12 17 25 46 90 200 400 AGE IN DAYS 4 8 10 15 25 43 64 MALE CHartT 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 occur- ring 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, hav- ing 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 CONSTANT PERCENTAGE 6. GMS. SPINAL CORD «| WEIGHT OF WATER. al ve SPINAL CORD y ALBINO RAT e ye ACCORDING TO ea ee CORD WEIGHT = Ye MALE——-_ [FEMALE ———— ; Ze 3 a 2 , E Jee 1 ; ity y ae Z SPINAL CORD WEIGHT a 1 2 3 4 5 6 7 Gms. 8 | 1 | ] ! ! ] | BODY Birth 15 28 45 70 105 150 225 305 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 follcwed by the percentage of water through the life cycle, and also 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 un- medullated, 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 rela- tive amount of white matter in the brain is much less than in the spinal cord (Donaldson and Davis ’08; 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 per- centage 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 constit- uents. 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 diminish- ing percentage of water. Such a statement however is purely formal. What probably takes place is this: Starting with a given per- centage 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 con- stant weight of water for the same increment of total weight. In this connection the analysis of the brain and cord should how- ever 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 inthecord. 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 per- centage 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 after 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 (Boughton ’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”’ conclu- sion 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 sub- stance, 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 myelination, 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 sub- stances 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 resultsdo not answer the question whether the forma- tion 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 concern- ing the relatively small reduction in the gray substance that the percentage of water in the cell bodies in this case is not progress- ing 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.’ From this it follows that the greater proportion of gray sub- stance 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 length- ening 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 oply 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 medul- lated as soon asits 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 in- creases 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 diminu- tion 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 conclud- ing that in the percentage of water we have a character corre- lated 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 asso- ciated with the span of life or longevity of the animal concerned. Broadly speaking, the changes in the percentage of water indi- cate 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 per- centage of water in the brain. In connection with such a compari- son, I have examined the entire literature on the percentage of water in the nervous system. ‘This literature needs to’ be sum- marized, 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 (cal- losum); (2) gray substance (corpus striatum); (3) gyrus (white PERCENTAGE OF WATER 141 and gray mixed); (4) cerebellum; (5) pons; (6) medulla oblon- gata. 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 tests 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 deter- mination for the water in the entire encephalon at maturity.’ 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 determina- tions by that author alone. 142 HENRY H. DONALDSON value for the percentage of water in the encephalon at two years 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 Percentage of Age Years Water Water Days Birth 88.3 87.7 Birth 2 years 81.1 81.3 26 days 9.5 years 19.2 78.6 115 days 25 years maturity ThA) 78.0 290 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 isto 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 correspond- ing 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. _Remember- ing 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 inten- sity of the growth process characteristic for each species. (Rub- ner 708 and ’08a). CONCLUSIONS 1. In the albino rat between birth and maturity, the per- centage 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 matur- ing 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 forma- tion of the medullary sheaths, is not a controlling factor in reduc- ing 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 ment 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 Boucuton T.H. The increase in the number and size of the medullated fibers 1906 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. Donaupson, 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. Donaupson, 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. Donaupson, H. H. On the relation of the body length to the body weight and to 1909 the weight of the brain and of tnespinal cord in the albino rat (Mus norvegicus var. albus). Journ. Comp. Neurol., vol. 19, no. 2, pp. 155-167. Donautpson, H. H. ann 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. Hara, 8. The effect of partial starvation on the brain of the white rat. Am. 1904 J. of Physiol., vol. 12, no. 1, pp. 116-127. Hara, 8. 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. Harar, 8. 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. Harat, 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. 169- 173. Kocu, W. AND Mann, 8. A. A chemical study of the brain in healthy and diseased 1909 conditions, with especial reference to dementia precox. Mott’s Archives of Neurol. and Psychiatry, vol. 4, pp. 201-204. RuBNER, Max. Probleme des Wachstums und der Lebensdauer. Wiener 1908 Medizinischen Wochenschrift, nos. 11-13. RvusBNerR, Max. Das Wachstumsproblem und die Lebensdauer des Menschen 1908a und einiger Saiugetiere vom energetischen Standpunkt aus be- trachtet. Archiv fiir 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. WeispacH A. Der Wassergehalt des Gehirns nach Alter, Geschlecht und Krank- 1868 heiten.. Med. Jahrbiicher, vol. 16, nos. 4 and 4, pp. 1-76. PLEASURE, PAIN AND THE BEGINNINGS OF INTELLIGENCE 8S. 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 ge- netic 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. Wundt in his well known tridimensional theory of feeling, which has secured a small fol- lowing, 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 ex- periments 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 re- sults which they regard as clearly at variance with Wundt’s doctrines. There is no agreement among psychologists asregards the phys- iological expression of the pleasant and unpleasant. TFéré, Leh- 156 S. J. HOLMES mann, Mentz, Zoneff and Meumann find that pleasant and dis- agreeable states are quite uniformly accompanied by certain characteristic physiological processes. Among the accompani- ments 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 de- crease 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 applica- tion. 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 dif- ferent 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 supply to the brain, and unpleas- ant sensations by the reverse effect. Angell and Thompson express themselves as 1utlows concerning the physiological concomitants of pleasant and unpleasant re- actions: 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, posi- tion 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 diguetes: able, produced vaso-constrictions. The results yielded by the study of affective states, by means of instruments for recording changes in circulation, pulse, and respir- ation 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 asso- ciated 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 prob- lems 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 andof pleasantness and unpleasant- ness 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 they 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 unneces- sary 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 attemptsat movements that really inhibit the taking the medicine? This is evidenced by the chick described by Lloyd Morgan, which after an experience witha nasty 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 particu- lar 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 con- cerned 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 con- currently 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 coin- mon. Similarly with certain reflex-ares arising from the skin of the pinna of the crossed ear. In them excitation reinforces that of the just men- tioned ares from the fore foot and opposite hind foot. The presence of savory food in a dog’s mouth causes the secre- tion of saliva and the movements of chewing and swallowing, and the stomach at the same time may be stimulated to secrete gas- tric juice. These activities are organically associated and they are usually preceded by seizing acts of various kinds. A particu- lar 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. ‘ 160° 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 move- ments 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 ex- ploring movements of the large chelipeds. The coédrdinated movements of the antennules, mouth parts and legs may beregarded as a complicated form of chain reflex. If now we apply a stim- ulus 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 ex- ploring movement of the limb. This is followed by similar move- ments of the other chelipeds and chewing movements of the mouth parts, and these by the twitching of the outer ramus of the anten- nules. 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 move- ments 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 crayfish 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 dan- ger, 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. wr __TIME_ a 13° ce : SERIES | | F | TIMES ean et aces si is Soe Fic. 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 withineachsquare. Inevery 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 muscula- ture, or the neuro-muscular codrdinations, 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 es- tablished, 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 com- plicated 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 apremium. In many instances the animals became increasingly nervous, and lost their heads. In problems whose solution can be effected, provided the ani- mal has enough time and makes sufficient movements, this com- plication 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 be- come so handicapped by fear, and the useless activity charac- teristic 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 LAO 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 intelligenee 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 eorrect 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 dis- turbs 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 path- way, 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 meas- urements, such an individual may seem to improve or to hold its own; if, on the other hand, we study the pathway alone, the ani- mal 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 wel As the objective records are at variance, it seems at first sight arbi- trary to fix on either one or the other, and to say this is a meas- ure 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 indi- viduals 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 direc- tion by superiority in another. The latter might even win, but if we were offering prizes, justice would demand one for improve- ment 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, though 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 whirl- ing movement. There was but one corner opening through which escape might be effected. The problem was thus essen- tially 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. We OTTO C. GLASER ANIMAL I. ANIMALS II. ANIMAL III. Trials | Time Intervals ‘Trials Time | Intervals | Trials | Time | Intervals 192072 0 ine Uae 850) 0 1 6.0") “130” QF 3 oe 0 a a a(t) N) 2 910°C) 110” 3) |) S2eOuar een 3 9.0” 1 3 8.4 10g 4 | 5.0" | 15! 4 | 5.07 | 0 4 | 26.07| 2/03” 5 Seo ae 0 BN 20) ist a ls Be ON ie 00 6 2a Bont 220% 0 (aor) e074" Hours RPT Oe i 17) e382" | | 8 122 le GAO's | PT GUAGE eee 2 uy | | 10 DEA 0 a0" | 12 0'4) Wome 12 1.0" Characterization, Animal I: At first many trial movements, then stopped inves- tigating. 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 IT: Accidental delay at second trial. Numerous extra movements infourth. Inseventh 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 III: 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 Diese e804 1 55 oY 30’ 1 fol 30” 2 S277 a0” 2 aie 30” <2 [See 207 3 gio 30” 3 Bale” 30” 3 430) 30” 4 LOSE ee 30" 4 6.37 30” 4 Fal” 30” 5 Beall. 300 5 Eye 30” 5 aes 30” 6 Seren: 20"! 6 apes 20” 6 Bhs 20’ 7 eae we 3) 7 ape 30” 7 LOR 30” 8 AA E30" 8 6.0” 30” 8 ea he 30” 9 Mae 30” 9 1222 30” 9 DOL StE 30” 10 a2 30” 10 Lon oY 30’ 10, es 30” 11 oS es betas 11 4.0” 30” Li era 30” 12 BuO! ee 12H aeons 12 | Asst! 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 estab- lished. 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 in- volve 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 ean construct the following table for comparisons: FIRST HALF RECORD SECOND HALF RECORD Averages Averages Seconds Seconds JNTOUTOTEY [te Ena on A oe ea 8: GRE serie those 3.66 ANGI UTOTNEET lAti Ll Le eae aay rene Pe ee ae Ld SOM ee ere eat hs aun Shoe PNooiie ng | By UY age aed Race, eae oe ee A. Oe eI ne eek tt Fs gee 1.60 PANTATIINS ee AVS EI Ss GRO? ea cies, ocks a x «al ON COREE SORE a ieee ee 4.88 PATNA EV ener or sek nie iac nck cs 5 LOO ee eet eres oe 8.11 Genenaleaveragerree 5-55... 11.23 General average..... 4.32 Aglance 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 under- went the experience of the tank. The individuals whose perfor- mances 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 8) 30)" Dl? 1 +5) (0) ee 2 Ane AA’! 2 13” Be 3 NO IQ)” 3 14/7 ale 4 Meta AY 4 10’’ De 5 5.0” 280 9) Ao! BUge 6 PO 45”" 6 Oleg ile a 0 apy a GO” oe 8 Loa ABU 8 5)“ aA ee 9 oe NOM 9 Ae She 10 0” alts) 10 2.07” 58” Ul Oia 56” iil 5 by Soi 12 2.0" 12 620% 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 £75 FIRST HALF RECORD SECOND HALF RECORD Averages Averages Seconds Seconds PATiari ct lgeeA Gas ee eet ee ae ent S.A ID Rees eee are ot NS Spears 1.80 “Amiga 18) Sls Se.0 he Ee ere SS DOR: ae es avs. ws REA CIs 4.75 JNA: KOb BS Owe se ee he OS SE OD eae 2 Bs 2) ee Seca ee 4.80 Avene SID) sc! ate se teks 5 ee GRAS ons, Sree Go ss Be 4.54 General average................. 8.338 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 Soe 42” 1 10.3 | 24”’ 2 33).0)" Do ae 2 Sno | 3h) 3 ROM 20% 3 Sate ) ibailoy? 4 SRO” 34” + Does 40’’ 5 9.0” 36” D 6.0"" 47" 6 4.0/" ee 6 20M 56 7 a0)" iL” a Om 40’’ 8 (0) ee Le (0) ee 9 SROm 43" 9 Seow et) 10 DSi ee 10 6.0” 58’” 11 AO 282 | 11 Ate | 42” 12 10 2 Deo Characterization, Animal C: Animal between three 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 nosemblance 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. Rela- tively 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 hovrs, 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 dis- tinguish the three young animals by small letters, the six adults by means of capitals. ANIMAL DATE TIME TRIALS AVERAGES Seconds Ce || eS 4/26 2:15 pan. Seen 4.58 Oe? 6 Aas oe 4/26 3728ip.m., “|. soma 1.34 a aos 2 rr 4/27 2 15pm. |. one 4.69 lage et: ee GR ae 4/26 2230spsmie, | eG 5.15 Dew Aas FAP Retry 4/26 3235ip.m..< wr AG 4.75 ORM <6 Sa 4/27 PSU oniy le 8 3.30 Dy ON ace hs 5/9 10.200am. | 6 5.65 a2 rpgk at RENE RN Noes, 5/9 10.45 a.m. 6 2.10 i eect SN» A PE 4/26 2.50pm." 116 13.40 CR Oe er eA 4/26 3.43 p.m. | 6 4.79 CO ee ocd eee a 4/27 2.50 p.m. 6 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 experi- ments 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, ean, however, be safely compared. THE FORMATION OF HABITS AT HIGH SPEED Leer | DATE DATE DATE | DATE ANIMAL | | | Jan. 15 Jan. 20 Jan. 22 Jan. 28 Seconds Seconds Seconds Seconds WO Fas co ot 10,40 1.66 Heol le 2208 1B eS nets Aaatone Raateeee eee eo eee 6.90 10.25 5.83 6.00 (CHa ne ence cleissks 8 yest eke 5.90 9.08 [3275 10.50 DL corre teens ARC ote Rane eae 6.41 | 5.66 6.08 6.08 Be terre. ater es 6:08) |) \e32082 11.41 9.75 Dee Loco | 19.00 16.00 10.16 General average........... 8.58 13.08 9.16 W233 The results from the young animals are not especially conclu- sive. Taking the records as they stand, however, it may be said that two out of the three seemed to show the effects of their expe- riences twenty-four hours afterward, and one of the animals, after twelve days, very quickly recovered and in the end actually bet- tered 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 further- more, 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 ¢ 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 de- tailed 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. GLASER . INTRODUCTION OF ADDITIONAL STIMULI The ‘“‘tank experience’? may be complicated by the introduc- tion 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 consider- ably 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 gotten 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 SAND 2 apa Reed eld Mee 6 7.45 6 Syl dl BBA ote eee: 3 eed 3 8.2 48 VANES etc toca 6 10.8 6 32 69 1510) ARR ee tee 6 13.03 6 5.26 60 ce 6 9.8 6 faa" 26 dds ero ne 6 16.65 6 12.66 24 CO eo, 14 Senor ye 6 13.3 6 7.2 46 ffi. & > ¢ Me dian Antavror Cerebral AiTery i ap ¢j or. 2p Mt 9 mmm rrr gt L CCR = UC J 3p TEER SCIPS ere Pavace b> tee nora 2cun s To ats Ve | lobe RQ, lobule prec s fy N Qo 7 Tm ©To : sib He q rots gs i ETT “ To Y, Dy a, Fig. 5. Diagram of the Circle of Willis with a typical median anterior cere- bral 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 Griinbaum and Sherrington, quoted from above, was to call attention to the surgical relations of the ine- qualities 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 intra- cranial 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 obstruc- tion of the anterior communicating artery and the left posterior cerebral artery. Fic. 6. Photograph of the base of the brain showing the origin of a large med- ian 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 Willisti 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 direc- tion of reversion to the primitive type, and that it is reasonable that the frequently found arteria termatica is one of these rever- sions or survivals such as are not uncommon in other organs. Accepted by the Wistar Institute of Anatomy and Biology, March 24, 1910. Printed July 7, 1910. CONTRIBUTIONS FROM THE ZOOLOGICAL LABORATORY OF THE MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE. EK. L. MARK, Drtrector. No. 208. 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, Cam- barus 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 altera- tions in the roots of the severed nerve were considerable. In long-standing cases of amputations, these authors reported altera- tions of medullated fibers, and a reduction in the size of the bun- dles of fibers. These observations led to experiments on animals for more direct evidence on the questions involved. Homén (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 Walle- rian degeneration. Marinesco (’92) found that on cutting the spi- nal 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 nor- mal animals. He also claimed that the remaining cells showed more or less atrophy. In experiments in which the ganglion nodo- sum 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 thedisappearance 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 experi- ments 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 condi- tion, since he found more cells at the end of four months than were seen at the end of ten days. Van Gehuchten (’03) found degenera- tion taking place centrally and declared it to be a true degenera- tion, 7.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 cer- vical 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 most convincing evidence 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 rea- sonably 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 cer- tainty, 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 pos- terior to each anterior group of cells, though not distinctly sepa- rated 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 le 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 gan- glion 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 num- ber 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 sec- tions 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 fibril- le. These fibrille 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 gan- glion cells following the cutting of their fibers could be readily observed. It was not known how well the animals could endure the in- juries 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 nor- mal 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 sternites. 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 tis- sues, and the ganglion was thus isolated as regards its nerve con- nections. The animals were then numbered for purposes of identi- fication, 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-platino- osmic 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 10u. 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). Fibrillee in the axis cylinder were usually visible, though not all the cells revealed them. That they were not always seen, is be- lieved to be due to this method, which, as is well known, is not a wholly satisfactory one for demonstrating fibrille. The Nissl ‘‘flakes”’ were present, though somewhat less distinct than in the ganglion cells of vertebrated animals. In some cases there ap- peared, 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 exhi- bited 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 con- trasted the conditions found in the corresponding ganglion cells from a fifth ganglion (fig. 2) which had been nervously isolated about twelve days, but otherwise hadreceived identical treatment. The cells from the isolated ganglia showed distinct alterations. The nerve fibrillee 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, many 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 quiteevident. Insomecases 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 1s regarded as strong evidence of true degeneration. On examining the results obtained in the first series of experi- ments, it became evident that, although degeneration could be demonstrated, the Niss] 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 structures of these ganglion cells in their normal condition, it was thought that this treatment might also show the changes occurring in connec- tion with the loss of function and consequent degeneration. Ac- cordingly 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# 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 differen- tiated in black and white. The fibrille (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 sur- rounding the nucleus. Numerous flake-like bodies, apparently having some relation to the Nissl bodies, le 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. Thenucleusshows the charac- teristic 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 fibrille (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 cellfrom ganglion isolated twelve days; nucleus collapsing, axis-cylinder area reduced, fibrille indistinct. 3. Large ganglion cell showing normal condition of axis cylinder, fibrillz, nu- cleus, and ‘‘flakes;’’ vom Rath method. 4. Large ganglion cell from ganglion isolated twelve days; tortuous and nodular fibrille. Compare with figs. 2 and 3. 5. Ganglion cell from ganglion isolated fifteen days; nucleus and axiscylinder reduced, fibrille 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 dis- integrating (seen in only a few cells), cytoplasm faintly granular. 9. Ganglion cell from ganglion isolated twenty-nine days; nucleolus has dis- appeared; cell-contents faintly granular and apparently very fluid. GANGLION CELLS OF THE CRAYFISH Hawnsrorp MacCurpy PLATE I THE JOURNAL OF COMPARATIVE NEUROLOGY AND PSYCHOLOGY—VOL. 20 No. 3 ied ; i ' , iia alia nyt os “if we vi e big Ai 4 vi aOR) " : ry ay YF ey) ae 4 Ans Cera) , : “4 ; ’ hb ; | | r ite : a ap) Nei ai j ‘. ed a id Ai Rats 2 | vue) GANGLION CELLS OF THE CRAYFISH 205 the cells were less clearly defined, 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 fibrille 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 partsof 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 demon- strated by the Nissl method and observed in the spinal-ganglion cells of dogs by Lugaro (’87), and in cats, dogs and rabbits by Késter (03). In eighteen to twenty days the fibrille and axis cylinder usually had entirely disappeared (figs. 5, 6), and the con- centric 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 MACCURDY 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 evi- dence 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 milli- meters 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 examina- tion 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 degenera- tion. 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 daysafter the connective had been severed. A smaller number of nerve cells in a similar condition of disinte- gration 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 ganghon 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 gangha which had their connectives severed, and were found in everysuch case. That the changes are those connected with true degeneration finds support in the fact that they resemble in their essential points the histological changes in the corresponding parts of the nerve fibers and cells of vertebrated animals as described by Moénkeberg 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 fibrille 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 spi- nal ganglion cells. In ten days vacuolar degeneration and shrink- ing 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 fibrille 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 thirty- nine 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 condi- tions 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 gan- glion 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 which pertain to the medullary sheath, which is absent in Cambarus. Accepted by the Wistar Institute of Anatomy and Biology, April 17,1910. Printed July 8, 1910. GANGLION CELLS OF THE CRAYFISH 209 BIBLIOGRAPHY Fiemine, 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 thecord. Edinburgh Med. Jour.,n.s., vol. 1, pp. 174-182, 279-288, plait: FRIEDLANDER, C., und Krauss, F. Ueber Verinderungen der Nerven und des 1886 Riickenmarks nach Amputationen. Fortschr. d. Med., Bd. 4, pp. 749-764, Taf. 7. Homén, E. A. Verinderungen des Nervensystems nach Amputationen. Bevirdge 1890 z. path. Anat. u. allg. Path., Bd. 8, pp. 304-351, Taf. 18-19. Kuerist, K. Experimentell-anatomische Untersuchungen tiber 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 Yaqrra, K. Experimentelle Untersuchungen tiber den Ursprun ? 5] 1905 des N. Vagus. (Quoted from Ranson 1906.) Koster, G. Ueber die verschiedene biologische Werthigkeit der hinteren Wurzel 1905 und der sensiblen peripheren Nerven. Neurol. Centralbl., Bd. 22, pp. 1093-1102. : Krause, F. Ueber aufsteigende und absteigende Nervendegenerationen. Arch. 1887 f. Anat. u. Physiol., Physiol. Abt., Jahrg. 1887, pp. 370-376. Luaaro, E. Sulle alterazigni 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 Verainderungen 1892 der Nerven und des Riickenmarks. Berliner klin. Wochenschr., Bd. 39, p. 988. 1898 Sur les phénoménes de séparation dans les centres nerveux apres la section des nerfs périphériques. Presse med., année, 6, semestre 2, pp. 201-206. MonkeserG, G., und Beruz, A. Die Degeneration der markhaltigen Nerven- 1899 fasern der Wirbelthiere unter hauptsichlicher 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 Rerzius, G. Zur Kenntniss des Nervensystems der Crustaceen. Biol. Unters., 1890 N. F., Bd. 1, pp. 1-50, Taf. 1-14. Van GexucutEN, A. L’anatomie fine de la cellule nerveuse. La Cellule, tome. 1897 13, pp. 312-390, pl. 1. 1903 La dégénérescence dite rétrograde ou degénérescence Wallérienn indirecte. Le Névraze, vol. 5, p. 1. (Quoted from Ranson, 1906.) THE DEVELOPMENT OF THE SYMPATHETIC NERVOUS SYSTEM IN MAMMALS ALBERT KUNTZ From the Laboratories of Animal Biology of the State University of Towa WITH EIGHTEEN FIGURES CONTENTS Fee Loa LOC CUO NNN BORED PSE 2 SUC ct ees aps wn Oe tee ree ea ev iakans «stake ols DP Maisto onll Swi fo oe ae ee aero eet Teer one eee 214 Te. Met hadciota lanes tale a thors sy tee ey Posies ays, 2 ow ngn ce Seles «pais = Seen gee te 218 ANDY, DAG) SstSray se Grivel A een hee fae ur ded a oo Tena Su eh caine ne, lok, 3 eee 219 Pe oympathetic trum tens iar eras ee UR ete eee eon ae 219 qotanly, development it. ecm ee 0 -miad las Weeroe le Sek 219 bene ramigera Geena cees. 1 ee pet tee Rote cel 9 a rs ae ci, «rae 222 @xalaaiter Ceviel Gps aon ak oe pete otek ort coke Srey cert sop eee!) Gd. Nature Of Miprabitip Gellsesermererri. ook tic om: Jeter 230 momPrevertebral plexuses 20). fic catheter ees ig ts oe aid ele eset eee 233 He, MOL NAO) ANE we ae ba go ac ch cho nest mem eoneodeaccu ne Bene 233 b. Histogenetic relatiouships.. 2429 -miis tne ee ht k= = 234 Boe Vara svinpathetic plexuses... see cfa se eas rec is alam PBS) Gia, WUDOC LIOR Aang tac endo nda on ar riea bole subse cs Ane 235 b. Myenteric and submucous plexuses............--....+--- 236 6 mpulmonary: plexuses 4.) . an ete a erie cad cede ts ast 239 d. Cardiac plexus........ Ee a ee Pn ae. 5 ZOU . Cell migration along ines VAIb salar oa eects art: 241 Vi, Discussion of Results. and Conclustons:2 9... 02. os) -e see oe 246 ae Migrationofmedullary cells:... 262 ssusweltacieece ett oe Fe 246 iD. INN ma uballocaaewes soso mee Bae eee oso oS cod odse eo dan tea mere one 248 c. Sympathetic excitatory and sy mpathetic sensory neurones..... 249 d. A wider application of Schaper’s conception........ Rhee ool e. Relation of the sympathetic to the central nervous sy een Oe ke 251 Poe UMCtOMAlTelaiONS.. gece s..:/- «cee eee setae oa oes eto aya Se 252 WR USSIESRIADETEN ope cy A eae one Se Sn es 2 fg a ee Ee ero 253 | SN SUMO T3P2 A Ft Cece ene 2 cg nee IRR re le -cPin ee ai imme Ra gare 256 PA lp ALBERT KUNTZ * I. INTRODUCTION The present investigation of the development of the sympa- thetic nervous system in mammals was carried on in the labora- tories 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 develop- ment of the sympathetic nervous system, our knowledge concern- ing 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 ofthe development of the sympathetic system in 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 important 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-brain 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 inspira- tion 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 valua- ble 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 ele- ments 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 ad- vance toward the aorta, each, however, retaining connection with its respective nerve by a fibrous branch which becomes the com- municating ramus. These ganglionic enlargements are at first independent of each other, but becomeunited later by longitudinal commissures. These observations on the selachians were sub- stantiated 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 devel- opment 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 verte- brates. 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 cell- groups 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 peripherally and become the anlagen of the peripheral sympathetic plexuses, including the sympathetic plexuses in the walls of the digestive tube and the sympathetic componentsrelated 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 iS face of the aorta. These are the beginnings of the primary sympa- thetic 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 spial nerves. The cells giving rise to the primary sym- pathetic trunks migrate thither from the spinal ganglia, along the spinal nerves and the communicating rami. The anlagen of the secondary sympathetic trunks are separated fromthe spinal ganglia only by the fibers of the ventral nerve-roots. Neuroblasts may be found in the ventral nerve-roots, caught apparently in migra- tion from the spinal to the sympathetic ganglia. After the appear- ance of the secondary trunks, the primary sympathetic trunks become resolved into the various ganglia and nerves of the pre- vertebral 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 out- growths 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 sela- chians, but differ very materially from the findings of later obser- vers for amphibians and birds. In his later researches on the urodeles, Hoffmann (’02) found conditions of development differing widely from those in sela- chians. 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 pos- terior 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 com- municating 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 syneytial 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 sympa- thetic 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 histo- genesis 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 (’04) showed experimentally that in amphibians the cells giving rise to the neu- rilemma 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 sym- pathetic nervous system. In his summary he expresses the opin- SYMPATHETIC SYSTEM IN MAMMALS PVE 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 systemasan 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 cell- aggregates 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 migra- tion, 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 rep- tilian type, the sympathetic anlagen arise as short cellular out- growths on the spinal nerves, which early show ganglionic enlargements at their distal ends (Neumayer). In birds the pri- mary 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 cell- aggregates just median to the ventral roots of the spinal nerves (His, Jr., Lillie). Inmammalsthesympathetictrunks ariseasa 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 sym- pathetic 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 iron- - hematoxylin 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 thick- ness 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 hematoxylin 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 inter- est 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-hematoxylin 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 service- able in tracing migrant nervous elements. The silver reduction method was used with good results in the later stages. IV. OBSERVATIONS I. SYMPATHETIC TRUNKS (a) Karly development.—The earliest traces of the sympathetic trunks appear in the thoracic regron 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 com- municating 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 protoplas- mic processes, or are included in small syncytia. These struc- tures 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. Fic. 4. Diagrammatic transverse section of an embryo 12 mm. in length through the suprarenal bodies, < 50. SYMPATHETIC x 210. SYSTEM IN MAMMALS s Fie. 1. municating ramus in an embryo 9 in mm. length, 221 Section showing the path of the com- (See reference letters below. ) REFERENCE LETTERS OF FIGURES (Hacept 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.c.-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. c.m.d.n.r.—Cells migrating into dorsal nerve-root. c.m.pv.—Cells migrating from sympa- thetic trunks into prevertebral plex- uses. c.m.v.r.—Cells migrating into nerve-root. c.m.vag.r.—Cells rootlets. d.n.r.—Dorsal nerve-root. e.l.m.—External limiting membrane. f.pv.Fibers extending into preverte- bral plexuses. g.c.-Germinal cells of His. i.c.c.r.-Indifferent cells in communica- ting ramus. ventral migrating into vagus i.l.m.—Internal limiting membrane. mes.—Mesentery. m.r.f.—Motor root-fibers. m.s.p.-Anlage of myenteric and sub- mucous plexuses. nb.—Neuroblasts. oe.—Gsophagus. o.rec.n.—Origin of recurrent nerve. p.a.—Pulmonary artery. pv.a.-Anlagen of prevertebral plexuses. rec.n.—Recurrent nerve. sp.g--Spinal ganglion. sp.n.—Spinal nerve. s.r.f.-Sensory root-fibers. supr.—Suprarenal bodies. sy.-Anlagen of sympathetic trunks. t.-Trachea. vag.—Vagus trunk.« vag.b.-Branch of vagus nerve. vag.r.—Rootlets of vagus nerve. v.n.r.—Ventral nerve-root. 22? 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 grow- ing 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 Fre. 2. Section showing the path of the communicating ramus in an embryo 10mm. 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, 1-@sCene) 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, e.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 Fic 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, < 275. there is a notable degree of individual variation in the rate of development. (b) Cell migration.—In a recent paper! 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. Ana- tomical Record, vol. 3, no. 8, pp. 458-465. 224 ALBERT KUNTZ L_-- f= 3 -— al Jo spot. deena Hic. 4. Transverse section of the neural tube and the anlage of the sym- pathetic 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 con- ditions. 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 limit- ing 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 nerve- roots enter the neural tube. In transverse sections, numerous breaches may be observed in the external limiting membrane in Fic. 5. Transverse section of the ventral part of the neural tube of an em- bryo 9mm. in length, showing cells migrating into the ventral nerve-root, 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 dor- sal nerve-roots (fig. 3, c.m.d.n.r.). Further evidence for the migration of medullary cells into the dorsal nerve-roots is pre- sented 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,¢.m.v.r.). This fact has recently been observed by Carpenter and Main (’07). I have been able to substantiate Fic. 6. Section showing the path of the communicating ramus in an embryo 7 mm.in length, < 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 dor- sal and the ventral nerve-roots seem to have their origin in more or less distinct regions. In the dorsal region most of the migrat- ing 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 DG 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 mi- grate 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 nerve- roots from those which become separated from the spinal ganglia, after they have advanced beyond the point of union of the sen- sory 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 to- ward 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 aeecompany- ing 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 communicat- ing 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 SUPT. Fic. 7. Diagrammatic transverse section of an embryo 9 or 10mm. 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. w.b., Wolffian body. indicates that the rate of migration reaches its maximum in em- bryos 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 Neu- mayer have attempted to account for the cells giving rise to the sympathetic nervous system, by local differentiation of elements SYMPATHETIC SYSTEM 1N 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 inves- tigations of Cajal, Harrison, and others, the neurone theory | eo ee ee SAS See ee ae Oe ¥ Fic 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 ‘‘accom- panying” 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 ‘‘in- different’’ 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 tothe 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 an- lagen of the sympathetic trunks are rapidly becoming fibrous. They still appear as continuous cell-columns showing little evi- dence 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 nerve- roots as they traverse the marginal veil. The spinal ganglia are also becoming more sharply 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 1N MAMMALS 2Zal Fie. 9 Fia. 10 Fie. 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 dotsin center = neuroglia cells; block cells = nerve cells. Fie. 10. Ganglion cells, neuroblasts, and indifferent cells, « 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. f Neuroblasts in the anlagen of the sympathetic trunks. 2p) of ectodermal origin undergoing mitotic division near the inter- nal 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 trans- formed either into neuroblasts or into embryonic supporting cells. In the higher vertebrates, however, many of these “‘in- different”’ cells retain a capacity for further propagation by di- As 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 communi- cating 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 usu- ally 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 in- termediate positions show varying degrees of elongation. This elongation cannot be accounted for mechanically, as by the squeezing through a narrow aperture in the external limiting membrane. These apertures are usually broad enough to per- mit free passage of the cells. It is probably due to more subtle forces which are operative in the process of migration. Further- more, it is by no means certain that such a change of shape al- ways takes place. Rounded ‘‘indifferent’’ cells are sometimes observed far distally along the course of migration, while elon- gated cells are present in the motor niduli and in the distal ends of the spinal ganglia. SYMPATHETIC SYSTEM IN MAMMALS Zoo 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 sympa- thetic 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 he 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 sym- pathetic trunks. These facts establish a direct genetic relation- ship between the sympathetic and the central nervous system. Weare 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 thespinal ganglia. Doubtless, many arise by the mitotic division of ‘‘indifferent”’ cells along the course of migration. The sources of these migrat- ing 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 cell- aggregates lying along the ventro-lateral surfaces of the aorta in the dorsal and the lumbar regions. In these regions the sympa- thetic 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 sur- faces of these bodies which, at this time, appear as compact cell- columns 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-aggre- gates 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- Fic. 11. Transverse section through the sympathetic trunks and the supra- renal bodies of an embryo 9 mm. in length, * 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 pos- terior region. In transverse sections through this region, traces of one or the other of these plexuses appear in nearly every sec- tion. 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 pro- nounced. 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 com- pletely 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 an- lagen 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 prever- tebral 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 “‘in- different’’ cells along the courses of migration. The develop- ment of the prevertebral plexuses is, therefore, entirely analogous with the development of the sympathetic trunks. Ill. VAGAL SYMPATHETIC PLEXUSES (a) Introductory—Under the term vagal sympathetic plex- uses, we shall consider those plexuses, usually regarded as sym- pathetic, 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 Fic. 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 sym- pathetic trunks and the prevertebral plexuses to the spinal gan- glia, could not derive the peripheralsympathetic plexusesfrom 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 re- lated 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,’ have shown that the cardiac plexus and the sympathetic plexuses in the walls of the visceral organs do not owe their origin to the sym- pathetic trunks as has hitherto been supposed, but that they 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 separate from the origin of the sympathetic trunks and the sympathetic plex- uses described above as prevertebral plexuses, they cannot prop- erly 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 sympa- thetic’’ is a departure from the established nomenclature, but inasmuch as there is no good collective term which could be ap- plied 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 pleruses.—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 character- istic chromatin structure of their nuclei. Many of them appear to become separated from the nerve-trunks and to wander into the walls of the esophagus 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 réle 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 cesophagus. 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 cesophagus 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 Fie. 13. Transverse section through the cesophagus and the vagus trunks in an embryo 7mm. in length, < 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 can- not 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 the cesopha- gus. 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 cesophagus except along the fibers of the growing branches. These courses are still plainly visible. The migrant cells in the walls of the cesophagus 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 prox- imal 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 con- tinue 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 cesophageal 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 cesophagus, the cells which have wandered in are aggregated into more or less distinct groups arranged in two broken rings. The myenteric and the submu- cous plexuses are thus becoming distinct. A similar arrangement, though less definite, is apparent also in the walls of the intestine. 240 ALBERT KUNTZ Fibrous connections have become established with the sympa- thetic 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 prob- able, 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 cesophagus, in the region of the bifurca- tion of the trachea, are carried out along the anterior and the dor- sal surfaces of the bronchi. These cells obviously give rise to the anlagen of the pulmonary plexuses. (d) Cardiac plexus.—The first unmistakable evidence of gan- glia 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 pene- trated deep into the angle between the aorta and the pulmonary artery. These cell-aggregates constitute the anlagen of the earli- est ganglia of the cardiac plexus. They are without fibrous con- nections 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, carn). 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 gan- glia 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 appar- ently free from migrating cells, the cardiac nerves having their SYMPATHETIC SYSTEM IN MAMMALS 2A TS). | @ 5 @ ere \e a p09 vo Fria. 14. Transverse section through the cesophagus, the vagus trunks, and the anlagen of the cardiac plexus in an embryo 13 mm. in length, * 100. ‘ origin in the sympathetic trunks still contain numerous “‘accom- panying”’ 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 sympa- thetic 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 carciac nerves having their origin inthe sympathetic trunks contain many cellular elements. He con- 242 ALBERI KUNTZ cluded, therefore, that the ganglia of the cardiac plexus are com- posed 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 eardiae 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.,? on the human embryo, except that the earliest cardiac nerves having their origin in the sympathetic trunks are less intimately asso- ciated 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 cone- shaped 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 mi- grating 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 dis- tal end and to wander peripherally along the vagus trunk. M1- § Abhdl. Math-physischen Classe d. Kénigl. 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 andthe 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 Fia. 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, < 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 anlagen of 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 nucle- olus 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 mob chl Fia 16. Diagram designed to show the relation of the vagi to the vagal sympa- thetic plexuses. The arrows indicate the course and the direction of the cells mi- grating from the walls of the hind-brain and the vagus ganglia into the anlagen of the vagal sympathetic plexuses. a.a., Aorticarch. au., Auricle. b.c., Buccal cavity. br., Bronchi. car.p., Anlagen of cardiac plexus. cb.l.,Cerebellum. /f.b., Fore-brain. g.t., Ganglion of the trunk. m.b., Mid-brain. m.ob., Medulla oblongata. n.t., Neural tube. oe., (Esophagus. oe.p., sophageal plexus. p.a., Pulmonary artery. st., Stomach. ¢., Trachea. ven., Ventricle. ven. 7V., Fourth ventricle. v.t., Vagus trunk. 10., Roots of vagus nerve. 11., Roots of spinal accessory nerve. 12., Roots of hypoglossal nerve. c.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 sepa- rate 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 ‘‘in- different” cells along the course of migration and in the anlagen of these plexuses. The vagus ganglia and the walls of the hind- brain, 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, < 1100. a., Neuroblast in the vagus rootlets. 6., Neuroblasts in the vagus trunks. c., Neuroblasts in the ganglia of the trunk. d., 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, there- fore, 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 gener- ally 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. Fur- thermore, 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 vis- ceral 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 sym- pathetic plexuses cannot be derived from the sympathetic trunks. These findings also give the vagi an importance in the develop- ment 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 con- tains frequent allusions to the migration of medullary cells ever since the time of Balfour (’75). That pioneer among the inves- tigators 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 observa- tior.s 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 amphib- ians, 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 Schiifer (’81) and Kolhker (’94) in the ventral roots of the spinal nerves in the cat. “Thompson (87) described cells which he interpreted as degenerating gan- glion 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 these 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 gan- glia along the fibers of the spinal nerves. He suggested that these cells take part in the development of the neurilemma. He be- lieves, 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. Car- penter (’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 neuri- lemma. In the oculomotor nerve, however, some of these ‘‘in- different’’ 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 ob- served migrating into the ventral roots of the spinal nerves in embryos of the pig become cells of the neurilemma and there sub- serve 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 ven- tral 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 ALBER1 KUNTZ cellsasafact. The present series of observations shows, moreover that the migration of medullary elements plays a far more im- portant 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 gan- glia. Furthermore, it has been shown that some of the cells which migrate peripherally from the neural tube and the cerebro- spinal ganglia give rise to the sympathetic nervous system. (b) The neurilemma.—An extended discussion of the develop- ment of the neurilemma is beyond the scope of this paper. Inas- much, however, as the histogenesis of the neurilemma is so inti- mately related to the histogenesis of the sympathetic neurones, its origin may be considered briefly at this point. As the ‘‘indiff- ent’’ cells migrate peripherally from the neural tube and the spinal ganglia, they migrate not only into the anlagen of the sympa- thetic 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. -hey cannot be accounted for in any other way. Not a few of the more recent investigators, including the advo- cates 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 differenti- ation on this point, but we must disagree with them as to the man- ner 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 neuri- lemma 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 pre- sents 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. Ac- cording to this interpretation, the cells of the neurilemma are homologous with the neuroglia cells in the central nervous sys- tem. (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 demon- strated. Froriep, like Langley, Koélliker, 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 codrdinated 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 contrac- tions 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 KUNTZ 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 cor- rect 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 indi- cates, so ne of these plexuses contain sensory neurones, it 1s proba- ble that these arise from cells which migrate from the vagus ganglia. While it is impossible by direct observation to trace either sym- pathetic 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 excita- tory 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 dor- sal nerve-roots. All these facts suggest that the cerebro-spinal ganglia have arisen from cells which originally lay within the neu- ral 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 Zit 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 Kohn- stamm (’00) render the existence of efferent fibers in the dorsal nerve-roots of the higher vertebrates extremely doubtful. Inas- much 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 inter- pretation makes the sympathetic neurones entirely homologous with the efferent and the afferent components of the other func- tional 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 com- mon 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 relation- ships 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 preverte- bral 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 sympa- thetic neurones, however, all have the same genetic relation- ships; they are the descendants of the ‘‘germinal”’ cells of His. 252 ALBERTI 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 acces- sion 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. (f) 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 investiga- tion. 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 cen- tral 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 plex- uses have their origin in the hind-brain and the vagus ganglia will probably be of even greater physiological and psychological im- portance. 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 phy- siological 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 thé greatest difficulties, but promises to be fruit- ful of the most far-reaching results. SYMPATHETIC SYSTEM IN MAMMALS 45 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 spi- nal nerves. Some of these cells deviate from the course of the spinal nerves and, migrating along the pathsof 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 re- gion 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 sys- tem which has hitherto been unrecognized. 4. The cells migrating peripherally from the cerebro-spinal system along the spinal nerves and the vagi are the descen- dants of the ‘‘germinal”’ cells of His; viz., the ‘indifferent’ 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 sym- pathetic system. Some become distributed along the nerve- fibers 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 sympa- thetic 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 sym- pathetic trunks and the prevertebral plexuses has not been demon- strated. 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 sympa- thetic sensory elements back to their specific source in the cere- bro-spinal nervous system, indirect embryological and anatomical evidence warrants the conclusion that the sympathetic excita- tory 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 interpreta- tion makes the sympathetic neurones homologous with the affer- ent and the efferent components of the other functional divisions of the peripheral nervous system. 8. Inasmuch as the cells migrating peripherally from the cere- bro-spinal nervous system are the ‘‘indifferent”’ cells and the ‘““neuroblasts’” of Schaper, Schaper’s conception of the develop- mental 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 vegeta- tive life. SYMPATHETIC SYSTEM IN MAMMALS PAS 5 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 Wistar Institute of Anatomy and Biology, April 21, 1910. Printed July 8, 1910. 256 ALBERT KUNTZ BIBLIOGRAPHY ¢ Aver, 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. Bautrour, 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. BarpEEN, ©. 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. Bearp, J. Morphological studies. 2. The development of the peripheral nervous 1888 system of vertebrates. Part 1.—Elasmobranchii and aves. Quart. Jour. of Micr. Sci., n. s., vol. 29, pp. 153-227. Casat, S. R. Nouvelles observations sur l’évolution des neuroblastes, avec 1908 quelques remarques sur lhypothése neurogénétique 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. Zooél., Harvard College, vol. 48, pp. 141-228. CARPENTER, F. W., anp Main, R. C. The migration of medullary cells into th. 1907 ventral nerve-roots of pig embryos. Anatomischer Anzeiger, Bd 31, pp. 303-306. Dourn, A. Studien zur Urgeschichte des Wirbelthierkérpers. Nr.17. Nerven- 1891 faser und Ganglienzelle. Histogenetische Untersuchungen. Mitteil. d. Zoédlog. Station zu Neapel, Bd. 10. 1901 Studien zur Urgeschichte des Wirbelthierkérpers. Die Schwann- schen Kerne, ihre Herkunft und Bedeutung. Erwiderung an A. v. Kolliker. Mitteil. d. Zodlog. Station zu Neapel, Bd. 15. Froriep, A. Die Entwickelung und Bau des autonomen Nervensystems. Medi- 1907 zinisch-naturwiss. Archiv, Bd. 1, pp. 301-321. Gurwitscu, 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. Sitzwngsber. d. Niederrhein. Ges. f. Nat. u. Heilkunde zu Bonn, 1904, B., 8., 55. 1906 Further experiments on the development of peripheral nerves. Amer. Jour. of Anat., vol. 5, pp. 121-131. SYMPATHETIC SYSTEM IN MAMMALS D5 His, W. Ueber die Anfinge des peripheren Nervensystems. Archiv f. Anat. u- 1879 Physiol., anat. Abt., pp. 455-482. 1888 Zur Geschichte des Gehirns sowie der centralen und peripherischen Nervenbauen beim menschlichen Embryo. Abhid. d. Sdchs. 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 Wirbelthieren- 1891 Abhdl. Math-physischen Classe d. Kénigl. Sdchsischen Gesell. d- Wiss., Bd. 8, Leipzig, 1897 Ueber die Entwickelung des Bauchsympathicus beim Hiithnchen und Menschen. Archiv Anat. u. Entwg., Supplement. Horrmann, C. K. Zur Entwickelung des Sympathicus. I. Die Entwickelungs- 1900 geschichte des Sympathicus bei den Selachiern (Acanthias vul- garis), Verhdl.d. Kénigl. Akad. v. Wetensch. te A msterdam. (Tweede Sectie), Deel vil. 1902 Zur Entwickelung des Sympathicus. II. Die Entwickelungsgeschichte des Sympathicus bei den Urodelen. Verhdl. d. Kénigl. Akad. v. Wetensch. te Amsterdam, Deel viii. Huser, G. Cary. Lectures on the sympathetic nervous system. Jour. Comp. 1897 Neurol. and Psychol., vol. 7, pp. 73-145. Jounston, J. B. The nervous system of vertebrates. Philadelphia. 1906 Jones, W.C. Notesonthe development of the sympathetic nervous system in the 1905 common toad. Jour. Comp. Neurol. and Psychol., vol. 15, pp. 1138- 131. Koun, A. Ueber die Entwickelung des peripheren Nervensystems. Verhdl. d. 1905 anat. Gesell. auf d. 19 Versamml. in Gemf. Ergdnzungsheft zum 27. Bd. d. Anat. Anz., Jena. 1907 Ueber die Entwickelung des sympathischen Nervensystems der Siugetiere. Archiv f. mikr. Anat., Bd. 70, pp. 266-317. Kounstamm, O. Zur Theorie des Reflexes von hinterer Wurzel auf hintere Wur- 1900 zel. Zentralbl. f. Physiol., Bd. 14, p. 457. K6LuiKER, A. Ueber das Vorkommen von Nervenzellen in den vordern Wurzeln 1894 der Riickenmarksnerven der Katze. Verhdl. Gesell. Naturf. w. Aerzte. 66. Versamml., Theil 2, Halfte 2, p. 363. 1905 Die Entwickelung der Elemente des Nervensystems. Zeitschrift f. Wiss. Zoél., Bd. 82, pp. 1-38. Kuntz, A. A contribution to the histogenesis of the sympathetic nervous system. 1909 Anatomical Record, vol. 3, pp. 458-465. 1909 The réle of the vagi in the development of the sympathetic nervous system. Anatomischer Anzeiger, Bd. 35, pp. 381-390. Lituin, F. R. The development of the chick, pp. 256-261. New York, 1908 258 ALBERT KUNTZ Marcus, H. Ueber den Sympathicus. Svtzwngsbericht d. Gesell. f. Morph. u. 1909 Physiol. in Miinchen. Neat, H. V. The development of the ventral nerves in the Selachii. Mark Anni- 1903 versary Volume, New York, pp. 291-3813. Neumayer, L. Histogenese und Morphogenese des peripheren Nervensystems, 1906 der Spinalganglien und des Nervus Sympathicus. Handbuch der vergl. u. experiment. Entwickelungslehre der Wirbeltiere, Bd. 2, 3. Theil, pp. 513-626. Onopr, A. D. Ueber die Entwickelung des sympathischen Nervensystems. 1886 Zweiter Theil. Archiv f. mikr. Anat., Bd. 26, pp. 291-313. Parerson, 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. 1847 Scuirer, 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. 348. Scuaper, A. Die friiesten Differenzierungsvorginge im Zentralnervensystem. 1897 Kritische Studie und Versuch einer Geschichte der Entwickelung nervoser Substanz. Archiv f. Entw.-Mech., Bd. 5, pp. 81-1382. ScHENCK, S., AND BrrpsaLtt, W.R. Ueber die Lehre von der Entwickelung der 1878 Ganglien des Sympathicus. Mitteil. aus d. embryol. Instit. iii, Wien. Van Wine, J. W. Ueber die Mesodermsegmente des Rumpfes und die Entwicke- 1889 lung des Exkretionssystems bei Selachiern. Archiv f. mikr. Anat., Bd. 33. Viena, W. Mémoire sur le développement des tubes nerveux chez les embryons 1883 de mammiféres. Archives de Physiol. norm. et pathol., 3. série, delesP aris: Tee CONTROL OF PHOTOTACTIC REACTIONS _IN HYALELLA BY CHEMICALS HARTLEY H. T. JACKSON From the Zoélogical Laboratory, University of Wisconsin Loeb! 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 photo- taxis 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 25m solution of ethyl alcohol, a m/6 solution of ether, and a m/25 solution of ammonium chloride, but a m/10 solution of boracice acid failed to produce such an effect. My own experiments were performed on another amphi- pod, 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 that Hya- lellas were made positive by boracie 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- 1LorB, J. The Control of Heliotropic Reactions in Fresh-water Crustacean, by Chemicals, especially COz. 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 posi- tive 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? 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 em- ployed 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 candle- meters 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 phototactie 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 .O1 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 *Lores, J. Comparative Physiology of the Brain and Comparative Psychology. New York, 1900, and l.c. 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 photo- tactic, 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 fig- ures 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 solu- tion. 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 thai it is not chemical change in the tissues which caused the reversal of reaction, for, if it were, it would be impos- sible to increase the concentration and still have the animals 262 HARTLEY H. T. JACKSON TABLE OF PHOTOTACTIC REACTION IN HYALELLA. a z i HYDRO- AMMONIT- roncure | um | “lem | "am | engi) cauome | "ewer gon ee 3 | = No. No. No. No. No. No. No. + — = +— + — a eS +— 0000 0 21 Q 17 0 20 0 18 120) Q 20 0 20 0008 0 21 Only 0 20 018 0).21 0 20 1 19 0016 Opi iel6 0 20 0 18 0 21 Q 20 0 20 * 0023 0 21 0 17 0 20 ike WZ 0 21 0 20 Q 20 0031 Oni 0 17 0 20 018 (0) PA 0 20 0 20 00388 0 21 0 17 0 20 0 18 0 21 0 20 0 20 0046 0 21 0 17 0 20 Ik WL 0 21 0 20 0 20 0053 0 21 OM 0 20 Ph W8) (0) Pall Q 20 0 20 0060 0 21 Om i it) 1183 0 21 0 20 0 20 0067 0 21 0 17 4 16 0 14 0 21 0 20 0 20 0075 (0) All 0 17 3 AG 0 9 0 21 0 20 0 20 -OLO0 (0) All 0 17 1 14 (0), 0) 0 21 1 19 0 20 O15 0 21 (0) sly 1 12 Q) 7 Q 14 0 19 Q 20 .022 0 21 (0) itz iL gli 0 4 Ons LNG 0 20 029 () Pal) LAS Dies ele2 0 20 .036 ORO OL, Q 12 0 20 043 0 13 0 17 OR 0 20 050 0 4 1 16 & 0 20 060 0 17 0 20 067 (0) 7 0 20 074 (0) 17 1 19 100 0 17 1 19 150 | ale 0 20 .220 ) (0) iy 0 20 290 0 17 1 18 360 | O 14 0 19 430 OS eG 018 . 900 0 18 .570 0 18 .640 O18 PHOTOTACTIC 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 read- ily 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 sud- den 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 8, 1910. ty? abe» ae % ‘a laedh iG “ ‘ 7 7 we ee j Cee aad , 7 1 8 aa vise fe 7 in n ahi : Rn 3 0 ia Ad Ty th 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 stu- dents 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 dissec- tion, 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 Cc. W. PRENTISS purpose in the present paper is to describe the rudimentary gang- lia which oecur 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 develop- ment 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 hypoglos- sal nerve. Anterior to this hypoglossal ganglion the neural crest was undifferentiated. Froriep and Beck (’95) found a precervi- cal ganglion present in the adult throughout those groups of mam- mals 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 commis- sure, 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.”’ Ina 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 commis- sure could not be subdivided into definite ganglia; it was charac- terized 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. Inembryos 10 to 13 mm. long the neural crest becomes differenti- ated 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: alae LENGTH ee, Sue If | 5- 7 mm. 6 mm. 5 II 8-10 mm. 8.5-9 mm. | Ill 12-14 mm. 13-13.5 mm. 10 IV | 17-20 mm. 17-18 mm. 10 V 28-30 mm. 28 mm. VI 41-50 mm. 2 BIRO G SG tees aoe alos co <1. A ieee 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 micro- scopic cell masses and trace the course of very small fiber bundles. The dissection was then transferred to aleohol and examined as 268 Cc. 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 reconstruc- tions. DESCRIPTION OF DISSECTIONS Stage 1. 6mm. In this embryo (fig. 1) the ganglia were con- nected from the glossopharyngeal (1X) back to the caudal region crist.neur -- Jang. sup. - 9ang jugu! = Ix. - gang petros XI gone | PN \\ % roriep ose qang.nodos. —-—~ XI Fig. 1. Dissection of a6 mm. pig showing in the hypoglossal region the undif- ferentiated 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 (C, C2) 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 (X/J/J). The spinal accessory (XJ), as an arched bundle of fibers, could be traced from the fourth cervical ganglion cephalad to the vagus. Dorsal to the ac- cessory and partly ensheathing it is a flattened band of cells, the c HYPOGLOSSAL GANGLIA OF PIG EMBRYOS 269 neural crest (crist. neur.), extending forward to the jugular gang- lion of the vagus (gang. jugul.). Opposite the posterior root of the hypoglossal a marked ventral loop and thickening in the crést (gang. Froriep) shows the position of Froriep’s ganglion. An- teriorly the crest of cells is broader and a few short proximal rootlets are present. A depression separates it more or less com- pletely from the cells of the jugular ganglion which is flattened and diffuse with 8-10 short proximal roots. The glossopharyn- geal 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 hypo- glossal 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 devel- oped. 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 jug- ular ganglia shows the most marked change. The slight enlarge- ment 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 resem- bles one of the two cell masses composing the first cervical gang- lion. 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, pl. 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 gang- lion is of more definite form and is pointed ventrally. Dorsal myelencephalon SWyouwpuy anne 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. gang. hypogl. <— = oe ws ' A r = gang sup. gang . 2 iii NS | B gang juqul. Froriep ~ fa ae BX. nod J -§-— gang.petros | en .Ae a QONGINOGoS ee ht XIL SE Seera Nee D.¢ Fie. 3. Dissection of a9 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 na the peripheral portion of the spinal accessory (XJ’). 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 if gang Jugul Fig. 4. Dissection of a 13 mm. pig to show a series 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 prox- imal roots are longer, however, and the neural crest near the jug- ular 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) Die Cc. 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. gang, hy pogl. gang. Froriep e gang. Sup. - gang jugul Se eee ANG Fic. 5. Dissection of a 13mm. pig showing a series of eight hypoglossal ganglia and the persistence of the neural crest from the superior ganglion to the first cer- vical. Stage 3. 13mm. 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 evi- dence. A third cell mass more cephalad shows a proximal root, is pointed at its distal end and projects ventrad to the spinal ac- eessory. Anterior to these three occurs a series of five cell masses, more diffuse, more closely united and elongate in the antero-pos- terior 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 distinetly double as in figs. 4 and 5. In the other two it gave evi- dence of-a double origin. An interesting point was the difference in structure exhibited on the right and left sides of the same em- bryo. 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 fig- ured 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 gangha 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 gang. hypog).. Us ates gang.nodos. See Sait Fie. 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 ats 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 con- necting with a small lateral spur. The distal root is large and gang hypoql, é gang Froviep -. f — gang. i gang i reel - eames 5 gang: jugul! Ps Gan petros Fie. 7. Dissection of an 18 mm. pig with three distal roots passing from the hypoglossal ganglia to the ventral roots of the hypoglossal nerve. The first cervi- cal 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 ex- tremely 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 jointhespinal 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. em- bryo shown in fig. 6. In this case three distal roots join the hypo- glossal 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 divisionof the first cervical gang- lion as seen in fig. 4, the resemblance is plain. From ten dissections at this stage we would note the following points: (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 connec- tion 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 gang- lion (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 DAF 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 myelencephalon 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 Cc. 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 hypo- glossal 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 rudi- mentary, 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 resemb- ling 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 mam- mals 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 spindle- shaped masses of cells. Other spinal ganglia show the same con- dition. The first cervical ganglion possesses always several proximal roots (4-5) and the distal root arises as two distinct bundles. 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 ganglha 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 hypo- glossal 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. Fur- thermore 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 accessory. 280 Cc. W. PRENTISS The present lack of segmental arrangement displayed by these ganglia does not preclude their metameric origin. Their develop- ment 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 hypoglos- sal 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 gang- lia located between the vagus and the first spinal nerve, the gang- lion of which would correspond to Froriep’s ganglion in mammals. According to this evidence, four dorsal ganglia have become rudi- mentary structures in mammals and the corresponding ventral roots have united to ferm 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 cervi- cal 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 glosso- pharyngeal 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 18 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 gangha ean 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 gen- erally 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 gang- lion. 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 evi- dence of their rudimentary or vestigial character. 7. One, frequently two or three, and in one case four hypo- glossal 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 neu- ral crest. It is similar to the connections which were found per- sisting between the roots of the spinal ganglia. 9. If we consider a pair of hypoglossal ganglia as the equiva- lent of a single spinal ganglion, four such ganglia would be repre- sented 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 develop- ment in embryos 17-20 mm. long, then retrograde, though gang- lia 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 Cc. W. PRENTISS BIBLIOGRAPHY Bremer, J. L. Aberrant roots and branches of the abducent and hypoglossal 1908. nerves. Jour. Comp. Neur. and Psych., vol. 18, pp. 619-639, 9 figs. Froriep, A. Ueber ein Ganglion des Hypoglossus. Archiv f. Anat. u. Physiol., 1882. Anat. Abth., 1882, pp. 279-302. Frorigp, A. uND Beck, W. Ueber das Vorkommen dorsaler Hypoglossuswurzeln 1895. mit Ganglien in Reihe der Siiugethiere. Anat. Anz., Bd. 10, pp. 688-696. Lewis, F. T. The anatomy of 212mm. pig. Amer. Jour. Anat., vol. 2, pp. 211- 1903. 225, 4 pls. Martin, P. Die Entwickelung der neunten bis zwélften Kopfnerven bei der 1891. Katze. Anat. Anz., Bd. 6. Merk, A. The encephalomeres and cranial nerves of an embryo of Acanthias 1909. vulgaris. Anat. Anz., 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. 83- 116, 4 pls.. 14 text figs. Accepted by the Wistar Institute of Anatomy and Biology, May 16, 1910. Printed September 9, 1910 EXPLANATION OF FIGURES All 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 anterior 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: Cl, C2, first and second cervical ganglia; crist. neur., neural crest; gang. Froriep, Froriep’s ganglion; gang. hypogl., hypoglossal ganglia; gang. jugul., jugular gang- lion; 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 neive. THE DEVELOPMENT OF THE SYMPATHETIC NERVOUS SYSTEM IN BIRDS! ALBERT KUNTZ From the Biological Laboratory, State University of Iowa TEN FIGURES CONTENTS NSN EG GUE DOIN PES ERCGNS Aaa ie God oarc¥ aves are, SE MND RIA uiallatad Bye 283 KO DSE VAITON Rpts me ene i Se re aa kw kage ok OES. Moms ce i Seok 284 OV ma pet Me itOm TRUM Ka whem a ss 23s ea oS do eL, < wwhee See 284 QyMELOGUCLOT Ym aare eee: 202 0s vce lo a ees os eae ed. 284 ba Primary: Svan pAGhewenbRe@kes: 2): .. © a: oc e Raye eet ch cases ie 286 Ce COG anyas vp abe bi ChLEUIMIKS |): = 5. ee eens eee ree Se 287 ae) & IES HOY ETS BETS) SIME Aa ler oe So, ha a ae ea 288 2. Prevertebral plexuses.............. MP ccc, ae ate car eee hes ot 293 Jee Cram OMy OLMEVCTIA AK Nee yee tafe say's 6 2G «2 acdc EO ec Ses ty ee 295 4, Vagal sympathetic plexuses................; i OMe Pan in ae reg 296 Go AMtroguevory:.2ah ot canes oe. - 5 ee 296 o. Myenteric-and submucous plexuses... 2...mete esse sca ee ee 296 Gebulmonary plexusesie as aitr ced.) so) oo cis Se oe atta oe 299 daz ancillary ple xS| tae. pease ter. ics senisiahs. » arn ioe eee ieee ae ee 299 CPMLLIST OZCMESISS, 6 115.2 oy aruce Gad, ons 9 + +, ico 8 ERR pete aiuto 300 Discussion.or results, and Conclusions .......... 0. ..Geeeeeeeto st ee oe ee 301 Nf UPTRW ORES ei a oe ea, a ag ag i A AE 305 Bibliographwarn wea ositenc Sooo ok cs bois. «2. . cack ene te eke 308 INTRODUCTION The present investigation of the development of the sympa- thetic 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 morphogene- tic differences in the development of the sympathetic system in birds and in mammals, with a view to their phylogenetic signifi- cance. Birds and mammals have become specialized along divergent lines. Their special habits of hfe have brought about modifica- tion 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 vegetative functions, has not escaped the modifying influence of specialized habits. It is hoped, therefore, that a more exact knowledge of the development of the sympa- thetic 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,? only such references will be made to the liter- ature 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-aceto- formaldehyde. The sections were cut to a thickness of 10 micra and stained by the iron-hematoxylin method. This method, as indicated in the earlier paper, was found best adapted for pur- poses 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 Comparative 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; ne., notochord; p.sy., primary sympathetic trunks; sp.g., spinal ganglion; sp.n., spinal nerve; s.sy., secondary sympathetic trunks. to His, Jr., the cells giving rise to both the primary and the sec- ondary 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-ageregates 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 haveassumed the appear- ance of loosely aggregated cell-columns (fig. 2, A, p. sy.). Well marked ganglionic enlargements do not occur, but the cell-col- umns 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 mesen- chyme, either singly or in small groups, toward the sides of the aorta (fig. 2, A and B, 7. 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 dor- sally and recede a short distance from the walls of the aorta until at the close of the fifth day (J20-hour stage) they appear as con- spicuous cell-columns lying along the dorso-lateral aspects of the aorta a short distance from its surface (fig. 2, C, p. sy.). The pri- mary sympathetic trunks are now sharply defined in the anterior region and are connected with the spinal nerves by distinct cel- lular tracts. In the posterior region, the cell-aggregates are still loosely szattered 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 maxi- mum development. During the course of the sixth day, they decrease materially in size until at the close of the sixth day (144- hour-stage) they have almost disappeared. Their complete dis- La 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 sympathetictrunks.—Theanlagen of the secondary 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 Sp.n. 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), * 200. C., Primary and secondary sympathetic trunks (120 hours incubation), X 200. D., Secondary sympathetic trunk (144 hours in- cubation), 100. ao., aorta; c.r., communicating ramus; 7.c.c.7r., cells migrating from spinal nerve to primary sympathetic trunk; 7.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 longi- tudinal commissures. Like the anlagen of the primary sympa- thetic 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 (iece2 ©) 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 secon- dary 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 sympa- thetic 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 gan- glionic enlargements become connected by the fibers of the longi- tudinal commissures. (d.) Histogenesis—Asalready indicated, thesympathetic trunks arise from cells which migrate peripherally from the cerebro- spinal 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 PCUMNT. 1G. Fig. 3. Transverse section of the neural tube and the spinal ganglion of an em- bryo of the chick (105 hours incubation), * 190. ¢.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 KUNTZ process may be observed as early as the 64-hour stage. During the fourth and the fifth day, after the spinal ganglia have be- come well differentiated, a few cells may be observed migrating from the dorsal part of the neural tube into the dorsal nerve- roots (fig. 3, c. m. d. n. r.). It is probable that cells do not mi- grate from the dorsal part of the neural tube in any considerable numbers after the spinal gangha have become differentiated. Cell migration into the dorsal nerve-roots is probably only a tran- sient 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 peripheralsympathetic plexuses. In view of the comparatively enormous development of the pre- vertebral plexuses and of the ganglion of Remak in birds, this is ob- viously the fate of the elements composing the primary sympa- thetic trunks in the posterior region of the body. There is no evi- dence, however, of the peripheral migration of cells from the prim- ary 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. Thelast 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 Fig. 4. Neuroblasts drawn with the aid of the camera lucida, X 825. a., in ven- tral nerve-root inside external limiting membrane (105 hours incubation); 6., in ventral nerve-root outside external limiting membrane (105 hours incubation) ; c., in spinal nerve (105 hours incubation); d., in communicating ramus (105 hours incubation); e., in ventral nerve-root (96 hours incubation); f., in spinal nerve (96 hours incubation). the motor niduli, and medullary cells may be traced without diffi- culty from the motor niduli into the proximal part of the ventral nerve-roots (fig. 3, ¢.m.v. r.). Numerous accompanying cells are present in the spinal nerve-trunks as farasthelattermay 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 prac- tically ceased. The great majority of the cells migrating peripherally along the spinal nerves are characterized by very little cytoplasm, and 292 ALBERT KUNTZ by large rounded or elongated nuclei usually having their chroma- tin 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 Scha- per. 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 rela- tionships as the cells which give rise to the neurones and the neu- roglia cells in the central nervous system. They are all the de- scendants of the ‘“‘germinal’’ cells (Keimzellen) of His. These observations are in full accord with the writer’s observ- ations on mammalian embryos. There is a marked difference, however, in the chromatin structure of the embryonic medullary cells in birds andin mammals. In mammalian embryos, the mi- grant medullary cells are usually quite readily recognized by the chromatin structure of their nuclei. They also usually take a slightly deeper stain than the cellsof thesurrounding 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 comparatively 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 gangha. Cajal (’08) described cells which he recog- nized 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 de- velopment of the sympathetic trunks actually migrate as such from their sources in the cerebro-spinal nervous system. Doubt- less, 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 pri- mary 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 plex- uses. In the sacral region, the aorta is soon completely surrounded ventrally by a ring of loosely aggregated sympathetic cells (fig. a 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 sym- pathetic 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- 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., hypogas- tric 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- Fig. 6. Transverse section of an embryo through the sacral region of an embryo of the chick (130 hours incubation), 110. ao., aorta; g.r., ganglion of Remak; hyp., hypogastric plexus; mes., mesentery; 7., 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 296 ALBERT KUNTZ terminates in the region of the genital ridges, in a slender cellular cord which Remak has called the intestinal nerve (Darmnery). 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 prepara- tions show conclusively that the cells giving rise to the ganglion of Remak are derived directly from the anlagen of the hypogas- tric 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 my 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——tIn 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 andthe vagus ganglia. (b.) Myenteric and submucous plexuses.—In embryos of the chick in the 130-hour stage, the vagus trunks may be traced posteriorly along the walls of the cesophagus just a little below its ventral SYMPATHETIC SYSTEM IN BIRDS 297 level. The ganglia of the trunk lie close to the walls of the cesoph- agus 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 bifur- cation 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 cesophagus. 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 cesophageal 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 cesophagus between the latter and the bronchus. These branches may be traced poster- iorly but for a short distance at this stage. At the close of the sixth day, the vagi have become more con- spicuous. In the anterior region, definite lines of cells may be traced from the vagus trunks into the walls of the cesophagus where they become aggregated into more or less distinct groups arranged in two broken rings (fig. 7, m.s.p.). Posterior to the bi- furcation of the trachea, cells may be traced dorsally from the vagus trunks into the walls of the cesophagus (fig. 9, m.s.p.). The vagus branches lying between the walls of the cesophagus 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 cesophagus. During the seventh and the eighth day of incubation, the sym- pathetic plexuses in the walls of the digestive tube become well established. Branches of the vagi may be traced into the walls of the cesophagus, and the cell-groups constituting the anlagen of the myenteric and the submucous plexuses assume a more defi- nite 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, how- ever, 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- ‘Vag. 9 “G.5. Fig. 7. Transverse section through the esophagus 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., cesophagus; ¢., trachea; vag.n., vagus trunks. Fig. 9. Transverse section through the cesophagus and the anlagen of the car- diac 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; ve., esophagus; 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 pre- vertebral 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 submuc- ous 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 cesophagus and the bronchi. Cells wander out along these Fig. 8. Transverse section through the region of the lungs of am embryo of the chick (seventh day of incubation), X 80. ao., aorta; 0e., esophagus; 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 pul- monary 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 throughthe head-region of em- bryos 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 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; 7.l.m., in- ternal 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 sympa- thetic 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 intes- tine, 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 mi- grate peripherally along the spinal nerves; viz., they are the ‘‘in- different”? 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 sym- pathetic 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 cere- bro-spinal ganglia. The cells giving rise to thesympathetic 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 suc- ceeded 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 sympa- thetic 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 hind- brain and the vagus ganglia along the vagi. As I have pointed out in an earlier paper, experimental evidence indicates the exist- ence 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 func- tional 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 periph- erally from the cerebro-spinal system. Thegrowingnerve-fibers, doubtless, constitute an important factor in the peripheral trans- portation 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 anlagen of the sympathetic trunks before fibers are present in the communicat- ingrami. Likewise, cells migrate ventrally from the sympathetic trunks into the anlagen of the prevertebral plexuses before post- ganglionic 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 in- vestigation, such a position is untenable. My preparations show conclusively that medullary cells migrate into the motor nerve- roots 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 eell-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 sympa- thetic neurones. I have found no evidence of cells migrating from the neural crest into the ventral roots of the spinal nerves in em- bryos of birds and mammals. Cell-chains connecting the neural crest with the cell-aggregates in the ventral nerve-roots, doubtless, do oceur in embryos of the lower vertebrates. I have observed such cell-chains in embryos of Amblystoma. This does not, how- ever, 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 ‘‘ germ- inal” cells of His; viz., the ‘‘indifferent’’ cells and the ‘‘neuro- blasts’? 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 sym- pathetic neurones are homologous with the afferent and the effer- ent components of the other functional divisions of the nervous system. The histogenetic relationships of the sympathetic neu- rones were considered at some length in my paper on the develop- ment of the sympathetic nervous systeminmammals. 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 signifi- cance. ‘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 counter- part in mammals. Minor differences also occur in the develop- ment 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 homol- ogous 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 cer- tain directions, which have arisen in response to peculiar vege- tative functions. SUMMARY 1. The primary sympathetic trunks in the chick arise about the beginning of the fourth day of incubation, as a pair of cell- columns 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 commis- sures. 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 essentially 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 aggre- gated 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 pos- teriorly. They are derived directly from the primary sympa- thetic 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 inmammals. Such departure is no more than should have been expected in the specialized avian branch of the vertebrate series. 308 ALBERT KUNTZ BIBLIOGRAPHY Cajat, S. R. Nouvelles observations sur |’évolution des neuroblastes, avec 1908. quelque remarques sur l’hypothése neurogénétique de Hensen- Held. Anat. Anz., vol. 32, pp. 1-25 and 65-78. Froriep, A. Die Entwickelung und Bau des autonomen Nervensystems. Med. 1907. naturwiss. Archiv., vol. 1, pp. 301-821. Hewp, H. Die Entwickelung des Nervensgewebes bei den Wirbeltieren. Leipzig. 1909. His, Jr., W. Ueber die Entwickelung des Bauchsympathicus beim Hihnchen 1897. und Menschen. Archiv Anat. u. Entwg. Supplement. Kunrz, A. Acontribution 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. Acomparative study of the development of the sympathetic nervous system in birds and mammals. Abstract of paper read before the American Society of Zodlogists, 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. Onop1, A. D. Ueber die Entwickelung des sympathischen Nervensystems. Archiv 1886. f. 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 MGT OCeet ON 4, nets Re eas oie seis sea 25. wo «chk 4 ORTON Sth cs ee ge 309 IS LOTIC ANS kebe ber tas eres aso oes nd <3 2 2 2 sia. eee pore oa eke oko ee ise 312 Matenialandimethodameee reat ch 5 sav As sek Pee Mee een ae 322 The differentiation of the neural plate...... EES an AA Se nIe ae 325 he ditierentiationiot the Gasserian ganglion... .-- qe sees es es eae 330 Ehearicinotthielaperais y Line anclia -~.... +. aseeeeeermaes o-otes sen vee ee eee ‘Thesaudibory- vesicle and auditory ganglion. .. °..c942 22.55. - fs o-s ene ee he 334 the tateior the preaucitoryplacode. .: <2: :.2cscgenee maa ess 2 +e age ee eee 339 he originionpehe seniculate Can GuGN: :,. » ae | yap. Ke | xX | SSS Gen. Cut....) Gass. | sparen oe Mielec Awe: | namie (22 5 apes (Sheeran ae ee Jugular Agus. Latent.) 23 Lat. VII. ) Pres.,| Press| eee lexis lst abetake Pres. Communis .. jobee oo Pres. Desc sthie | Pres.| Pres. | Pres.| Pres.| Pres. | This table is slightly different from Herrick’s description. There is a lateralis ganglion in the region of the I[Xth nerve which he seems to attribute (’01, p. 208) morphologically to the Xth, although it is associated with the root of the [Xth. 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 VIIth, where there are two divisions, the dorso-lateral and the ventro-mesial; and in the VIJIth, [Xth, and Xth. The visceral ganglia are found in the VIIth, [Xth 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 distinc- tion 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. aoe 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 thematerial 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 rela- tiveagewasknownbuttheabsoluteagewasnotknown. Theseries 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 peri- pheral 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 elon- gated 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. Optie vesicle with slight cavity. Epidermis not thickened preparatory to the formation of alens. Auditory vesi- cle with a slight cavity and the lateral walls of the vesicle in con- tact. 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 sur- rounding 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, 7.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 prepara- tory to its detachment from the epidermis. Nuclei of cells in lens not yet located in the periphery of the cells. First epibran- chial 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 dparee om fie series. Stages VIII—IX correspond closely to a “eine 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 7 thick and it will be convenient to ex- press the spatial relations of structures in sections. THE CRANTAL GANGLIA IN AMEIURUS a25 THE DIFFERENTIATION OF THE NEURAL PLATE Ameiurus presents a rather striking difference from the descrip- tions usually given of the formation of the neural cord in its rela- tion 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 sub- stantial agreement among all the earlier descriptions of the cere- bral 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, text- figs. 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 coming 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 epi- dermis 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 theneural crest and dorso-lateral placodes are not differen- tiated from each other at first but appearasa largelateral thicken- ing lying on either side of the neural plate. This I have desig- nated 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 1s never recognizable 326 ' ¥. 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 indi- cated 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 Hal ness Pe ee as : neural plate e (nl FeG10n mesogers lat mass vas = Jresoderm endoverm veural kee/ endoderr 9 dor. lai mass vir lateral) ec /| nevral | re. av, placod 4 7? / . A C (a cord VB Ny : SSS SS \ Bee ea er ee, = \endode rm C D Fie. 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. Fic. C. A camera tracing four sections anterior to fig. 9, showing the relation of the lateral mass to the neural cord. Fic. 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 afew 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 dis- tinguish 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 Sot 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 pos- terior 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 form- ing 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), 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 [Xth 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 (dorso- lateral placodes). The fate of the lateral mass, as a whole, variesin different regions of the head. In the regions between the optic stalk and the Gas- serian 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 shouldbecompared. 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 w vhich there are numerous intercellular spaces. The later history of this mass shows that it is converted com- pletely into a very loose mass of mesectoderm with large intercellu- lar 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 VIIth the lateral mass is converted chiefly into mesectoderm, except its ventral border which repre- sents 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 VIIth 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 VIIth 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 audi- tory vesicle is reached (fig. 6.) The change in size of the preaudi- tory placode is almost imperceptible. It extends farther and farther dorsally until the condition of fig. 4 is reached. The audi- tory 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 character- istic of the auditory vesicle. The future lateralis VIIth 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 pos- teriorly 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 Gas- serian 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 VIIIth ganglion and the lateralis [Xth 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 repre- senting the primary derivatives from this structure. THE JOURNAL OF COMPARATIVE NEUROLOGY AND PSYCHOLOGY, VOL. 20, No. 4. 330 F. L. LANDACRE As menijomed above. the lateral] mass has ai first a periecily stalk throughout the whole lengeth of ithe head. the Interval mass aniemor to the vesicle is characierized by the resence of Inferceliuler spaces and the partial detachment of the the Gassetam ganzhon is not recognizable. li becomes recogniz- tenor to tirather than by any change m the region of the ganglion. a eee ee re Goes mot le exacily parallel with ithe long axis of the body, iis postenior, which Ges above the level of the middle of the neural eam2al Theres noihmg resembime 2 root ai this siaze. The Whole of ithe posterior end of the gamgiion is somewhai nearer the ead than ihe epadermis, but metther the root nor ihe trunk of the nerves from tht: ganghon appear for some hours. Anierortly, og edge ihe gangbonie mass passes gradually into Diesode formed m connection with this ganglion. The contact where Ii occurs is to be mierpreted 2s a failure of the lateral mass to separate completely from the epidermis. The same is true of the leteral portions of the ganghonie mass. Fig. 10 (Stage V) of the sanghome mass shortly after % ean be first recognized. THE CRANIAL GANGLI4 IN 4MEIUETS E<3! Its general shape is, in transverse section, at this timecireular except at its anterior end where it is dongated dorso-ventrally and lies quite close to the epidermis. The denser appearance of the cytoplasm. with indefinite cell boundaries, is characieristic of nearly all the early ganglionic masses im Amefurus. The later history of this ganglion is quite easy to follow. It becomes more definite in outline with cleaner borders and is quite distimet from any other ganglionic masses and cannot be confused with them. While in the early stages the posterior end of the Gasserian ganz- lion overlies that portion of the lateral mass from which the gent eulate ganglion is derived, they are not m contact until about the 86th hour im A. melas (fig. 83). For 2 long time after the aghty- sixth hour, while the two ganglia are in contact. their outlimes are quite distinct. Preceding this stage they are not even m contact with each other. The fibrillated root of the Gasserian ganglion appears at its posterior end where the ganglionie mass les near- est the brain as deseribed above. and ean first be detected In an embryo of 75 hours (A. melas). The development of this ganglion from the Isteral mass i 2 very definite feature of the embryology of Amefurus. There can be no doubt that it does not use all of the laters] mass m its forms- tion and that it does not come from 3 neural! erest as that term is generally used, although some portion of the lateral mass may be homologous with the neural erest of other types. The origin of structures which in other types come from the neural erest, and of dorso-lateral placodes from 3 common primordium im Amejurus seems to be due to the failure of these structures to differentiate in the early stages of the Isteral mass, and if it were not for the fact that so much of the lateral mass becomes converted Into me ectoderm one might speak of the neural erest region of the lateral mass; but in no region is all of the Iateral mass converted into 2 ganglion and since the specific structures derived from the Isteral mass are gangha, mesectoderm, auditory vesicle and plscodes, it would only introduce confusion into the deseription te refer to specific portions of the lateral mass as neural eresi, although m other types some of these structures are Known to come from the neural crest. BIS F. L. LANDACRE THE ORIGIN OF THE LATERALIS VII GANGLIA The origin of the lateralis VIIth ganglion resembles closely the origin of the Gasserian with the exception that the lateral mass (figs. 4 and 9, Stage ITI) giving rise to the lateralis VIIth 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 VIIth ganglion at first. Its anterior end, as in the case of the Gasserian, is situated more ventrally than the pos- terior end and comes into contact with the mesoderm of the mandi- bular arch anterior to the middle region of the arch. The anterior portion of this lateral mass later forms part of the geniculate gang- lion of the VIIth nerve, but at this time the anterior limit cannot be outlined, not, in fact, until the placodal portion of the VIIth is detached from the epidermis, which is some hours later (A. nebulosus, Stage VIII). The whole ganglionic mass has an up- ward 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 VIIth 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 deter- mined is not overlapped by the Gasserian ganglion but is over- lapped by its root. Its posterior end, just where it comes into con- tact 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, hes four sections anterior to the auditory vesicle. The lateralis VIIth 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 1s 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 endas 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 VIIth ganglion, the dorso-lateral and the ventro-mesial, can be deter- mined; these boundaries are shown in figs. 34 to 37 and in fig. 83. While these two divisionsof 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 upto 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 boun- daries in the lateralis VIIth ganglion comes from the fact that the root of the geniculate extends throughout the whole length of the lateralis VIIth 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 VIIth 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 VIIth, which is a pure acustico-lateralis ganglion, from the lateral mass in adjoining 334 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 ’05b). 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 VIIth have similar modes of origin. Owing to the intimate relation of the auditory vesicle and plac- odes with the lateralis VIIth ganglia, structurally and in point of time, I shall leave that portion of the lateral mass anterior to the lateralis VIIth and posterior to the Gasserian ganglion, 7.e., the communis VIIth or geniculate ganglion to be taken up in connec- tion with the epibranchial ganglion of the VIIth nerve. From the time that the lateralis VIIth assumes definite form until the visceral VIIth 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 dorsally 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 [Xth gangha. I find that the whole of the lateral mass in the auditory region is con- verted 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 VIIth that THE CRANIAL GANGLIA IN AMEIURUS 335 it is not possible to determine definitely whether there are lateral mass cells in the VIIIth or not, since it does not become distinct from the lateralis VIIth 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 postaudi- tory 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 audi- tory vesicle. At the extreme posterior end where the differentia- tion 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 charac- terized by having elongated radially arranged cells with their nuclei situated at the periphery. Ventro-laterally the elongated 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 em- bryo, 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 connect- ing 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 (’05b) 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 ight 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 con- clude 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 pre- auditory lateral mass. The lateral mass there furnishes the two lateralis ganglia of the VIIth nerve. The exact mode of origin of the VIIIth 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 VIIth ganglia and is so closely fused with the ventro- mesial ganglion of these two ganglia that it is impossible to be cer- tain 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 audi- tory ganglion; this, when it appears as a definitive mass, is situated as a cap of cells adhering closely to the ventral and ventro- mesial 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 later- alis ganglia of the VIIth nerve. At this stage before the appear- ance 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 pro- liferated from it. However, shortly after the separation of the preauditory placode from the vesicle the walls of the anterior THE CRANIAL GANGLIA IN AMEIURUS 307 end of the vesicle lose their characteristic regular outline, while throughout the remainder of the vesicle its walls are still charac- terized 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 VIIIth 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 VIIth ganglion, and seems to grow forward into this mass on its dorsal and mesial wall, one would be inclined to think that the whole audi- tory ganglion came from the vesicle. The conditions here are identical with those at the posterior end of the vesicle where the lateralis [Xth 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 con- tribute 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 cer- tain that the whole ganglion comes from the vesicle. Since, how- ever, 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 VIIIth ganglion and two later- alis VIIth ganglia for some time after this, it is impossible to say positively that the whole of the VIIIth 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 cleav- age separating the VIIIth from the lateralis VIIth which appear later, as well as by the manner in which the VIIIth 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 acus- tico-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 VITIth, however, does not affect the homo- geneity of this system of ganglia, since the acustico-lateralis system is based on anatomical and physiological characters. We have lat- eralis ganglia, asin the case of the VIIth, 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 cen- tral connections of this system in the brain and not on embryolo- gical 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 (VIIIth and two lateralis VIIth) become separated from each other at any given stage. Sometimes the dorso-lateral VIIth is free at one or both ends, while in other embryos of the same age one or both ends may be incompletely separated from the adjoining ganglia. There is also much varia- tion in the relative lengths of the two divisions of the lateralis VIIth, the dorso-lateralis VIIth 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 pos- terior boundaries of the vesicle can first be determined by the rad- ial arrangement of its cells and by the size of the vesicle, the plac- ode is represented by a slightly differentiated region in the ventral portion of the lateral mass. Fig. 4, Stage III, is taken four sec- tions 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 plac- ode 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. Slghtly 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 preaudi- tory 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 preaudi- tory 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 an- terior 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 theembryo. It is not likely that the placode has moved forward bodily, since the anter- ior 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 inter- vening between the posterior end of the placode and the anterior end of the vesicle where the placode has been converted into mes- ectoderm. 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 com- pletely 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 some- what 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 VIIth 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 epi- branchial 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 VIIth 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 sec- tions 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 notice- able, 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 x 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 VIIth 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 VIIth nerve, it will be more conven- ient to describe the origin of the geniculate ganglion here than to follow the natural order and take up the differentiation of the post- auditory 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 VIith is just forming. Fig. 19 is taken at the point of contact of the hyoid endodermal pocket, with the epidermis and corre- sponds in position to fig. 18. It differs in two 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 VIIth 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 be- comes detached, forming a portion of the geniculate ganglion, is quite easy to follow. Of the fate of the small mass of cells (z) shown in figs. 21 and 22, I am notcertain. They may enter into the ganglion, but it is more probable that they become con- verted into mesectoderm. They lie alittle 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 mitotie figures are beginning to be quite numerous. Fig. 24 is taken four sections posterior to fig. 23 and shows the transi- tion 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. neb- ulosus, 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 be- tween the two histologically. All of the preauditory placode pos- terior 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 differ- entiates 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 [Xth 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 ar- ranged 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 epi- dermis. In the succeeding section (fig. 28) the proliferated mass of cells, the anterior portion 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 prolifera- tion changes somewhat in appearance, being longer and its at- tachment 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 ap- pearance 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 [Xth 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 VIIth 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 fol- lowing 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 VIIth 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 OF 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 epi- dermis 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 he- comes 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 deter- mine 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 ar- rangement 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 out- line. 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 gen- eral 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 ('84b). These combine and later separate to give rise to the maxil- lary 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 VIIth also appears in these sections. The roots of the geniculate and Gasserian gangliaare 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 VIIth ganglion (V. LZ. 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 VIIth 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 de- tected. The arrangement of the roots of the geniculate, later- alis VIIth 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 VIIth ganglionic complex become more intimately fused so that it is not surprising that nerve trunks which are at first undoubtedly pure trunks, 7.e., contain only one component, later become mixed and contain two or more components. The con- tiguity 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 VIIth ganglia is in a par- ticularly 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 acustico- facialis 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 fu- sion 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 hyomandib- ular 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. Vand VII which, owing tothe contiguity of its. gan- glia, 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 Jateral mass is at first entirely homogeneous. Fig. 3 represents the appearance throughout its whole extent at an early stage, except that the lat- eral 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 con- tact 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, ex- cept at its ventral border; here where it comes into contact with the mesoderm it remains slightly denser. In the region where the Xth ganglion forms and possibly where the [Xth forms the process of conversion into mesectoderm does not go so far and the lateral mass in these regions is in some em- bryos slightly denser than in the surrounding regions. These two regions however are only 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 com- pletely 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 ill- defined 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, (Y), lying just over the mesoderm. The loose mass of cells shown in fig. 48 (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 indica- tion of a migration of cells from that structure. THE CRANIAL GANGLIA IN AMEIURUS apt 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, re- versing the relations to the medulla. These figures are taken from an embryo in which the condensations in the region of the TXth 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 identi- fied 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 dis- cussion 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. LANDACRE situated low in the body ona 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 meso- derm 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 [X ganglion will be taken up with the lateralis [Xth. The relation of the visceral Xth ganglion to the jugular or general cutaneous Xth which appears later and is situated near the medulla intra- cranially, 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. Inthe 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. Fromevidenceto be brought out later it appears that the lateral mass contingent of the gen- iculate 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, 2.e., the general cutaneous and general visceral ganglia of the head have a common derivation from the neural crest, whichis 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 ex- tent to which the general cutaneous component and general vis- ceral component, which are both represented in the spinal ganglia, THE CRANIAL GANGLIA IN AMEIURUS 300 are separated in the head. In the more generalized Xth the general cutaneous is intra-cranial and the general visceral extra- cranial, while in the more highly specialized region of the Vth and VIIth both are intra-cranial. THE ORIGIN OF THE LATERALIS Xtra 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 re- mains as the placode. The portion which disappearsis 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 gan- glion 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). Through- out the whole region of the overlap the placode is contributing 304 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 con- dition is duplicated in all my series in which the ganglion and pla- code overlap. ~In an embryo of 563 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 ques- tion as to whether the postauditory placode moves bodily away from the auditory vesicle or whether the placode at successive dis- tances from the vesicle represents a localized differentiation at those points has been variously answered’ Theanteriorend of theplacode seems to be partly converted into ganglion cells and partly to re- vert 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 pla- code. The lateral line organs which appear long after the p la- code has passed a given point appear approximately along theroute 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 de- tached 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 cellsto the ganglion the ganglion ends bluntly posteriorly and there is no strand of cells connect- ing the ganglion and placode. The trunk of the lateral line nerve of the X th grows out as a new structure from the ganglion. THE FATE OF THE POSTAUDITORY PLACODE AND THE APPEAR- ANCE 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 gan- glion 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 posi- tively 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 VIIth 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 ves- icle at this time is beginning to form the semicircular canals. The second organ, innervated by the ramus supratemporalis [Xth, 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 inner- vated 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 em- bryo 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 char- acteristic 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 le posterior to the large branch of the Xth nerve, while the placode les anterior to it. Owing to the fact that structures fre- quently 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 301 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 epi- dermis scarcely perceptible at first except by its location at the seat of the future organ. This thickening becomes more pro- nounced by the elongation of the deeper cells and by their assum- ing 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 elonga- tion 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 Platt on Necturus (95). The fact that anterior to the ear the preauditory dorso-lateral placode disap- pears 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 gang- lia 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. LANDACRE 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 fol- lows 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 founda- tion. This common foundation has the shape of a long furrow (ecto- dermic) 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 lat- eral line, the middle part becoming the auditory 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 (’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 ex- amination 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 onthesalmon. 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 differen- tiation, and in the transformation of its ill-defined anterior ex- tremity 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 lar- val life one or two sense organs are found in this region andit is extremely probable that they arise from the dorso-lateral 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 auditory vesicle to the preauditory placode, and to the postaudi- tory placode in part, we are in essential agreement, 7.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 pre- auditory 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 post- auditory 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 pre- auditory 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 nocommon 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 epider- mis on which the lateral line organs later appear (p. 28). There is a bare possibility that the relation of the anterior sen- sory ridge to the branchial sense organ as described by Wilson in Serranus is one of contiguity only and not of genetic relation- ship. If this be admitted, and it is the point about which Wilson is least definite, then there are no insuperable difficulties in har- monizing 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 or- gans arise separately and have nothing at all to do with the post- auditory placode. This conclusion was reached by Hoffmann (94) and by Miss Platt (’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 exten- sions the pre- and postauditory placodes, and, second, do thelateral 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 indivi- dual organs, the question arises as to which is the more primitive method. If the appearance of individual organs as in Ameiu- rus, each of which is homologous with sensory areas of the audi- tory 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 defi- nite 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. Itishard to see how organs that primitively come from a com- mon primordium could later in phylogeny arise singly. The dif- ficulty is increased when we remember that the lateralis VIIth 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 con- fusion 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. LANDACRE tory vesicle. If these organs arise as differentiations of sensory ridges which grow out from the anterior and posterior extension of the auditory 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 VIIth 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 563-hour embryo (A. melas) is taken just posterior to the point of contact of the visceral pocket with the epidermis. The placode is recog- nizable at this stage as a simple thickening of the epidermis ex- tending from the level of the dorsal border of the future gill slit down to the level of the branchial artery and longitudinally occu- pying 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 associ- ated with the placode or with the gill slit is not clear. The pla- eode 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 mesec- toderm. 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 de- fined mesial border can be seen and, while it could not be said that no cells slip individually from this placode into the mesecto- derm, 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 posi- tion 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 mesecto- derm and its boundaries can usually be located with no difficulty except at the posterior end which is ill defined, except in the ear- liest 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 sztu 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 ean benoticed. 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 gan- glionic mass and the epidermis with which it is still attached. The epidermis immediately dorsal to the attachment of the pla- codal ganglion has resumed its normal appearance, having a rough epi branchial! placode Ge: /_branch.arl a ge E medu//a _qud vesicle aud gan glion Fie. E. A camera tracing corresponding to fig. 56, showing the position of the epibranchial placode. NO. notochord; b.v. blood vessel. inner border composed of cells which seem to be in process of be- coming 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 gan- glionic 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 gan- glion 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 [Xth is first completely detached from the epidermis. This section is taken through the point where the ganglion lies nearest the epidermis, 7.e., near the anter- or 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 1X th as compared with the length of attachment (A. melas). | | Age:in: HOUPSL< dha a a a ees A | 69| 75! 81} 86] 93] 99] 105/ 113 Length of attachment to placode in | | | BECHONS thie c uta Cheat ee « a8 |) 2) le Re BOE OF oles. 0 Length of epibranchial ganglion in | | SEC ELONS Ree stare cys tpt S oie aret ss 3 | OV T2219 | 20% is |) 15 In the first series (69 hours) the mesially projecting mass is no longer than the thickening, but in the following series the gan- glion 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 re- sembles 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 gan- glion arose. 366 F. L. LANDACRE The introduction of the term ‘branchial sense organ’ by Fro- riep 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 Platt (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 post- auditory placodes to the specific lateral line organs finds a strik- ing 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-and postauditory placodesof Ameiurus), although they receive their innervation from ganglia derived from these placodes in the case of the X, [X 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 cesophagus 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 appear- ance. 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 super- ficial 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 the 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 113- hour embryo; a length of time sufficient to suggest tht the appear- ance 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 [Xth 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 especi- ally 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 VIIth nerve and the point of origin of the ganglion of the VIIth nerve precludes such an explanation. The suggestion offered above that there is some coordination in the growth of the peripheral fibres and the appéar- 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 evi- dence. 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 [Xth 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 dem- onstration. The fact that the VIIth, I[Xth and various portions of the Xth ganglia possess gustatory fibres approximately in proportion to the extent to which the epibranchial placodes con- tribute cells to these ganglia materially strengthens the conclu- sion. The geniculate ganglion of the VIIth 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, contain- ing only small contingents from the reduced fifth and sixth epi- branchial 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 cere- bral ganglia give rise to gustatory fibres which contain cells derived from the epibranchia] 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 epi- branchial ganglion acquires a connection with the proximal por- tion of the lateralis [Xth (figs. 58 and 59) before losing its connec- tion with the epidermis. However, the growing point of the epibranchial root is extremely thin. It passes up internal to the eardinal 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 [Xth 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 [Xth GANGLION After reaching the level of the auditory vesicle in its dorsal extension, the growing point of the root of this epibranchial gan- glion comesinto contact with a mass of cells derived from the pos- terior and ventral portion of the auditory vesicle, which later forms the lateralis ganglion of the I[Xth. It is not derived from 370 F. L. LANDACRE nor directly related to the Xth, nor is it at first in contact with the auditory ganglion proper, but arises separately and can be fol- lowed 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 [Xth comes into contact with the motor root of the [Xth 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 [Xth ganglion are: first, the cells derived from the epibranchial placode situated in the distal por- tion 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 [Xth 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 I Xth ganglion in its later stages. The early stages of the lateralis [Xth 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 pre- auditory lateral mass, and the ganglion would have a double composition and would resemble the VIIIth, if this ganglion con- tains lateral mass cells or the lateralis VIIth, rather than the later- alis 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 AMEIURUS oil 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 [Xth 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 illus- trated in fig. 638. The lateralis [Xth after it assumes a definite form is carried posteriorly by the backward growth of the audi- tory vesicle until it comes into the region of the slightly denser mass of mesectoderm, which at the time the lateralis [Xth 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,@), but in only one is it at all definite or does it present the appearance of an early stage of aganglion. I take it to be the mass of cells which other authors have described as neural crest cells that enter into the [Xth gan- glion. ‘ The early appearance of the definitive lateralis [Xth 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 [Xth 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). Se F. L. LANDACRE There is a well defined space, however, between the anterior end of the lateralis [Xth and the posterior end of the auditory gan- glion. The backward growth of the vesicle carries the ganglionic mass of the lateralis [X 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 [Xth or not. If they do, the lateralis [Xth 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 VIIth 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 VIIth, 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 long axis of the body, and is derived exclusively from the postauditory placode. Intermediate between these two extremes lie the audi- tory and lateralis [Xth, 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 specializa- tion, the ganglia and nerves appearing first in the head and extend- ing from there to the body, and having its center in the cranial region. The change from the generalized lateralis ganglion of the VIIth 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 [Xth ganglion, except that it forms its attachment to the brain in conjunction with the re- maining 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 [Xth 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 epibran- chial 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 (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 re- mainder of 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 rootofthisnerve, asmentioned above, lies lateral to the cardinal vein, as do all the other placodal rootsof the Xth. The appearance of the epidermis near the placode resembles closely that in the region of the [Xth. Anterior, dorsal and posterior to the placode the epidermis is usually one or two cells thick and quite ragged, presenting the appearance of pro- liferating cells into the mesectoderm. The thickening of the plac- ode 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 occa- sional 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 prolifera- tion of cells dorso-mesially to form the body and root of the gan- glion, and later the complete detachment of the proliferated mass from the epidermis. The attachment disappears last at the ex- treme anterior end in an embryo of 99 hours. The dorsal exten- sion of this placodal ganglion resembles closely the dorsal exten- sion of the placodal ganglion of the IXth nerve; but the fact that it grows back parallel to the dorsal extension of the fourth epi- branchial 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 mesecto- derm make it more difficult than in the case of the [Xth to deter- mine whether there are lateral mass cells in its composition. Gen- eral 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 epi- branchial ganglion of the [Xth. 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 pos- terior extension of the placodal ganglion has come into contact with the lateral mass ganglion situated just over the fourth gill sht 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 [Xth, 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 attach- ment to the epidermis. Fig. 69 is taken two sections posterior THE CRANIAL GANGLIA IN AMEIURUS BS 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, Hp. 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 com- position. It becomes completely detached from the epidermis in an embryo of 105 hours; while both the third and fourth epibran- -chial 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 ease of the fifth and sixth epibranchial placodes condi- tions 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 appear- ance 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 interpreta- tion 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 men- tioned above, however, I have been unable to separate this gan- glionic 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 sec- ond, 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 dis- appear some time before this, perhaps between the 105 and 113- hour 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 aaa 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 ease of the fifth epibranchial placode. The location, time of ap- pearance, 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 tran- sient 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 reduc- tion in the number of gills in the bony fishes as compared with cy- clostomes and elasmobranchs. The time at which the contact occurs and the length of the at- tachment of the [Xth 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 attach- ment. EPIBRANCHIAL PLACODES TIME OF FIRST TIME OF LENGTH OF TIME APPEARANCE | DETACHMENT | OF ATTACHMENT hours | hours hours Recondn eee oi en 563 | 93 373 SU ias eee diye Se ne a 69 99 30 Momnche oy never! sac aianiel 75 105 30 LTT 7 ee oo ee eee ee ey 93 = 119 90) Sinthis wr dete Rite oath otc 105+ | 113+ +-—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 PSYCHOLOGY, VOL. 20, NO. 4. 378 F. L. LANDACRE The condition here seems to throw some light on the interpre- tation given these placodes by former workers. They have so often been described as arising at the time the neural crest gan- glion fuses with the epidermis that I am unable to reconcile the conditions existing, under which the placodal ganglia of the [Xth and first two divisions of the Xth arise in Amelurus, with these descriptions, unless there are neural crest cells in the [Xth 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 erest 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 composi- tion. The presence of a neural crest in the region of the I[Xth 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 [Xth nerve in Menidia (Herrick, 99) and apparently in Gadus (Her- rick, ’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 Xtra 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 jugular ganglion thereis 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 IN AMEIURUS 379 fig. 73 which, however, is not taken through the largest portion of the ganglion. Herrick (’01) does not give a description of the com- munis 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 cesophageal rami of the vagus. He also calls attention to the fact that in Menidia 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 dis- tinguished 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 intracran- ially, and an extra-cranial visceral ganglion in which it is not pos- sible 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. LANDACRE 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 sep- arated 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 amass of cells surrounding the root of the Xth nerve and extending from near the medulla down around the root of the nerve to the extra- cranial portion. Preceding the stage of 69 hours, and particularly before the formation of a fibrillated root, it is not possible to dis- tinguish a definitive jugular ganglion from those cells which will form the extra-cranial portion. The history of the jugular gan- glion seems to be briefly as follows: First there is a loose mass of cells extending from the brain ventrally to a point where the epi- branchial 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 differenti- ated from the extra-cranial communis ganglion after 118 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 an- terior 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 reot. The extra-cranial portion of the lateral mass I shall not describe further in detail. It is so large and in- creases 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 com- ponents 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 cor- responds to the neural crest ganglion of other types, differentiated into two ganglia, a general cutaneous, the jugular, situated intra- _ eranially, 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. Un- like 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 con- tains beside the dorso-lateral placodes the homologue of the neural crest of other types, but the greater portion goes to form mesecto- derm. Of the components found in the adult ganglia, the special vis- ceral 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 VIIth come from the lateral mass. The auditory and lateralis [Xth come chiefly from the auditory vesicle but may have lateral mass cells in their composition. They show a some- what 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 gan- glia. The geniculate ganglion was seen to be composed of con- stituents 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 [Xth nerve seems to be a pure placodal ganglion. The first two epibranchial ganglia of the Xth resemble the [Xth, 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 Vi | VII | Vill IX x x x x COMPONENT | r | | | ( Pres- | Pres- Pres- General | ent | ent ent cutaneous | Dorsal | portion | n | ‘ ¢ : % | Pres- ? ? Acustico- | | . = | | ent lateralis _ 4 | a 2 | Dorso- | | | 3S! | | i . | lateral | Pres- Pres- Pres- Pres- Acustico- | placodes | | ent | ent ent ent lateralis eae, | ; 2 Ventral | Pres- | ? ? | Pres-| Pres- |General | . | | . | portion ent | ent ent jvisceral Epi- | | branchial | Pres- | Pres- | Pres-| Pres-| Pres-| Pres- |Special placodes ent ent ent | ent | ent ent visceral | | | In this table the general cutaneous, general visceral and acus- tico-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 later- alis Xth extends however much posterior to the visceral portion. A comparison of this tablewith 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. 3821) are broken ‘down in the table V into those portions derived from the epibranchial placodes, 1.e., the special visceral or gustatory ganglia, and the portions derived from the lateral mass, 2.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, 2.e., the lateralis VIIth and possibly portions of the VIIIth ganglion and of the lateralis [Xth, and those portions derived secondarily from the auditory vesicle and placodes, 2.e., all of the lateralis Xth and most if not all of the auditory and later- alis [Xth 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 dor- so-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 1s 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 constit- uents not being separable shortly after fusion. 4. Just anterior to the auditory vesicle a portion of the lateral mass gives rise to the lateralis VIIth ganglia which does not dif- ferentiate 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 VIIth, 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 [Xth 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 [Xth nerve, where a slight condensation of lateral mass cells is sometimes present, and it may possibly contain lateral mass cells. 7. Between the [Xth and Xth ganglia and for some distance posterior to the anterior end of the Xth, the lateral mass breaks down completely Jnto 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 com- pletely 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 pre- auditory 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 pro- liferation 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 epi- dermis, sometimes singly, sometimes two from a common pri- mordium 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 AMEIURUS 387 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 num- ber of mitotic figures. The thickening of the epidermis is fol- lowed 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 lying dorsal and posterior to the first gill slit. The thickening is accom- panied by active mitosis, proliferation of cells, and finally by com- plete detachment, en masse, of the proliferated cells. The pro- liferated epibranchial ganglion is in this case, however, apparently a pure placodal ganglion, since no lateral mass cells could be detected entering intoits composition. Itsdorsal extension comes into contact with the remainder of the [Xth 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 think- ing 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 placo- dal cells derived from the small fifth and sixth epibranchial plac- odes 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 Amelurus 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 VIIth comes entirely from the lateral mass and in other types apparently from the neural crest. The auditory and the lateralis [Xth 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 acustico- lateralis system as a special cutaneous system finds a strong sup- port in embryological evidence. The lateralis VIIth ganglia rep- resent an intermediate stage between the general cutaneous gan- glia, and the acustico-lateralis ganglia of the [Xth 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 plac- odes represented by the auditory vesicleand 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 Auuis, 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. ii, nos 23: BalLey, F. R. anp Miuuer, A. M. Text-book of embryology. Chapters 17 and 1909. 18, by Oliver S. Strong. New York. Baurour, F.M. Onthe 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. Lon- don. 1885. Comparative embryology. London. , BaRBIERI VON Crro. Ricerche sullo svillupo dei nervi cranici nei teleostei. 1907. Morphologisches Jahrbuch, vol. 37. Bearp, J. On the segmental sense organs of the lateral line and the morphology 1884. of the vertebrate auditory organ. Zodél. Anz., Jahrg. 7. 1885-86. The system of branchial sense organs and their associated ganglia in Ichthyopsida. A contribution to the ancestral history of ver- tebrates. 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. aw 1888. Morphological studies II. The development of the peripheral nervous system of vertebrates Part I. Elasmobranchii and Aves. Quar. Jour. Micr. Sc., vol. 29 (n.s.) no. 114. 1888a. 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 Ichthyop- sida. Part I. Raia batis. Zool. Jahrb., Abtheil. ftir Anat. uw. Ontog., Bd. 9. BeckwitH, Cora J. The early development of the lateral line system of Amia 1907. ealva. Biological Bulletin, vol. 14, Dec. Bunker, F.S. Onthestructure of the sense organs of the lateral line of Ameiurus 1897. nebulosus. Anat. Anz., Bd. 13, nos. 8-9. Cuapp, CorNELIA M. The lateral line system of Bactrachus tau. Jour. Morph., 1899. vol. 15. no. 2. Coxe, Frank J. On the cranial nerves of Chimzra 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. LANDACRE Cote, 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, Zodl. 7. Dourn, A. Bemerkungen iiber den neuesten Versuch einer Lésung des Wirbel- 1890. thierkopf-Problems. Anat. Anz., Jahrg. 5, Heft 2 u. 3. Froriep, A. Ueber Anlagen von Sinnesorgans am Facialis, ete. Archiv fiir Anat. 1885. und Physiol. Gorette, A. Entwickelungsgeschichte der itinke. Leipzig. 1875. GoronowitscH, N. Das Gehirn und die Kranialnerven von Acipenser ruthenus. 1885. Morph. Jahrb., Bd. 13. 1886. Untersuchungen tiber die erste Anlage der Kranialnerven bei Salmo fario. Société Impériale des Naturaliste, de Moscou, Tome 16. Harrison, Ross GRANVILLE. Ueber die Histogenese des peripheren Nerven- 17 6 | Systems bei Salmo salar. Archiv mikr. Anat. und Entw. Bd. 57. Herrick, C. Jupson. 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 si- luroid 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 Anfinge des peripherischen Nerven-Systems. Archiv f. Anat. u. Entwickelungsgesch. Horrmann, C. K. Zur Entwickelungsgeschichte des Selachierkopfes. Anat. 1894. Anz., Bd. 9, pp. 638. JOHNSON, ALICE J., AND SHELDON, LituiAN. Notes on the development of the 1886. newt. Quar. Jour. Mic. Sci., 5, No. 4, (n.s.), July, 1886. Jounston, J. B. Hind brain and cranial nerves of Acipenser. Anat. Anz., Bd. 1898. 14, no. 22-23. 1902a. The brain of Petromyzon. Jour. Comp. Neur. Psych., vol. 12. 19026. An attempt to define the primitive functional divisions of the cen- tral nervous system. Jour. Comp. Neur. Psych., vol. 12. THE CRANIAL GANGLIA IN AMEIURUS 391 JounstTon, J. B. Das Gehirn und die Cranialnerven der Anamnier. Merkel und 1902c. Bonnet’s Ergebnisse. Bd. 11. 1905a. The morphology of the vertebrate head from the viewpoint of the functional divisions of the nervoussystem. Jour. Comp. Neur. Psychs, vole 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. Kourzorr. Entwickelungsgeschichte des Kopfes von Petromyzon planeri: Ein 1902. Beitrag zur Lehre tiber Metamerie des Wirbelthierkopfes. Moskau. KupFrFrer, C. von. The development of the cranial nerves of vertebrates. Jour. 1891. Comp. Neur. Psych., vol. 1. 1894. Studien zur vergleichenden Entwickelungsgeschichte von Ammo- coetes planeri. Mtinchen und Leipzig. : LanpacreE, F. L. On the place of origin and method of distribution of taste 1907. buds in Ameiurus melas. Jour. Comp. Neur. Psych., vol. 17, OO, Ue Linuiz, F. R. The development of the chick. New York. 1908. Locy, Wm. A. Contribution to the structure and development of the vertebrate 1895. head. Jour. Morph., vol. at no. 3. MarsHALL, A. M. On the early stages Oi 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. Sev. Se LAY, I CES —~ 1881. On the head centefs and associated nervesin 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 Seyllium. Quar. 1881. Jour. Mic. Sci. MirropHanow, P. Etude embryogénique sur les Sélaciens. Arch. de Zool. 1893. Expér. et Gen., sér 3, t. 1, pp. 161-220. pl. 9-15. 1891. Sur la formation du systéme nerveux périférique des vertébrés. Comptes Rendus, Paris, Tome 113, no. 19, pp. 659-562. Nea, H. V. The segmentation of the nervous system in Squalus acanthias. 1898. Bull. Mus. Com. Zoél. Harvard College, vol. 31, no. 7, Onopi, A. P. Ueber die Entwickelung des sympathischen Nerven-Systems. 1886. Archiv. fur Mikrosk. Anat., vol. 27. av€ it 392 F. L. LANDACRE PouuarD, 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. Puart, Junta 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. RaBu, 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 occipi- 1904. tal region of the human embryo. Amer. Jour. Anat. Srrone, O. 8S. The cranial nerves of the amphibia; a contribution to the mor- 1895. phology of the vertebrate nervous system. Jour. Morph., vol. 10. Wuitman, 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. WisuHeE, J. W. Van. Ueber die Mesodermsegments und iiber die Entwickelung des 1882. Nerven des Selachier-Kopfes. Amsterdam. 1886. Ueber die Kopfsegments und die Phylogenie des Geruchsorganes des Wirbelthiere. Zool. Anz., Bd. 9. Wiuson, H. V. The embryology of thSsea bass (Serranu satrarius). Bull. of 1891. the U.S. FishCom. Vol. 9, Washington. Wiuson, H. V. anp Marrocks, J. E. The lateral sensory anlage in the salmon. 1897. Anat. Anz., Bd. 13, no. 24. Wriaut, R. R. On the skin and cutaneous sense organs of Ameiurus. Proc. 1884a. Canadian Institute, n.s. vol. 2. 1884 b. On the nervous system and sense organs of Ameiurus. Jb7d. ABBREVIATIONS Au. G., Auditory ganglion. Ep., Epidermis. Au.G.,TandII. Firstandseconddivi- Ep. Pl. 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. Pl. II, Second epibranchial placode chial nerves. (epibranchial placode of the IX En., Endoderm. nerve). D. L. VII, Dorso-lateral portion of the Ep. Pl. III, Third epibranchial placode lateralis VII ganglion. (first epibranchial placode of the D. L. M., Dorso-lateral mass. X nerve). THE CRANIAL GANGLIA IN Ep. Pl. IV, Fourth epibranchial placode (second epibranchial placode of the X nerve). Ep. Pl. V, Fifth epibranchial placode (third epibranchial placode of the X nerve). Ep. Pl. 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. III, 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 Pl., Postauditory placode. Pre. Pl., Preauditory placode. R. X, Root of the X nerve. R. Au., Root of the auditory nerve. R. C. D. X., Ramus cutaneus dorsalis vag. AMEIURUS 393 R. D. L., VII, Root of the dorso-lateral portion of the lateralis VII gan- glion. R. Gass., Root of the Gasserian gan- elion. R. Gen., Root of the geniculate gan- glion. R. L. G. X, Lateralis root of the X nerve. R. S. J, Ramus supra-temporalis J, glossopharyngei, (Herrick, ’01). R. V. L., VII, Root of the ventrolateral portion of the lateralis VII gan- glion. A. A., Ramulus acusticus ampule anterioris. A. E., Ramulus acusticus ampullee extrene. L., Ramulus acusticus lagene. N., Ramulus acusticus neglectus. P., Ramulus acusticus ampulle pos- terioris. Sac., Ramulus acusticus sacculi. N., Ramulus acusticus recessus utriculi. . Gass., Trunk of the Gasserian gan- glion (supero-lateral strand of Wright). T. Gen., Trunk of the geniculate gan- ganglion of the VII nerve (in- fero-mesial strand of Wright). T. V. L. VII, Trunk of the ventrolateral portion of the lateralis VII gan- glion (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. Hope wPRR RP 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 4mm. eye-piece, X 8 Spencer, reduction to 3. Total magnification, 186 dia. 1 (A. neb., Stage I) is taken at a stage when the medullary plate is slightly dif- ferentiated into the neural cord (N. C.) and the lateral mass (L. M.). The sec- tion 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 liesin 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, twenty- nine 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 post- auditory 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 regionin whichthe Gasserianganglionforms. 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 ves- icle. It shows the reduction in size of the preauditory placode as one reads for- ward from the position of fig. 4. THE CRANIAL GANGLIA IN AMEIURUS 395 WONT Sy Sees ©) COG 396 F. L. LANDACRE EXPLANATION OF FIGURES 10 and 11 (A. neb.. Stage V) arefromthesameembryo. Fig. 10is 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 hes 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 lat- eral 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 preaudi- tory 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 ‘s 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 ex- tends 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 disappear- ance 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 (XY) 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. 397 AMEIURUS LN] THE CRANIAL GANGLIA — — —Gass.G. ———--Mesre E @ (ie Mrsree-- —-“S= ee — és A : <3 too} Fe hE [ane eae 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 por- tion 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 28 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 whichis 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 VIl 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 p26 9 O90 B HG 0,5 9- HO) <6 sand AO? gs 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 [X 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 P vic, hw R.D.L vu--- = --—--D.Lvit 7 MO ti — — —-—B.Gen. Vs: 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 [X 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 disappear- ance 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 grad- ually 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 be- yond 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 twenty- seven 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 the epibranchial 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 pos- terior end of the placode is detached from the epidermis and has formed an attach- ment 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 gan- glionic 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 AMEIURUS 403 § ees ‘ ‘yf f N XY. 70} ie 4 D, 4 Mrsec-———— Ed = Q & iT a. ¥ OC = eto: x 404 F. L. LANDACRE EXPLANATION OF FIGURES 57 is taken onesection posterior to fig.56at 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 [Xth 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 stagein 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. 64 (A. melas) 69 hours shows the earliest recognizable trace of the third epi- branchial 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 epi- branchial placode. This probably is not the earliest trace of the fourth epibran- chial placode, however. The preceding series is defective at this point. THE CRANIAL GANGLIA IN AMEIURUS 405 ve 0) =r ern as ES 406 F. L. LANDACRE EXPLANATION OF FIGURES 69 (A. melas, 93 hours) is taken through the detached portion of epibranchial ganglion IV, also shows the root of epibranchial ganglion ITI. 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 com- munis 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. III appear to be the posterior extension of this ganglion. 71 (A. melas, 93 hours) shows the point of union of the fourth epibranchial gan- glion with the lateral mass of the X (Gen. vis. X). 70, 72, 73 and 74 illustrate the conditions under which the fifth and sixth epi- branchial 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 com- munis 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, 118 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 AMEIURUS 407 SE o.C. D2. 2——Gen Com X 408 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 agiven 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 longitudi- nal 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, 81 hours; only three epibranchial placodes are present at this stage. 80is areconstruction 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 ge meral soma a general viseera/ \\ special somali THE JOURNAT. OF COMPARATIVE NEUROLOGY 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, 174 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 Att 82. ASCs Rdix----- -~--- -Gex Com x EeG wiz yi Se y Br ee) 83. Gass: GSS . #--TVLvir general SO/N asic gen era/ yiecera/ = specia/ somayi 7e special yisces al Hh TG Ass-7 ey 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 OF 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, 23, 5, 10, 15, 20, 25, 30, 35, 40, 45, 50, 60, 70, 75, 80, 90, 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 4, +4, ,'), ete. natura! slze. i 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 3, 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 4, it must be made 13 times the size of the engraving. 4. If the drawing is to be reduced 4, it must be made 14 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—eacb line of the drawing must be 40 times the length of the corresponding line in the object. (2) For 5 reduction—each line must be 80 times the length of the corresponding line of the object. (3) For 4 reduction—each line must be 60 times the length of the corresponding line of the object. (4) For 4 reduction—each line must be 50 times the length of the ov corresponding line of the object. HOW TO MARK A DRAWING TO BE REPRODUCED. (1) Full size (i.e., no reduction), (2) 4 off; (3) 4 off; (4) 4 off; (5) + off. Bs) THE MORPHOLOGY OF THE FOREBRAIN IN AMPHIBIA AND REPTILIA C. JUDSON HERRICK From the Anatomical Laboratory of the University of Chicago EIGHTY-FOUR FIGURES | (aK RrCOXS WUKGUVON OT SMEs c 3 See c crordislers tae Aenea a Cee Mee ceo cmc. circhc Shae s Soe aemeayo oeatee 413 J NASH OLN OTN Seve eae ex Sipe a thea ee ee Seth Red a Dh 415 HU) codes eph ere te aD O aces les! 3 soe ch SESE EE Ot ok ee 416 PATTANIOLAVATE OER URE eR oe 208 502 ack och 2 poe ER aS Se Ne ae . AIG (Oe ere OR H0v8 Vol Cee to er cer hee ca Rae ee 435 ANTM oa8 b BAP ete aslo ae ee ene ce ag clan ureee: 438 Comparisons switin Hsnese: ay. pte at sins 5s, «sa bier o eer ee oe se 449 Ee Patra ice were ce Bae ea eI L's. ss bgdgh wach ee eR ee 452 DISCUSSION Heo Wa ee TNE Gober is Sa abe ey oe DONE Ee ee: 466 Diencephalon;and.telencephalan: ............... 2 ce eee eee aes =: 466 The paraterminal body..... hs RD Heel Le nd US bh ok eee Ae aE A) 483 ihemorphology of. the cerebral cortex. ...<..3.09... meter eee te: 486 ‘Phe subdivision ofthe prosencephalon...! ....’- mec seeerciee: o.oo eee 492 SUAS TEED UT CAPR HAE MoO ate PR POU IR oe SSI Sl. oleh a ee 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 disappoint- ingly meager. We have, however, now reached a point where effective correlation has begun to take form and within the be- wildering 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 sim- plicity of this pattern Johnston has recently presented strong evidence that the tel- encephalon 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 com- prising 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 ven- tricle, 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 con- stitute 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 phy- logenetic) for more and more of the telencephalon medium to be evaginated through the interventricular foramina into the hemi- spheres. 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 contribu- tion of further data for the solution of this problem, I have ex- amined 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 cere- bral 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 recog- nizable 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 dien- cephalon. 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 modifica- tion 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 Embry- ological Collection, was prepared by Mr. P. 8. McKibben of the Department of Anatomy, University of Chicago, to whose kind- ness and skill Iam greatly indebted. I have also examined a se- ries of cross sections through the head of Petromyzon (Ichthy- omyzon 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 sig- nificance 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 Cc. JUDSON HERRICK eminentia septalis, ganglion basale, formatio 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, 799, 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 intermedius on the ventricular side and fissura arcuata on the superficial side. For reasons which will appear in the sub- sequent 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 lim- itans (lateralis et medialis), and the zona limitans may be present as a useful landmark in cases where the corresponding fissure is not externally evident. URODELA Amblystoma 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 by 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 hemis- phere 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 ependy- male 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 dor- sal 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. 0. 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 inter- rupted by the interventricular foramen and from this point cau- dad the ventriculus lateralis communicates widely with the ven- triculus impar of the forebrain (fig. 2) The wide membranous ‘roof over the ventriculus impar is a choroid plexus which is ex- tended laterally to form the plexus lateralis. Farther caudad this membrane is evaginated dorsally to form the paraphysis and back- ward into the third ventricle as velum transversum and dienceph- alice 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 for- ward from this point are three separate membranes: (1) the forebrain roof and the plexus lateralis already described; (2) the paraphysis (figs. 2 and 3); (8) the plexus chorioideus ventriculi tertii or diencephalic plexus. For further details regarding these membranes I refer to the excellent account of Necturus by War- ren (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 Cc. 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 dorso- medialis 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 olfacto- rius dorso-lateralis follow the whole length of the pars dorso- lateralis and reach the posterior pole. The basal forebrain bun- dle 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 compo- nents 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 prepara- tions by the presence of much coarser fibres in the former (fig. 3). The medial forebrain tract connects chiefly with the hypothala- mus, the lateral with the thalamus and mid-brain. Both de- cussate 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 pass- ing rostrally of the interventricular foramen and anterior com- missure between the primordium hippocampi and the pars ventro- medialis. Farther back this tract contains fibers for the com- missura pallii anterior (com. 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 con- verge at the posterior pole, where their distinctive character- istics 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 trans- versum passes over into the diencephalic plexus and, farther laterally and ventrally, into a ridge which forms a massive con- necting 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 ex- tends caudad through the whole length of the diencephalon, turn- ing ventrally behind to join the sulcus limitans in front of the tuberculum posterius. As we shall see beyond (pp. 481 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 Mon- roi. Kupffer (06, p. 181, fig. 193) designates in Necturus the suleus ventralis as sulcus Monroi, but this is evidently inap- propriate, for this ventral sulcus separates the hypothalamus from the eminentia thalami (see pp. 431 and 469 ff.). Immediately be- hind 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 hypo- thalamus. 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. Itis the sulcus dience- phalicus dorsalis. These relations come out much more clearly in older larvae and adults after the straightening ‘of the diencephalic flexure. 420 Cc. 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 transversum (figs. 6 and 7) illustrate clearly the relations of the velum (7) to the eminentia thalami and adjacent parts and the communication of the diencephalic and telenceph- alic 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 exam- ined 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 im- portant fibrous connections between these dorsal parts of the telencephalon and diencephalon all cross the dorsal barrier inter- posed 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 aot 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 lat- eral 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 gran- ular 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 gran- ular layer to accumulate on the median border of the hemi- sphere (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 tne 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. Far- ther 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 unmed- ullated 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. Ina 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 medullated 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 ven- tral division is mingled with the median forebrain tract. The tractus olfactorius dorso-lateralis arises from the whole length of the dorsal border of the olfactory bulb. The total number of medullated fibers is, accordingly, quite large (figs. 8 to 11). These medullated fibers are, however, all short, ending in the adjacent gray of the nucleus olfactorius anterior and pars dorso-lateralis of the hemisphere. The medullated 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 ac- cessorius 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 com- missure, 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 olfactortum of mammals. The dor- sal wall remains undifferentiated and is continued caudad into the pars dorso-lateralis of the hemisphere. The latter is, accord- ingly, 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 cau- dad to the posterior pole has the same structure and fiber con- nections as in the frog. Its cells are not arranged, as else- where 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 sur- face and descend to the ventro-medial angle, where they turn caudad and accompany the ventral forebrain tract to the hypo- thalamus (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 medul- lated 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 ventro- medialis 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 associa- tion 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 ventro- lateralis hemisphaerii. This contains, in addition to the tractus olfactorius ventro-lateralis already referred to, medullated and unmedullated ascending and descending fibers between the hemi- sphere and the thalamus, the lateral forebrain tract. A cross-section taken behind the olfactory bulb of adult Am- 494 C. JUDSON HERRICK blystoma resembles closely a corresponding section of the frog, save for the absence here of 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 me- dianus septi, becomes thinner (fig. 13) in the site of the larval septum ependymale, and for about 100 micra immediately ros- tral 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 proc- ess, 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 ventro- medialis is reduced in size and no clearly defined nucleus medi- anus 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 compo- sition 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 nu- cleus 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 hip- pocampi 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 “cor- pus callosum” of Osborn and his followers. The ‘dorsal commissure” and the ventral or anterior commis- sure are not so clearly separate in urodeles as in the frog and com- parison 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 hemi- sphere 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 imto 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 les dorsally of the interventricular foramen, but ventrally of the sulcus medius, 7.e., in the pars ventralis thalami, and it occurs in the same relations in Anura also. In Anmiota 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, ete. For the relations of A 426 Cc. JUDSON HERRICK these plexuses see beyond, p.431). Here in the eminentia thal- ami, 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, bear- ing a few medullated fibers (fig. 15) connected with the primordium hippocampi, divides into two parts, each with medullated 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 ventro- lateralis hemisphaerii (striatum complex). Both subdivisions have unmedullated connections with the posterior pole. Far- ther back (figs. 18 and 19) the stria medullaris has an unmedul- lated connection with the eminentia thalami and a strong medul- lated 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 media- lis reaches as far forward as the pars ventro-medialis hemisphae- ri 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 hippo- campi by way of a decussation in the commissura pallii anterior. These fibers may be the same as the ones which [ 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 fic. 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 cor- tico-habenularis medialis and tractus cortico-thalamicus. The hippocampal commissure is divided into two parts, the commis- sura pallii anterior and posterior. 2. The commissura pallii anterior includes commissural fibers between the dorso-medial parts of the hemispheres and decus- sating fibers of the tractus cortico-habenularis medialis and cor- tico-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 Cc. JUDSON HERRICK 5. The tractus olfacto-habenularis system. Primitively, as in eyclostomes, 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 decus- sating 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 prin- ciple. There are five components (paragraphs 6 to 10 below) of this system, of which the largest comes from the nucleus pre- opticus 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. 482) 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 hippo- campi directly into the stria medullaris. Some of its fibers prob- ably 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 FOREBRAINg 429 and apparently reach the pars ventralis thalami above and behind the optic chiasma. 104. The transito1y posterio1 chorioidal fold is seen at z and the first rudiment of the plexus lateralis at y. MORPHOLOGY OF THE FOREBRAIN 509 Pars dors. that Leminentia thalam; a CStria medullaris primordium Ecol lat : ay faenia Wy thal et forn. Yy 4. ; UG : ; A Sat ft b. if n.olfactorius THE JOURNAL OF COMPARATIVE NEUROLOGY AND PSYCHOLOGY, VOL. 20, No. 5 510 C. JUDSON HERRICK EXPLANATION OF FIGURES 7. YS pep. 47 Ger. dorsal sac r.po. qypAop.ch. CRD hypophysis 50 528 C. JUDSON HERRICK EXPLANATION OF FIGURES 51. Section 402 from the same series as fig. 50, taken 11 sections from the sagti- tal plane. X 15. 52. A parasagittal section through the brain of a specimen of Lacerta vivipara (?) with a total length when uncoiled of about 36 mm. (head 5mm. long). Har- vard Embryological Collection, series 603, section 201. X 24. On this figure is indicated the approximate planes of figs. 53 to 57. 53 to 57. A series of sections through the brain of a specimen of Lacerta vivi- para (?) of about the same age as the one last figured. Harvard Embryological Collection, series 604. X 24. For planes of sections, see fig. 52. 53. Section through the anterior commissure. 54. Section immediately ventrally of the interventricular foramen. 55. Section through the interventricular foramen. MORPHOLOGY OF THE FOREBRAIN 529 THE JOURNAL OF COMPARATIVE NEUROLOGY AND PSYCHOLOGY, VOL. 20, NO. 5. 530 Cc. JUDSON HERRICK EXPLANATION OF FIGURES 56. Section through the fimbria and nucleus of the commissura pallii posterior. 57. Section through the commissura pallii posterior. 58. Horizontal section through the commissura pallii posterior of a 25.2 mm. em- bryo of Sphenodon. Harvard Embryological Collection, series 1490, section 341. X 29. 59. Section 347 of the same series as the last, taken 70 micra farther dorsally. MORPHOLOGY OF THE FOREBRAIN 5031 GA str. med. a 532 Cc. JUDSON HERRICK EXPLANATION OF FIGURES 60. Cross section through the lamina terminalis and precommissural body of an advanced larva of Anguis fragilis. 75. After Kupffer (’06, p. 234). The refer- ence letters are those of the original. - 61. A transverse section through the brain of Phrynosoma cornutum (Harlan). Haematoxylin preparation. X 30. The section passes immediately rostral to the foramen interventriculare, the left side being slightly farther caudad. MORPHOLOGY OF THE FOREBRAIN 533 4 Sey ee 60 ortex dorsalis ortex lateralis a str iatilm: : . hippocampi parctiatds: a eee n. med.s. Perr a 61 534 C. JUDSON HERRICK EXPLANATION OF FIGURES 62. A section from the same series farther caudad. x 30. On the right side it passes through the interventricular foramen, of the left immediately caudad of it. 63, 64, and 65. Three horizontal sections through the brain of the lizard, Sce- loporus undulatus (Daudin). Haematoxylin preparation. The right side is slightly farther ventral than the left. X 20. 63. Through the dorsal part of the hemispheres. 64. On the right side the section passes through the interventricular foramen; on the left side above it; the precommissural body surrounds the foramen. MORPHOLOGY OF THE FOREBRAIN 535 See “siriatari eh, : iimoretiu fig Wiclempr \corpu - precommis: urale pramen, ee culare \ =... Z S commis. superior 536 Cc. JUDSON HERRICK EXPLANATION OF FIGURES 65. Immediately ventral to the interventricular foramina, illustrating the way in which the precommissural body forms the “‘bed’’ of the commissura pallii an- terior and commissura anterior, the latter crossing immediately ventrally of the point designated. 66 to 70. Sections from a transverse series through the brain of young Alligator mississippiensis about 25 em. long, stained by the method of Ramén y Cajal. Glo. 66. Through the caudal part of the septum, illustrating its medial and lateral nuclei and their relations to the primordium hippocampi. MORPHOLOGY OF THE FOREBRAIN 537 cortex dorsa cortey lateralis -7 / zz cortex dorsalis . ‘-¢ortex dorso-medialis 538 Cc. JUDSON HERRICK EXPLANATION OF FIGURES 67. Through the anterior commissure and commissura hippocampi. 68. Through the interventricular foramina. The nuc’eus of the commissura hippocampi (n.c.h.) is continuous with the nucleus of the tractus cortico-habenu- laris (tr.c.hab.) 539 MORPHOLOGY OF THE FOREBRAIN we atharine H uieeas Do Retioung gee eS = ” alis “medial -and ° m1) ‘g: : ees fo) nee ly a 3 ey, 2 ie = < av? °o [-¥) Zh aS os os 8 x tpg Oe 2 x PLY | ae ag w ra} SZ - 3 i's U i — chiasma 68 N. po. - 540 C. JUDSON HERRICK EXPLANATION OF FIGURES 69. Through the rostral end of the diencephalon. 70. Through the middle of the diencephalon, including the rostral end of the massa intermedia. MORPHOLOGY OF THE FOREBRAIN 541 cortex dorso-mediatis_-~_ 5.0 a -Paraphysis. J oe ans: Bas fi mbri ee-par gM. po. optic chiasma dorsal sac AH \ OCS habenula paraphysis pars dorsalis thalami 542 Cc. JUDSON HEPRICK EXPLANATION OF FIGURES 71. Diagrammatic cross section through the brain of young Echidna, after Elliot Smith, ’03, p. 468, fig. 16. a, dorsal part of the paraterminal body. 8, sulcus limitans hippocampi y, ventral part of the paraterminal body. 6, fissura hippocampi. 72. Diagrammatic median section of the brain of a hypothetical vertebrate ancestor, showing the probable relations of the rostral end of the neural tube be- fore the evagination of the optic vesicles and cerebral hemispheres. The site of the tissue which gives rise to the optic vesicle is shown at R, to the olfactory bulb at O. (See the text, pp. 468, ff.) MORPHOLOGY OF THE FOREBRAIN UY —— : — SS} —=—_ —————— CV 71 tectum :com post. ae : epiphysis 5 -5.d. . Cok, sup. <7. A sac. d. lamina. dorsalis aN hab... é P c.gen fat. Sy. ( Ot Sulcus Limitans Wi 3 ELE . Jamina ventralis chiasma opt." Ue 543 544 Cc. JUDSON HERRICK EXPLANATION OF FIGURES 73. A diagram similar in plan to that of fig. 72, reconstructed from actual sec- tions of the cyclostome brain, Ichtnayomyzon concolor. It is based on the same selies of cross sections illustrated by figs. 74 to 8! and the planes of the sections figured are indicated on this figure. (See the text, pp. 470, ff.) 74 to 81. A series of cross sections through the forebrain of a specimen of Ich- thyomyzon concolor (Kirtland) 120mm. long. Drawn from haematoxylin prepar- atious made by Dr. Charles Brookover. X 30. The plane of each of these sections is indicated on fig. 73. The series of sections is numbered, beginning with no. 1 at the lamina serminalis and the serial number of each section figured is given in the following descriptions, the sections being 15 w thick. Cf. the series of cross- sections of Lampetra figured by Johnston, ’02. 74. Section no. 3, immediately behind the lamina terminalis. 75. Section no. 4, through the recessus preopticus. 76. Section no. 7, through the rostral end of the optic chiasma. 77. Section no. 13, at the rostral border of the interventricular foramen. 78. Section no. 17, immediately caudad of the foramen and passing through the probable site of the embryonic velum transversum. 79. Section no. 23. This section is wholly diencephalic except the lateral hemi- spheres. The dorso-median ridge here attains its greatest size. MORPHOLOGY OF THE FOPEBRAIN 545 8! i 19 1811 1b 14 tlt as i i { sr i i 546 Cc. JUDSON HERRICK EXPLANATION OF FIGURES 80. Section no. 29, through the caudal border of the chiasma ridge. The dorso- median ridge and the pars dorsalis thalami cannot be clearly separated at this level. 81. Section no. 41, through the middle of the habenulae. 82. View of the inner surface of the brain of ahumanembryo of 6.9mm. Drawn from the wax model by Ziegler of the embryo Br 8 of the His series and lettered after His, ’04, p. 56, fig. 34. The region marked C.s. includes not only the corpus striatum but also the preoptic nucleus and other parts of the 1hinencephalon. (1) Maigo reuniens (rhinencephalon with rhinencephalon and pallium). (2) Margothalamicus(pallium with +halamus, site of the di-telencephalic fissure). (3) Margo peduncularis (corpus striatum with thalamus, site of the sulcus dien- cephalicus medius). (4) Margo hypothalamicus (corpus striatum with hypothalamus). 83. Diagrammatic cross section through the diencephalon of Urodela and Anura. On account of the diencephalic flexure, the section must be taken obliquely to the long axis ot the brain in the plane A-B of fig. 22, in order to pass transverse to the thalamic axis. The numbers 1, 2, 3 and 4 and the letters A, B and C mark cor- responding structures in figs. 83 and 84, the two figures being designed to illus- trate the way in which the cerebral hemispheres have been formed by the lateral evagination of the walls of the neurai tube; see the text. p. 477. A, sulcus diencephalicus dorsalis. B, sulcus diencephalicus medius. C, sulcus diencephalicus ventralis. D, roof plate. V, floor plate. 84. Diagrammatic cross section through the cerebral hemispheres in front of the lamina terminalis of the frog. A, dorsal angle of hemisphere. B, zona limitans lateralis and fissura endo-rhinalis. C, ventral angle of hemisphere. D-V, zona limitans medialis and fissure hmitans hippocampi. L.F.T., lateral forebrain tract. M.F.T., Medial forebrain tract. ae THE ROLE OF VISION IN THE MENTAL LIFE OF THE MOUSE KARL T. WAUGH From the Harvard Psychological Laboratory TEN FIGURES CONTENTS ie Problems imethousvand mesulitsee.: sess. «cla cee. sein vette ene nae een 549 Problem 1. Discrimination of light intensity .... 550 MX, Whavelereatavo lunar nilllomomnimneni@in es es od sooo anor aedcncocdesdooeose 550 ldpgoeronmarerms, (Gh) IBilevelie eyavel WINK... 05. 55k code des sacnsenee 552 Experiment (b) Light and dark varieties of a color......... 552 Experiment (c) Influence of background................... 553 Experiment (Gd) selechomoh yannsSacsse eae eee OOD IB, (Uline kop oliineevere mllllwaaauiineHNG@Iy, 52555 oa565densannassovccusevcudase- 557 Problem yam olondisenimmnatlOnn esas ae enna er eer 560 INS Whavolere thavelieevoth WN livia ks cecceanetacc acc eo scbcasesenbon 560 Experiment (a) Discrimination of colored objects.......... 560 Experiment (b) Selection of colored yarns ................ 564 Bae Underidinectwllluminaitaoniee eee eee es eens eee 566 Rroblemrse = Lerce pti OnE Otero nile i - ese ee ee eens 570 Problem 4. Perception of distance of object from animal........... 572 Problem 5. Perception of third dimension in objects............... BYE Dee Bino cular stom o-4 Wiviss lee Se ac ona eee ees 581 III. Kinaesthetic sensations, the guide to movement....................... 590 NEES tructuneroiihereyerol the amouses 1. ance eee en eee eee: 595 AES. AUTO OOIAY.& lected be oan ee Le ET he Sees cA aa eT ECO Lie 598 I. PROBLEMS, METHODS AND RESULTS The purpose of the investigation which is described in this paper was to answer the general question—What does the mouse receive from the outer world through the sense of vision, and of what importance in its life are the visual data so received? The experimental work was done in the Harvard Psychological Laboratory between January, 1905, and March, 1907. I am THE JOURNAL OF COMPARATIVE NEUROLOGY AND PSYCHOLOGY, VOL. 20, No. 6. 550 KARL T. WAUGH indebted to Professor R. M. Yerkes, under whose immediate super- vision the work was carried on, for the suggestion of the problem and for much helpful criticism and advice throughout the course of the work. I wish also to make acknowledgment to Professor Miinsterberg and Professor Holt for their advice and suggestions. The anatomical portion of the work was done in the Museum of Comparative Zodlogy at Harvard under the direction of Profes- sor G. H. Parker, to whom I am indebted for aid and advice. The general method pursued was that of presenting to the mouse a choice between two conditions, one being agreeable to the animal and the other disagreeable. In the majority of cases these were food and a slight electric shock. PROBLEM 1. DISCRIMINATION OF LIGHT INTENSITY A. Under indirect illumination In experiments (a) and (b) under this head, five animals were used. MOUSE COLOR SEX AGE DURATION OF EXPERIMENT INO RSE eer va hotest. c ack gray male 3} months Mar. 18—Apr. 17, 1905 INO MR ated or! Bertani ce black female 33 months | Mar. 23-Apr. 27, 1905 INO FRA rey 2 osc Morty hs black male 2months | Mar. 23—Apr. 27, 1905 DAs oo co ae ele te male 3months Jan. 21-31, 1906 Oren amr: oer || LOW male 3 months Jan. 17-26, 1906 1 All the white mice are albinos. APPARATUS: This consisted of a wooden box measuring 52 em. x 40 em. on the inside, and 18 cm. deep. At one end of the box was a small opening fitted with a sliding door, which, when lifted, permitted the mouse to enter from the nest box N (fig. 1). At the other end of the experiment box were arranged two round tin boxes X and Y, each measuring 43 cm. in diameter, covered with papers which differed in brightness, in color, or in both. The boxes were fitted into small wooden mounts fastened to two boards (224 em. x 184 em.). On these boards wires were placed, as is shown in fig. 1. These wires were then connected with a Porter inductorium. By closing the key, K, the experimenter could give a mouse whose feet rested upon two adjacent wires a slight shock. VISION IN THE MOUSE 551 Merxop: Food was placed in one of the boxes and the electric current was switched into the wires on the board which supported éO) rae —— | ot Fie. 1 Apparatus to test visual discrimination in the mouse. JN, nest box; X and Y, food boxes; 7.C. inductorium; C, dry cell; K, key. the other box. In this way inducement was given to cause the animal to make use of what discriminative power it might possess for the purpose of avoiding. the shock and obtaining the food. This, of course, presupposes the possibility of the formation of an association of shock with one intensity of light and of food with the other. The food used in this experiment was a small quantity of “‘force,’’ which has little or no odor to enter as a disturbing factor. An animal was placed in box N and the door was lifted. The Dae KARL T. WAUGH mouse would enter the experiment box and make a choice of X or Y, and after receiving a morsel of food or a shock, as the case might be, it would run or be driven back into box NV. The boxes X and Y would then be interchanged and the current switched into the wires on the other side. When all was in readiness, a door between N and the main box was raised and the mouse was per- mitted to seek the food again. It was considered a choice if the mouse touched the edge of either food-box. If he approached the wrong side first and received a shock and then ran over to the other side to get food, it was recorded as a wrong choice and the animal was forced toreturn to box NV before making the next choice. Twenty trials were made each day with each mouse, and each choice was recorded as right or wrong according as the mouse obtained the food or the shock on first running out. Resuuts: The following tables give the number of trials and the number of right and wrong choices: Experiment (a) Black and white. Black and white papers were pasted on the food-boxes. MOUSB FOOD IN NO. TRIALS RIGHT WRONG CHOICES CHOICES DD ere eee eee myvuive 100 73 27 (see curve, fig. 2) OMAR es tie cies white 100 83 17 (see curve, fig. 2) Experiment (b) Light and dark varieties of a color. Light and dark violet papers were substituted for the black and white. MOUSE FOOD IN | NO. TRIALS AMEHEU WRONG | CHOICES CHOICES eek a | SS a a INIG)S. diesotevenorce Light violet 370 252 118 (see curve, fig. 2) INO2 ees Light violet SO 42 38 INOW4 eee Light violet 230 152 78 (see curve, ‘fig. 2) CyrEck EXPERIMENT: In order to make sure that the animals were using the visual sense in discriminating one paper from VISION IN THE MOUSE 553 the other, and not the sense of smell, check tests were made. With a mouse that seemed to give good evidence of discrimination, the food was placed in the darker box after having been in the lighter for earlier choices. In the case of mouse no. 1, after a training of 370 trials favoring the lighter violet, and after the boxes had been washed and clean papers pasted on them, the resulting choices were, in ten trials: light, violet (no food)... 222... ..- 8... Darkivaolet (food)iG ee. 4... 2 In fig. 2 are given the error curves for the discrimination of inten- sity. The ordinates represent the number of wrong choices in each twenty trials given on a particular day. The succession of days is marked on the axis of abscissas. D and O illustrate experiment (a); no. | and no. 4 illustrate experiment (b). Iie. 2. Error curves for discrimination of light intensity. Ordinates repre- sent number of wrong choices in twenty trials; abscissas represent days -, Curve for mouse No. 1. = , Curve for mouse No. 4. ~ = eno -, Curve for mouse D. 0 S00 9 0 = (Cine OP Tmo ©). Experiment (c) Influence of background. Metuop: In this experiment two animals, O and D, were used. They were trained for white by one hundred trials, in which the back and sides of the experiment box were covered with cards 554 KARL T. WAUGH (18 x 18 cm.). These were interchanged when the food boxes were changed so that the black food box always had a black back- ground and the white box appeared against a white background. The choices during this training were not recorded. Upon the completion of the training, changes were made in order to obtain an answer to the questions: does the mouse choose by discriminating between white and black boxes, or is it influ- enced also by the illumination of the whole field? Is food associ- ated with object or with background? For this purpose white cards were placed behind the black box and black cards behind the white. Other conditions were later introduced for testing the nature of the association formed. RESULTS: White box against black background and vice versa ap MOUSE | NO. TRIALS RIGHT WRONG REMARKS ORE oc) s Semen were | 20 5 5) The last ten, 100 per cent | DEAE Fees Abc alas Met | 20 13 ff right, 7.e., white. | | Uncovered tin boxes used. Backgrounds only changed MOUSE NO. TRIALS | RIGHT WRONG REMARKS OR oe i eee 10 a 3 Great hesitation OD eR Gorn Poko 2 KuereeF 10 10 0 Mouse D was now tried without backgrounds but with the black and white papers on the food boxes. The result of twenty such choices was: Wine i box (meee aero 14 Blackbox cto eee 6 The next experiment was with plain tin boxes without back- grounds, but with strips of paper (4 em. x 18 em.) laid crosswise on the floor of the experiment box directly in front of the boards which earried the electric wires. Black paper was placed on the one side and white on the other. Under these conditions mouse O in ten trials crossed over the white paper to reach the tin box nine times, the black paper, once. VISION IN THE MOUSE 5a. In the next experiment, light gray paper was substituted for the white with the following results. MOUSE NO. TRIALS | LIGHT GRAY BLACK Dirge tee od Oe ne eee eed oe eit Oe 26 12 14 Doane soy Rae ae een re nn SEER et Ee er ee 10 6 4 The strips were now taken up and the tin boxes were covered respectively with light gray and dark gray papers. No back- grounds were used. The experiment resulted as follows: By mouse D, in 40 trials, the light gray was chosen 19 times, the dark gray 21 times. Haperiment (d). Selection of yarns. Preference of mice for ight or dark yarns obviously depends upon their power of discrimination. MernHop: Neither food nor electric shock was used in this experiment. Advantage was taken of the opportunity atforded by the instinct of a mother mouse to make a warm nest for her litter. A black mouse, X, about six months old, with five little ones, was taken as subject. When the young mice were a little over a week old, pieces of yarn were hung in the cage and some of the cotton was removed from the nest. The order in which the mouse took the different yarns to replenish the nest was recorded. Theyarnsused were white, black, and two shadesof gray. These were hung in the cage in a row, 4 em. apart and at a distance of 15 cm. from the entrance to the nest. The lower ends rested on the floor of the cage. After a set had been pulled down by the mouse and taken into the nest, four other pieces were hung up in a different order. After three or four sets had thus been taken they were removed from the nest and the experiment was repeated. Inanentireseries all the possible arrangements of the fourshades were made. There being six possible permutations of the four yarns and four places, there were 96 selections in the series, each yarn appearing in each place six times. The order in choosing the four positions was recorded also, in order to show what influence the habit of first seeking a certain locality might have. 556 KARL T. WAUGH Resuuts: The table gives the number of times each shade was selected first, second, third, and fourth or last. YARN FIRST CHOICE SECOND CHOICE | THIRD CHOICE FOURTH CHOICE — —. | — Blackay ycereen eee 8 7 | if 2 Darkeray..225a.... 6 6 7 5 inghteray.. 0... oe 4 2 | 6 12 5) Wiiiiemere reteset 3 6 9 4 $$ — _ - Me As seen from the table, the order of preference is (1) black, (2) white, (3) dark gray, (4) light gray. The nest was next turned through 180 degrees and the yarns were hung in the back of the cage, opposite where they had been before, and another complete series of records of choices was obtained in the same manner. The results of this second series are summarized in the following table: YARN FIRST CHOICE | SECOND CHOICE THIRD CHOICE | FOURTH CHOICE Backes - 5s. sates a a 6 8 3 Warkvonay sa. 4s 5 8 7 4 iniehiitonsyee sae 3 4 3 14 Welniteas: Ae 2 seygt beak 9 6 6 3 } The order of preference as shown from this table is (1) white, (2) black, (8) dark gray, (4) light gray. The leaving of light gray till last in both series was so interest- ing that it was thought well to make a check test to learn whether the taste or odor of that particular dye was determining the ani- mal’s choice rather than the shade. A gray yarn was made by twisting together a strand from the black and one from the white, the two preferred yarns, and a set of 24 choices of the three yarns, black, white and gray, was obtained. The results were as follows: YARN FIRST CHOICE | SECOND CHOICE LAST White....... Oa Se 5 0 1 Black Wy mye ome ee 1 3 2 Gray (black and white) | 0 3 3 | or ~I VISION IN THE MOUSE - (‘oncLUSIONS: (From the experiments under problem, 1, A.) That the mouse discriminates between light and dark objects under indirect illumination is evident from experiments (a) and (by) Experiment (c) shows that both object and background are influential in determining the reaction. The albino mouse was influenced by the environment more than the brown mouse. This is shown in the case of the white mouse D making 100 per cent right choices, 7.e., choices of white background when the food boxes were uncovered. This result is quite in har- mony with the biological theory of protective coloration. B. Under direct illumination Apparatus: In fig. 3 is shown a view of the apparatus used in these experiments. It consists of two parts, an experiment box (32 cm. x 52 em.) and a light box (382 em. x 98 cm.) Between these two parts is a slide carrying ray filters. A is a nest box (2953 em. x 18 em. inside) from which the animal can enter the compartment 6 through a door J. From B (20 em. x 17 em.) it cannot pass back into A directly, but must enter one of the smaller compartments in front, which open into alleys on each side. From one of these alleys the animal reaches the nest box by a gate O. The two small compartments, (each 143 cm. x 8 em.) which may be entered from B, are illuminated by the light from electric lamps in the light box, which enters the compartments at G and R& through two apertures each 63 cm. square. These aper- tures, in the experiments now to be described, were covered with ground glass. In the light box the lamps can be moved back and forth to give the required differences in intensity, their distance from the ground glass being measured on a scale S. The light box is divided lengthwise by a partition which insures the illumi- nation of each aperture by the appropriate lamp only. The slide carries three rectangular cells (15x16 x 6 em.) separated from one another by pieces of felt. ‘These, filled with colored media, can be used as ray filters for tests of color discrimination. During the present experiments they remained empty. ( KARL T. WAUGH CZZEEERSSSSSSSS (LLLLLLIZZIEEN eee. Ny ic. 3. Visual discrimination apparatus. A, nest box; B, entrance chamber; R, R, red ray-filter; G, green ray-filter; L, L, incandescent lamps in light box; S. millimeter scale on light box; J, door between A and B; O, O, doors between alleys and A. (Yerkes, The Dancing Mouse, p. 153. 1907. The Maemillan Co.. N. Y.) ~ VISION IN THE MOUSE 559 Meruop: The method is very similar to that previously des- eribed. Two animals were used: MOUSE } COLOR | SEX AGE Ue nee iit ne oe | black male About 4 months eg ee se white male 5 months These animals were trained to enter the brighter of the two compartments. Food was kept in both alleys near O, but the exit into the alley on the side of the weaker light was closed by a piece of glass which could be slipped into the partition just above the exit and pushed down. A shock was given whenever the mouse entered the darker compartment. From the brighter com- partment the animal was allowed to passunmolested into the alley, obtain food and enter the nest box through the passage O. After receiving a shock the animal was apt to run quickly out of the darker compartment and into the other one. The wire door at / permitted passage only into 6, and those at O, O, only into A. The mice soon learned that the food was to be obtained only by passing from B through one of the small compartments into the alley. Twenty experiments were made with a mouse each day. Some- times it was found necessary to urge the animal to make a choice by gradually narrowing the space in compartment B with a thin board, placed vertically across the compartment. The experiment was begun with one light at a distance of 34 cm. from the ground glass, and the other at 54cm. After each choice, or sometimes two or three choices, the light that had been at 34 cm. was moved to 54 em. and the other moved up to 34 em. thus making the other compartment the brighter. At the same time the piece of glass which had blocked the exit was removed and pushed into the other slip, blocking the exit on the opposite side. Resutts: (‘‘Right’’ indicates choice of the brighter of the two lights). D60 KARL T. WAUGH Lights at 34 em. and 54 em. ee | MOUSE ieee 2ND 3RD 47TH | 5TH TOTAL U {Right 15 aia |, 14 lew el. CS 81 oS eel aaa | Wrong 5 3 3 3 | 2 19 y {Right 14 14 13 17 12) | 74 A BRM Ra es 7 | Wrong 6 | 6 7 3 4 26 Lights at 34 cm. and 40 cm. MOUSE eae 2ND 3RD 47TH 5TH TOTAL Se : 4 =I “ U J Right 15 oe 13 13 14 57 ’ | Wrong 5 as 7 7 6 33 V | Right 11 15 13 12 13 62 i cae, ' | Wrong 9 7 i 8 ul 38 (oncLustons: The mouse discriminates between differences in the brightness of white light; the less the objective difference, the greater the difficulty in discrimination. The discrimination of the albino mouse is slightly inferior to that of the mouse with black eyes. PROBLEM 2. COLOR DISCRIMINATION A. Under indirect tllumination Experiment (a). Discrimination of colored objects. APPARATUS AND Mrruop: The apparatus used in this experi- ment was the same as that already described under problem 1, A, (a) and (b) (see p. 550). The colors chosen were an orange-red and a blue (Bradley papers). These were judged to be of equal intensity by several members of the laboratory. The papers were pasted on tin boxes and the experiment was conducted as in the intensity-discrimination experiment. The following eleven mice were used: ; VISION IN THE MOUSE o61 MOUSE COLOR! SEX | AGE? DATE OF EXPERIMENT INOS Les ct eit: eee eel eT ANY, male 2months Jan. 27—Feb. 20, 1905 INOW Sieh cies pe ciate cee ee OE a male | 7 weeks Feb. 17-25, 1905 \ YF al Scot ane ae MeO Di 5 white female | 7 weeks Feb. 2-25, 1905 IN(OMOReo 5 bic: Ebi eray male 9 weeks Nov. 20—Dee. 2, 1905 1B} ML eee Aeege wae!) Iconian male 2months Oct. 10-Nov. 8, 1905 SES rk a> Sais oes Se brown male 1 month Oct. 10-Nov. 8, 1905 AN S09 ae ee white male adult Dec. 2-4, 1905 SI tek on ees white male adult Dee. 2-4, 1905 (Ols3 5 SE ere eae white male 1 month Nov. 9-Dee. 10, 1905 IDs 5 AA ee ee ee white male | 5 weeks Nov. 9-Dee. 10, 1905 NP Nace stance ar white male 7 weeks Nov. 29-Dee. 10, 1905 1 All the white mice are albinos. 2 The ages are approximate. An adult mouse is over 2 months old. Age given is age at the time experiment was begun. Marked individual differences in the animals made it possible to get a much greater number of choices with some than with others. The ease with which a motor habit may be formed makes all the difference between a good subject and a poor one. Mouse no. | early formed the habit of running directly to one of the tin boxes, taking a piece of food (if he happened to select the box which contained food) and running back with the morsel to box N. He would be eating the last piece of food while I was changing the boxes and switching the current to the other board. Thus he saved me much time. He always made his full assignment of twenty choices per day. Some of the other mice would follow along the edge of the experiment box, pause a long time in the corners and wash their faces, or they would take the food in such a way as to give no satisfactory evidence of choice. Throughout these red-blue discrimination tests the food was kept in the red box. At frequent intervals the papers were removed and fresh ones pasted on, in order to obviate the influence of odor. The ex- periment box was so adjusted with reference to the windows of the room, that the two halves were equally illuminated. Resuuts: The following table gives a summary of the results of the red-blue discrimination tests: 562 KARL T. WAUGH MOUSE NO. TRIALS RIGHT (RED) | WRONG (BLUE) PER CENT RIGHT INOS RU. sees Sapna eee peewee 480 320 160 67 INOMOe Bete ceemoen en. 59 os iN?/ 70 Wie ee ere th Revers. es xa 185 104 $1 56 INOMPO Mao eee ese ar 91 49 42 a4 JIS Pees oa ecole oe Omens noes 210 160 50 76 BES Sere ehsorcronns itis: dana ats 28 15 13 54 AE ees reer rtess PRE ac ac.bebcheg 12 9 3 75 | BaP oc. cel tohtes cht Arar Mec 10 6 4 60 CRE eS MG ete yi aks cones gs D9 29 26 53 | Bees ethan hcl Ga nce rae CHE SS 4 22 22 a0) ieee nr eee “Airs . cunt ae atcleks 120 6S 52 57 The figures in this table represent the total number of right and of wrong choices during the training. The improvement in discrimination and the stage of success finally attained are shown in the curves of fig. 4. The results for one mouse are pre- sented as typical. On the axes of ordinates are represented the | Fre. 4 Curves of learning. Number of errors made daily is represented on the axis of ordinates. Number of days is represented on the axis of abscissas. Lower curve is the error curve for one mouse in red-blue discrimination. Upper curve is the curve of dissociation and association for the same mouse. q /o Ws 12 43 /4. 15 /e / VISION IN THE MOUSE 563 number of errors made daily in twenty choices. Successive days on which sets of choices were made are represented on the axis of abscissas. The horizontal line at 10 may be regarded as the line of no discrimination. CHECK TESTS: To make sure that vision was being used by the animals in discriminating, rather than smell, food was placed in the blue-covered box and the red box was emptied. With mouse no. | after the training for red of 400 trials, the result was, that, in the first hundred choices, the mouse selected the empty red box first 59 times in preference to the blue which contained the food. Keeping the food in the blue, the experiment was continued in order to discover how rapidly an association of one color with food, formed through training, could be changed to an association invol- ving a preference for the previously avoided color. The dotted curve of fig. 4 represents the result of 380 trials. It is an error curve in which choices of red are considered ‘‘ wrong.”’ With another mouse, B M, a series of experiments was made, in which the check test consisted in changing the food into the blue box for from three to five choices, after each ten choices with the food in the red. This was done after 190 choices in training for red had already been made. The results are given in the follow- ing table. The first column gives the number of the test; the second gives the number of right choices out of ten with the food in red; the last two columns give the results with the food trans- ferred to the blue. FOOD IN RED FOOD IN BLUE NO — aed a a DS Right Choices No. trials Red chosen 1 8 5 5 2 9 By 3 3 8 a) 5) 4 10 5 4 5 7 5 3 6 8 5 5 if 4 5 3 8 9 3 0 law] o at ie) — or) eS) oo (0) bo 100) 564 KARL T. WAUGH Experiment (b) Selection of colored yarns. Mernop: The apparatus as well as the method here used was the same as that used in the experiment on the selection of gray yarns (see p. 555). Three animals, all of them mothers of litters, were subjects in the experiment. MOUSE COLOR SEX AGE DATE OF EXPERIMENT KG 4-4 6 bole: otelteaataea Ne meioee white female | 3 months | Mch.—Apr., 1906 DOA lan fos ae hoa black female | 6months Jan., 1907 VR erie er rsa.) ah che white female | cir.1 year Jan., 1907 Four colors, blue, red, yellow and green, of as full saturation as possible, were selected from among the yarns used in the laboratory for testing color blindness. These were hung in a row in the back part of the mouse’s cage, all at the same distance from the opening into the nest. After the four pieces had been pulled down and carried into the nest one after another, another set was arranged in a different order, as in the previous experiment with gray yarns. Resutts: These are presented in tables, each table represent- ing a series of 96 selections of yarns. Mouse K NO. OF TIMES SELECTED COLOR -— = 4 eee ORDER OF | PREFERENCE Ist 2nd | 3rd 4th Bliet ita o. Te ee 6 3 | 10 5 3 1 BAEY 0 lanes cbechles ce Bue RL aa ea 10 5 6 1 Mello watts cee ace 6 11 2 5 2 Greciise.o iho : 2 5 9 8 4 Mouse X NO. OF TIMES SELECTED COLOR = ORDER Or PREFERENCE Ist 2nd 3rd 4th Blues ee ae Bea yee i 5 A S 3 LEY eyo luis fete MMR ng rrerta Men set 6 11 3 4 2 Wellowid.ae eee ces 11 3 7 3 1 GreCTIP LAYS eee Mea 0 5 10 9) 4 VISION IN THE MOUSE 565 The order of preference is obtained by comparing the numbers which represent the preference-value of each color. Such a number may be obtained by adding together the number of times a color is chosen first * 4, the timesit is chosen second X 3, thetimes chosen third < 2 and the number of times chosen last. Thus for K, the preference valueofblueis: 6X 4+3x*38+10x*2+5 = 58. In the same way the preference value of red is found to be 67, of yellow 66, and of green 49. For X the values are: blue 59, red 67, yellow 70, and green 44. This preference-value is usefulin showing the constancy through- out the series. Thus the whole set of results for X was divided into halves and the preference values for each half were as follows: FIRST HALF SECOND HALF WHOLE Sic eee) oS oem | 26 33 59 LEY line aah Goat eh bh Rk ee 5. rea 34 33 | 67 PYG O Wee oer een Meas? 35 35 | 70 Gree Si eee eR a! 25 19 44 Concuusions: The mouse can learn to associate food or elec- tric shock with red or blue objects. The connections thus formed may be disassociated and an association formed with another color. In albino mice, color discrimination is poor. Red and yellow are preferred to blue and green, and of the latter two, blue is preferred to green. Whether the discrimination involved is true color discrimina- tion, after the fashion of that in human beings, can not be dis- covered, but we may eall it color discrimination so long as it answers the practical purposes of the mouse in distinguishing between such objects as it is likely to meet with in its habitat. If it be claimed that colors of the red end of the spectrum are preferred by the black mice because they seem to them the darker, we would suggest a correlation with the results obtained in the matter of the preference for gray yarns (see pp. 556). The same mouse, X, appeared in the two experiments. White and black were both preferred to either of the intermediate grays. THE JOURNAL OF COMPARATIVE NEUROLOGY AND PSYCHOLOGY, VOL. 20, No. 6. 566 KARL T. WAUGH B. Under direct illumination Apparatus: The same as that in problem 1, B, fig. 3. The colors used were red, blue and green, obtained by means of the following ray filters and glasses: For green—a saturated solution of nickel nitrate. For blue—solution of copper ammonia sulphate. For red—pieces of ruby glass placed in the cell. The three cells in the slide between the light box and the experi- ment box were filled with the solutions, the two on the outside containing the same solution. By moving the slide a little to one side, the relative positions of the colors were changed, the one previously on the right now appearing on the left. The cell containing the ruby glass was filled with water in order that by the effect of the refraction it might appear to be the same distance away as the other colors. In all, fourteen animals were made use of in this experiment: MOUSE COLOR SEX AGE DATE OF EXPERIMENT Ore eae rt ee white male adult Apr. 23-27, 1906 Rae eo cane eo white male adult Apr. 17—May 17, 1906 IN Sires eee: black male adult Apr. 28-May 7, 1906 Ree cr ee: white male adult Apr. 12-May 25, 1906 Code Viktor: *. brown female adult Apr. 5-May 19, 1906 Ca ett AL ted white male 5months Mar. 26—-May 24, 1906 IVT OSE Pie oe brown male ? (wild) Apr. 6-9, 1906 iO cia ety white male 7months Mar. 26-May 24, 1906 1S ee a eee brown male adult Apr. 9-—May 19, 1906 Desh cht apace es, 8 gray male 4months | Novy. 9-Dec. 17, 1906 Wie anew. black male 4months Nov. 25-Jan. 14, 1907 XMa.............|gray and white!) male 1 month Jan. 9-21, 1907 OP ER Red white male 6 months Noy. 16—Dec. 10, 1906 Fitts SA Ae ee white male 6months Noy. 14-Dec. 10, 1906 |The gray and white mouse had pigmented eyes. Meruop: ‘The first two colors used were green and red. One of the electric lamps was moved until the intensities of the two colors appeared equal. Judgment was given on this point by five members of the laboratory, after each had been in the dark- room for five minutes. lard VISION IN THE MOUSE 567 A preliminary test of 50 trials was made with each animal to see 1f any natural preference for either of the two colored lights existed. The mice were found negative in this respect. Under the circumstances, it is evident that the problem resolves itself into the question whether the mouse can be trained to prefer one quality of ight to another. This is to be done by the association of pleasure or pain with whatever distinguishing characteristics the lights may present to the eye of the animal. Just what the distinguishing characteristics are—what factors the mouse uses in discriminating—may be suggested by the use of check tests, in which the relative brightness of the lights is varied. In this respect the apparatus of this experiment is more satisfactory than that of the first series under this problem. Yet it leaves something to be desired, for if the animal’s choices are determined wholly by intensity, this fact would become apparent, but if they are determined by more than one factor (e.g., quality and intensity) as seems probable, then we can hope for results only on the supposition that a point of non-diserimination may be found, or better, a point of least discrimination, where a certain intensity tends to counteract the quality which would act in part as a determining factor. RESULTS: Red-green discrimination ’ l MOUSE TRAINBD FOR | NO. CHOICES | RIGHT WRONG : et ae Oe ene se 6 es sina s green 100 | 57 43 6.) ayey ee Kee Gee ae red 100 | 60 40 Es oon ELS Aanisigica ane eee red 100 58 42 | Green-blue discrimination ' | MOUSE TRAINED FOR NO. CHOICES RIGHT WRONG 1 es Ea ee a AES pe ae green 100 42 58 COR iet See rea ae. Se ES green 100 50 50 Ochs ei etakestA ene beer a OER a nike se green L100 49 51 IN Ug ele pe nie aed eater ret ch me green 60 29 dl ID ose ote 5 sae ee eta sioae cine) green 100 50 50 Iie rca eek ke By nce, Ghose A or eaareae green 80 41 39 568 KARL T. WAUGH The next problem undertaken was that of ascertaining whether the mouse discriminates between white and red lights when they are of about the same degree of brightness. For this experiment four incandescent lamps were used as sources of light. Two, measuring 4 ¢.p. each, were used back of ruby glass to yield the red light, and two measuring 13 ¢c.p. each back of ground glass to yield the white light stimulus. The posi- tions of these lamps in the light box were changed as indicated in the following tables of results. Lamp of 13 c.p. 34 em. from red glass; lamp of 4 c.p. 80 cm. from white glass (This yielded a red and a white stimulus which seemed of equal intensity to the human observer) | MOUSE | COLOR NO. OF TIMES CHOSEN TOTAL Dg oe ae Red | 4 6 eB Eb iba i ne : : : | (Ud) Roe Witte| 6 4 | 8 Wem yl 7 | 6 ls 1ic6 7 |. "Gs 2 Tata aesmtie 3) 4 NOM AR an ial \ MOUSE COLOR NO. OF TIMES CHOSEN | TOTAL eee White} 8| 9| 8| 8| 7 7/8] 8 10, 8| 9| 10} 100 Wieser Reds) (27) 002 | os) Sere aa 2a 22-0) eo eaten ee) | | Ete ae White| 8| 6| 7 | 7 28 SI Reet Reds 2) e4 113 SS Sa ae ete | | 12 Lamp of 13 e.p. 54 em. from the red glass; lamp of 4 c.p. 70 em. from white glass MOUSE COLOR | CHOICES TOTAL | MOUSE | COLOR | HOICES | TOTAL | | | | 7 | | | | _—, Spo eepee White; 7 | 10 7 9 | due ena! Winabe |: 18 9 17 ORsiwecens: Red | 730] 70 St wees e Red | 2 I 3 Lamp of 13 ¢.p. 14 em. from red glass (one of the two thicknesses of ruby glass used previously removed) ; lamp of 4 c.p. 20 em. from white glass MOUSE COLOR CHOICES TOTAL U2 eee et oe White 8 s 16 | BP ee A ne ee ee ALS 2M Be Red 2 Ps 4 VISION IN THE MOUSE 569 Lamp of 13 e.p. 14 em. from red glass (single thickness) ; no light through white glass MOUSE | COLOR CHOICES TOTAL (is ae ae os White Got esala 6 15 Wepre eee Red AM | ite 15 Lamp of 16 c.p. 14 em. from red glass; lamp of 1.69 c.p. 90 cm. from white glass. CHOICES | TOTAL MOUSE COLOR | chy ae iene (Sis erie | White | Sai ae 17 iW: 5, Renee eee Red | Por I he ie 3 [ i? Lamp of 16 c.p. 14 em. from red glass; Lamp of 1.69 c.p. 90 cm. from white glass. (With glasses at exits, so placed as not to reflect light) MOUSE | COLOR | CHOICES | TOTAL | | } [Gia eades SI eS White 49 6|5/6/6/6|7|-| - 49 [Li OAR tees Red |6/1|)4|5/4/4/4/3| -| - 31 Sy. od a ole pel White |7/6/5/8|5/5|5/4|/5/6) 56 oe ee | Red [3)/4/5/2/5/5/5|/6|/5/4) 44 Lamp of 16 c.p. 34 em. from red glass; lamp of 1.69 c.p. 60 c.m. from white glass MOUSE COLOR | CHOICES | , TOTAL ia ; "3 lie l aa Da ne os. Oy Sts ae a White- 27 esan), 4 | 6 15 | 5 | Gn | 4 15 OX ete eho eee ec caters Red Lamp of 13 c.p. 34 cm. from red glass; lamp of 4 c.p. 80 cm. from white glass (Brightnesses equal for human eye) » MOUSE COLOR CHOICES | TOTAL , Te eee oe | White |5/5/ 6/6 65 8)4)5/7 57 eo ee a ee We MRed) 95 | 5 | 4 )-4i) 45a 256) aaa 43 ssh een EN | White |6/8|/4/4/7/3/5/5/5|6/ 58 Teme TON SOA MES | Red (4)2/6/6|/8/7/5/5/5/4) 47 570 KARL T. WAUGH ConcLusions: Our results do not indicate any ability to dis- criminate between red and green or between blue and green. The black-eyed mice discriminate between red and white light when they are of equal brightness to the human eye. When the red is made darker the discrimination is shightly improved. When the red is made brighter the discrimination is not so good. The power to discriminate seems to fall away as the brightness of the red relative to the white increases for the human observer. PROBLEM 3. PERCEPTION OF FORM APPARATUS AND Meruop: In these experiments the appar- atus used was the same as that already described in connection with the problem of the discrimination of direct light stimuli. (see p. 557) Pieces of black cards were substituted for the liquids in the cells, and in these cards were cut holes of the forms desired for the test. The forms selected were, a circle, 4 em. in diameter, and an X-shaped figure which was inscribable in a square of 4 em. diameter (fig. 5). I made these two figures of equal area in Fig. 5 Cards for study of perception of form. order that difference in the amount of light passing through them might not serve as a condition of discrimination. Two incandescent lamps were used as sources of light. Each of these was of 12 candle-power, as tested by the Lutnmer-Brod- hun photometer. The lamps were placed in the light box at a distance of 50 cm. from the apertures. Additional pieces of ground glass were placed against the black cards inside the cells, in order VISION IN THE MOUSE Ge to make the illumination of the apertures more homogeneous, as well as to reduce the light. The electric shock (as punishment for a wrong choice) was given in the compartment with the X-aperture, the side exit on that side being closed with a piece of glass. When the animal entered the compartment marked by the cirele, it was allowed to pass into the alley and to obtain food. The circle and X were interchanged at intervals, usually after three or four choices had been made. This was accomplished by sliding the carriage a few inches to the right or left so that another eard with an X-aperture would come into the field, taking the circle’s place, the circle having moved into the position of the former X. Ten choices were obtained each dav. Four mice were used in this experiment. MOUSE COLOR SEX AGE VE hati tates Oa Abner tin fh a white female 3 months A Aleit ane eA AR al white female 3 months EXERT We ee ct BTA ee ee black female | 5 months XG re me a eh ae Bote black female 6. weeks Resuuts: In the following table, R stands for right choices, t.e., circle; W for wrong, ?.e., X-aperture. Ya Yi xX Xd oe R Ww R Ww R Ww | R | Ww ist... 4 6 4 6 3 a 5 2 SATO Weratoh tle & 7 3 a 3 6 4 4 6 3rd.. 8 2 + 6 6 4 8 2 AG his Saepesty eae 8 2 10 0 6 4 7 53 LAO, - 8 2 6 f 5 5 6 4 6th 6 4 6 f = = 6 4 TAT Nepelst seecae ane S 2 Ss 2 - - = = Sth. a 3 6 4 — - = = Ohrake Wa earo note 7 3 a 3 — = - 10th 6 4 - — - — = I kilotse wee mane 7 3 - - - ~ WA soc oee iG 3 = = = _ — BSN, a oo ee a 3 = = = -- - Wotals/e-. =. 90 40 | St eal mney. 2608 oe eso aimee Ly pe KARL T. WAUGH ConcLusions: As may be seen from the table, perception of form does not seem to be very well developed in the mouse. Only in the case of Ya, may we feel justified in concluding that discrimination was present, and even here errors were not elimi- nated up to the 130th choice. That the mouse is able to form an association of object with food or shock is shown by the experi- ments in intensity discrimination. Evidently the mouse tends to depend upon the size of the illuminated area or the intensity of the hight. When a strange mouse is introduced into the cage of another mouse there seems to be no recognition of the nature of the in- truder by vision. Not until the home mouse can touch and smell of the stranger, does there seem to be any knowledge of whether he is an enemy or the mate of the home mouse. One might suppose that differences in the form of the animal would be noticed at some little distance. It is by form that human beings known one an- other, different expressions of the face being in the last analysis minor differences in form. PROBLEM 4. PERCEPTION OF DISTANCE OF OBJECT FROM ANIMAL APPARATUS: A wooden disk 10 em. in diameter, supported in the center by a column 23 em. in diameter, which passes through Fic. 6 Table for study of perception of distance. « round hole in the top of a bench. The height of the disk above the bench can be adjusted by sliding the column up or down. VISION IN THE MOUSE 573 Merruop: (OMS Bn to hie eS Ee oe 6 1 5 KER Rad ae Pty Sa est ot Ae. 6 2 4 XC er ar ay Ae ee teeter tN hi 6 2 4 SNC Rs Ata e ME tec S 3 5 In all of the white mice there was observed a peculiar head movement from side to side while crouching in the passageway, preparatory to running forward. Since the results in the matter of depth-perception favor the white mice—for the black-eyed VISION IN THE MOUSE 581 mice the number of right choices being 47 per cent of the number of trials, while for the albino mice it is 53 per cent— it was thought that this head movement might be serviceable to the animals, as giving several points of view of the object, each from a different angle, thus possibly rendering a perception of depth easier. Following this idea, a record was kept of the number of times the head movement was observed and this was compared with the right choices. I present as typical the results inthe case of one mouse, Kh. This animal made 14 right choices and 16 wrong. Of the right choices the head movement was observed in 8 and not observed in 6, while of the wrong choices there was head move- ment in | and no head movement in 15. ConcLusions: Judging by the number of errors, we may con- clude that the mice do not make use of visual perception of depth. If they have the anatomical equipment necessary for the percep- tion of depth, their important muscular sense controls their actions, making them take the same course they took on the pre- ceding occasion. II. BINOCULAR VISION The question of binocular vision in the mouse suggested itself in connection with the investigation into the perception of depth, and an attempt was made to find how far the structural condi- tions are fulfilled which would make it possible. The conditions which must exist in order that binocular vision in the psychological sense may be present are: (1) Itis necessary that the eyes be so situated in the head as to have a portion of the field of vision common to each. (2) There must be consensual movementsof theeyes. The lines of sight of the two eyes must be capableof moving approximately parallel to one another so that the imagesof an object may fall on corresponding points of the two retinae. (3) There must be a chiasm of the optic nerve and a portion of the fibres from one eye must mingle with a portion of those from the other, that is, there must not be a total decussation of optic fibres at the chiasm. THE JOURNAL. OF COMPARATIVE NEUROLOGY AND PSYCHOLOGY, VOL. 20, NO. 6. 582 KARL T. WAUGH If the animal possesses the requirements estimated above, there is good reason to suppose that it has binocular vision in the proper sense. A piece of anatomical work undertaken to supplement the experiments on behavior was that of measuring the angle of divergence of the optic axes, and determining the angular field of possible binocular vision. The results appear in fig. 8. In ahead from which the eyes had been removed the optical axis was ob- tained by drawing a line from the fundus of the eye-socket through Fie. 8 The field of vision of a mouse. 7, right monocular field; /, left monoc- ular field, rl, binocular field; 0, 0, optic axes. a point equally distant from all points in the rim of the opening, outward into space. The limits of the field of vision were deter- mined by the points from which the eye could be plainly seen. Although the axes of the conical eye-sockets in the mouse diverge greatly, forming an angle of about 100 degrees, yet, owing to the prominence of the eyes themselves, it is quite possible that they may receive images from the same object simultaneously. VISION IN THE MOUSE 583 I experimented by standing in front of the mouse where, with one eye closed, I could plainly see both eyes of the animal, and then moved my head from side to side in order to discover how great was the angular field from which the two eyes might be seen. I found that I could move my head through an angle of 70 degrees without losing sight of either eye. Although the possibility of seeing the ball of the eye in a human being does not always mean that the eye sees the observer, yet we know that when the eye is turned as far as it can go toward the observer, he is seen on the extreme border of the field of vision. Fia. 9 Diagram showing relative size of lens in eye of mouse. ch, choroid; l, lens; c, cornea; 2, iris; 0, optic nerve; r, retina. This of course, is made possible by means of the refractive power of the lens and by the fact that the retina extends all the way around the inside of the eye right up to the ciliary process. If the retina were of uniform acuteness throughout its area, or if the lens were larger, it would not be necessary to move the eye to its limit in order to see the observer who can just see the eye- ball. These are just the conditions which we have good reason to suppose exist in the mouse. In the first place the size of lens, 584 KARL T. WAUGH compared with the size of the eye, is much larger than in man (see fig. 9); and, second, the homogeneity of the retina is demonstrated in a later section (IV) of this paper. From these considerations we feel justified in concluding that there is in the mouse a portion of the field of view which is shared by the two eyes. Concerning the second requirement for binocular vision it is exceedingly difficult to secure data from observation of the ani- mal. The uniform appearance of the surface of the eye makes it almost impossible to detect movements of the eye. The black glistening beadiness of the eye is always the same. Several attempts were made to obtain a point of reference on the eye-ball by using a small square of paraffined Chinese white, but without success. The animal would close the eye immediately and dis- lodge it. Finally on close observation under a very bright light a faint line of the pupil could be distinguished, and when the light struck the eye at a certain angle, its movements could be observed. The animal was held on the palm of the hand facing the ob- server and trials were made by moving a finger back and forth on different sides to learn the nature of the eye movements. When the finger was brought slowly above the level of the palm, within the field of vision for one eye, the mouse would turn its head shghtly toward the finger. When the fingerwas moved toandfro, it was not followed by the eyes. Under these conditions the ani- mal always reacted by remaining perfectly still with the eyes in a fixed position. A more rapid movement of the finger would elicit only a slight further turning of the head in the direction of the movement. The finger was then lowered and raised on the other side of the animal’s head. These conditions were alternated and repeated about twenty times. The reaction was in all cases the same. Fear cannot reasonably be suggested as a cause for the move- ments not being followed with the eyes, for the animal was per- fectly tame. It was used to being handled daily and would run to 1 It might naturally be supposed that the uniform beady appearance of the eye is of biological value in permitting the animal to make movements of the eye which, while enabling it to see its enemies, would themselves remain undetected. VISION IN THE MOUSE 585 my hand at any time. Further, the fact that the head was moved would make the view that fear inhibited the eye-movements un- tenable. Such slight eye-movements as do occur seem to be rather for the purpose of getting more light from the general direction of the object than for getting a clearer image. There appears to be nothing in the vision of the mouse which compares with fixation in the human being. If the image of the object is cast upon any part of the retina, all the conditions are fulfilled which make vision useful to the mouse as a protective sense. On the third requirement for binocular vision, the dividing and ‘crossing of the optic fibres, positive statements can not be made from a mere study of the anatomical structure. The chiasm exists in all vertebrates, while in mammals, birds and to some extent, in reptiles the nerves unite, so that each is made up of fibres from both roots. In fishes they cross without uniting. Whether fibres from corresponding points in the retinae unite so that their excitations are carried to the same neural channels in the brain is practically undiscoverable, but the fact that thereis any union at all rather than total decussation, indicates an intimate degree of relation between the two eyes. Harris,? who worked by the method of degeneration upon the optic fibres of the lizard, ventures the statement that decussa- tion in rats and mice is complete. More satisfactory knowledge on this point would be gained from the cumulstive evidence of the degeneration method and the method of myelogenesis as used by von Gudden. The amount of mingling of optic fibres would not be great considering the fact that only small portions of the retinae can receive images from the same object. Cajal® working by the method of degeneration upon the optic chiasm of the rab- bit and the mouse, has in fact shown that the crossing of the fibres is not complete. 2 Harris, W. Binocular and stereoscopic vision in man and the other verte- brates, with its relation to the decussation of the optic nerves, the ocular move- ments and the pupil light-reflex. Brain, part cv, p. 107. 1904. 3 RaMON ¥ CaJAL, S. Textura del sistema nervioso del hombre y de los verde- brados, vol. 2, p. 652, Madrid, 1904. 586 KARL T. WAUGH If there were no mingling at all of optic fibres, then we should be urged to the absurd conclusion that a single object which casts double images in the eyes is not interpreted as one, or else that there is an alternation of attention, a sort of psychical rivalry in which the sensation from one eye intermittently inhibits that from the other, a view which is not in accordance with the law ot parsimony and is most improbable. As a third possibility it might be claimed that one of the images becomes the dominating stimulus while the sensation from the other is entirely inhibited. We find such a possibility proposed as a theory by Morgan.t He supposes, in the case of animals like the rabbit, where the eyes are so situated that they cannot combine in binocular vision, that “the image that falls on the most sensitive area or yellow spot of one eye suggests the focal impression, while that which falls on a similar spot in the other eye is marginal to its conscious consentience.’’ The existence of such a yellow spot Morgan as- sumes. The need of a theory of this sort might be vindicated, were it shown that animals whose eyes diverge at so great an angle possess a fovea. Schafer’ states that only man and some primates have optic axes capable of convergence and a single central fovea. The theory proposed above would be adequate for the explana- tion of the conditions which obtain in those animals, which in attending to an object turn one side of the head toward it, thus inhibiting any sensation from it by way of the other eye by practically excluding it from the other visual field, as do most birds. The mouse, however, turns toward an object enough for it to be clearly perceivable that lines from the object strike both eyes with a generous margin. In reasoning from binocular vision in man to that possible in the mouse extreme caution is necessary, because in man many of the phenomena of binocular vision are closely connected with the central point of most acute vision in each retina. Convergence is meaningless unless we have reference to some definite pomt in 4 Morcan, C. Luoyp. Introduction to Comparative Psychology. chap. 10, p. 160, London, 1902. 5 Scuirer, A. KE. Text-book of Physiology. vol. 2, p. 1148, 1900. VISION IN THE MOUSE 587 the retina, where the lines of sight which converge on the object terminate. Although it is true that corresponding points function most accurately in the region of the central spot, as is shown by the difficulties attending the experimental determination of the horop- ter, yet it need not be true that they owe their existence to the presence of a central spot. The existence of corresponding points is quite possible in a retina, all portions of which are similarly organized. The mouse, which has no fovea, might have certain portions of the eye adapted for binocular vision. These are the extreme pos- terior areas of the retina, which correspond to the temporal seg- ments in the human being. It is here that images from one object ‘an fall on both retinae and, therefore, here corresponding points must have been developed, otherwise the animal would perceive objects double. The conditions here differ from those obtaining in the human being only in degree. In the latter a considerable portion of the nasal area of each retina can not function in binocular vision on account of the prominence of the bridgeof the nose, and there- fore a point in this region can have no point in the temporal por- tion of the other retina corresponding to it. In the mouse this area is much more extensive. If, for convenience, we call this area the monocular area, as distinguished from that in which corresponding points exist—the binocular area, then in an animal like the mouse the centre of the retina lies in this monocular area. If there were a pointof clearest vision near the center of the eye, it would be merely a fixation point which might function in Jen- ticular accommodation. However there is no structural sign of a fovea in this region and obviously such a point in the monocular area could not function in convergence. We must conclude that the optical axis of the mouse’s eye has no functional significance. If we expected to find a fovea which serves the mouse as ours serves us, we would look for it in the extreme posterior portion of the retina, in the binocular area. I made a number of observations upon squirrels in connection with this problem. Inthe squirrel the position of the eyes is some- 588 KARL T. WAUGH what less favorable to binocular vision than in the mouse. The snout of the mouse is proportionately longer than that of the squirrel, but the bridge between the eyes is so much lower that there is a large field of vision for two eyes above the head, which the squirrel does not possess. When the squirrel under observation was approached from the side, he would sit on his haunches, lift up his head and show all signs of attention. When I would kneel on the ground within six feet of him and make no movement, he would remain with one side of his head toward me, using only one eye. When a movement was made by waving the hand back and forth, he would turn his head directly toward me in a position where both eyes were equally visible. Thisis a reaction very similar to that in the mouse. Two explanations of this reaction suggest themselves: (1) The movement is made for the purpose of getting perspective which would aid in the perception of distance. Here convergence would be implied in the way that it occurs in human beings. (2) The head movement is made for purposes of orientation pre- paratory to turning the body in the direction of the stimulus and, perhaps, approaching it. The forward movements and turnings of the animal are executed with reference to a median plane, to which the precise relation of an object in space is more easily determined when the object is seen with both eyes. For deter- mining the space relation of twoobjectsboth in the median plane, the one factor of location of images on the retina is adequate. Obviously a crawling animal like the mouse is moreconcerned with accuracy in its right andleft movements thaninmovements in the vertical plane. Of the two explanations given the latter is the more likely to prove correct. It is in accord with our experimental results in depth perception in the mouse, and further, the idea of conver- gence is not entirely compatible with a retina without a fovea, homogeneous throughout and of which the habits of the animal would demand only the function of communicating in a rough way the general nature of the object and its direction. Perception of distance adequate to the animal’s needs may be obtained through a synthetic correlation of retinal impressions and motor impulses of monocular accommodation. VISION IN THE MOUSE 589 Loeb® suggests that some animals may localize by means of changes in the form of objects, which must result from a marked astigmatism existing in the eyes of certain animals. Cole’ distinguishes four types of animal response to photic stimuli: A. Response of eyeless forms. B. Response of forms with direction eyes. C. Response to size of luminous field. D. Response to different objects in the visual field. For our purpose we would make five types, dividing the last into two. Of these the first is to be found in those animals which perceive the presence of objects and afew general characteristics concerning them. The other is in those animals which have a distinct perception of form by means of a fovea or central spot of most acute vision. The mouse would be included under the first of these two. Under the latter would come man and the rest of the primates and some other animals such as the chameleon and certain birds of prey. Animals of the former class, destitute of a fovea, although they may have a more delicate perception of the existence of objects in the field of view than we, yet do not see the form of objects regarded, as distinctly as we do. A faint star is best seen with averted gaze owing to the fact that the rods are functioning and are susceptible to faint light stimuli but not to distinctness of outline. This is the case with the mouse, in whose retina cones do not exist. The visual conditions existing in the mouse as revealed by our study of it are in accord with a view expressed by LeConte:§ “In lower animals, especially those which are preyed upon by others, it is far more important to see well in every direction than to fix attention exclusively on one point, therefore, the advantages of exquisite microscopic distinctness of the centerof the field are sacrificed for the much greater advantages of moderate distinct- °’ Lors, Jacqurs. Arch. f. d. ges. Physiol., vol. 41, p. 371. 7 Coxe, L. J. An experimental study of the image-forming powers of various types of eyes. Proc. Amer. Acad. Arts and Sciences, vol. 42, p. 410, 1907. 8 LeConte, J. Sight, chap. 5, New York, 1881. 590 KARL T. WAUGH ness over a very wide field. The most important thing for them is a very wide field and the equal distribution of attention over every part. Hence their eyes are prominent and destitute of a central spot so that they see all parts with equal distinctness.”’ III. KINAESTHETIC SENSATIONS, THE GUIDE TO MOVEMENT In the human being who is at all introspective, kinaesthetic sensations often come into the focus of consciousness, but presum- ably more often they do not. In the latter case, strictly speaking, they are hardly to be called sensations. They exist merely as neural modifications. Traces are retained within the nervous structure in the form of facilitated transitions across the synapses, or, of increased permeability of the neurones. These neural pro- cesses operate in controlling the actions of the body without nec- essarily involving consciousness at the time. Sometimes they emerge into consciousness late, as when we become aware of having had our limbs in a certain position and know that they are not now in that position. Again we may have kinaesthetic images of bodily actions we are about to perform. Our theory of the guiding sense in the mouse may be introduced with an illustration from human psychology, the phenomenon of alternating personalities. The normal person may do many things of which he is wholly unconscious, e.g., he may lay an object in a certain place and, after a while, search for it, entirely unconscious of having put it anywhere himself. Later when another person- ality is dominant, it may occur that the knowledge of the location o1 the object is present to consciousness and there is no difficulty whatever in finding it. The second personality remembers putting it in the place in which it is found. This phenomenon may be explained on the supposition that during the incumbency of the for- mer personality, the kinaesthetic sensations from the movements of the limbs are unable to emerge into the conscious field because other psychical processes, viz., those to which the normal person is attending, have control of the system of neurones whose excite- ment is accompanied by consciousness. Association of the kinaes- thetic sensations in question with the present perceptions is VISION IN THE MOUSE 591 inhibited by the draining of the nervous energy of the association neurones of the higher senses in the direction of the frontal tracts of the brain. Now under these circumstances, it isevident, some- thing must become of the nervous currents coming from the mus- cular sense organ. They seize upon certain motor neurones of the Rolandic area whose connection with the systems operating to reinforce the higher functions is not so direct, and consequently are not so thoroughly drained of energy. Such systems are mainly those which have functioned in bringing about the movements which gave rise to the kinaesthetic sensations. In them theresist- ance is low. Thus it comes about that motor circuits of the sec- ond level are formed. The passing of the synapses in these circuits is rendered progressively easier on account of the rever- berating of impulses through the kinaesthetic-motor system. This system and the higher systems involving attention are for the time being mutually inhibitive. The animal in choosing between alternatives is guided mainly by kinaesthetic sensations which have been registered in the nervous system, just as is the case with the subconscious personality. When the animal enters a compartment where it must choose between situation A on the right and situation B on the left, two internal factors tend to determine action. One of these is visual, the other is kinaesthetic, Of these two the visual is the one em- phasized by Thorndike in his experiments with cats. Observation indicates that in the mouse, the visual stimuli are not of so great importance in guiding the animal as the kinaesthetic. The latter is relied upon wherever possible. Smell seems to be the next in unportance. These considerations prompt us to adopt a law of parsimony in studying the senses of the mouse: we are warranted in inferring a case of visual determination only when there is no possibility of the muscular sense being used for the discrimination in question. Training a mouse to discriminate always involves training him away from a reliance on the muscular sense. This law may be applied to the muscular sense and smell, or, to smell and vision. There is a suggestion here of a possible criterion for erading the intelligence of the animal series, viz., the relative im- portance of the various senses in directing movement. 592 KARL T. WAUGH When Thorndike put his cats into a cage, the process of learning to open the latch consisted in the gradual association of a certain movement with certain sense impressions under certain condi- tions of hunger. Under the instinctive excitement caused by the situation the cat makes many movements; those in each part of the cage are guided by the visual impressions of that part of the cage acting by way of the visual cortex. Each group of visual impressions would thereafter, in accordance with the law of neural habit, tend to lead to the same movements more readily than before. One of these impressions acquires increased intimacy of association with a certain movement more readily and certainly than the rest, viz., the visual impression made by that part of the cage in which the latch is situated, with that movement which results in the falling open of the door. The intimacy of this par- ticular association is increased each time this particular movement is made, until, as the cat casts its eye over the cage, the visual impression of that part of the cage at once evokes the movement. The explanation which Thorndike gives as to why the association between the particular sensory path and the particular motor disposition becomes fixed while other possible motor dispositions do not become fixed, is, that the former association gets ‘‘stamped in’’ by the pleasure resulting from it, while the other is “‘stamped out”’ by the pain of failure. The kinaesthetic sense undoubtedly has a tendency to deter- mine the cat’s behavior, but vision operates more quickly, for the ‘at directs its attention to the visual stimuli, rendering possible readier association between the object seen and the motor mechan- ism. The condition here differs only in degree, not in kind, from that obtaining in the mouse. The relative importance of muscular association and visual association may be well shown by the analysis of the actions of a mouse used in problem 2, B. If we go on the assumption that the mouse’s action was associated with the result of the action immedi- ately preceding, then we divide the whole series of choices of the animal into two kinds of sequences: position sequences, in which the mouse turns to the same side, left or right, where it received food in the preceding trial, and avoids the side where it received VISION IN THE MOUSE 993 a shock, and color sequences, in which the animal goes to the color where it received food in the preceding trial and avoids the one where it received the shock. The turnings to left or right were recorded in all the experiments, and in the case of mouse Q, in 100 choices we have the following results: MOSUL TOMES OCI OLIGORIA yainysy5. skys cise. isa ais vs.a! «ce eee ee 69 RB inlintEBECUGH CES yeryen = Beis wee fw a.oaesChed gl ns See eeracme ne 53 Position sequences in opposition to color.....................+-+-- 26 Color sequences in opposition to position.....................0.... 10 We interpret the behavior of the animal thus: it enters the com- partment where a choice is necessary. Its attention may be on the idea of the food which it expects to receive or upon the pain of the electric shock or anything else you please—if we admit the possi- bility of such attention in the mouse—but this attentive con- sciousness is not directive unless it is associated with the idea of moving toward the food or away from the shock, and not then unless this idea is accompanied by the actual movements, at least in their incipiency. To put it in physiological terms, there must be a connection between cortical centers for representation and the motor areas, and this must be sufficiently energized to drain the energy from the kinaesthetic system. This, as we have been led to conclude from observing the behavior of the animal, is not generally the case. The governing of movements is turned over to the motor circuits, and thus it happens that time after time the mouse runs into the same compartment, the compartment in which it may receive a shock, but still the compartment which it entered on previous occasions sufficiently numerous for a motor circuit of a certain degree of stability to become established. The reverberating of excitation through the motor circuit in the animal may be likened to a fly-wheel, which carries the animal in one direction or another by its momentum. The cue for the revolving of the wheel is afforded by the sensations which are constantly coming from the incipient movements of the animal. These may be reinforced by tactual impressions received from the floor and walls of the compartment. The mental state of the mouse on entering the puzzle box may 594 KARL T. WAUGH be conceived as similar to that in which one finds one’s self when learning to operate some little mechanical device: I go to my locker after a long absence, having forgotten the combination, and as soon as the muscular sensations come in from handling the lock, I find that I am turning the knob to the partic- ular succession of numbers which will open the lock. I become aware of the succession and by attending to the movements I am making, learn the combination from myself. In this case I have been observing the effects of the operation of a motor circuit. Supposing I am not particularly anxious to relearn the com- bination; my attention is on the paraphernalia within the locker. It is possible that I may open the lock without learning the com- bination from myself and may come to the locker again the next day without an idea of the key. Instead of being wholly interested in what I seek, it is only when I fix my attention on the means of attaining it that I am in a position to learn; but I can accomplish my ends perfectly without paying attention to the means, letting the motor circuits do the work, and this is the way, we conceive, that the mouse does in the majority of cases where it is successful. Now suppose that after I have become used to a certain combi- nation and can work it unconsciously, the combination is changed and I am told a new one. I receive the new series of numbers and commit it to memory. The next day I go to the locker and find, after working with the knob for some time, that my fingers have been using the old combination. I may even make this same mis- take for several days. This experience we may compare with that of the mouse when the lights are changed. The creature is guided by the effects of the motor circuits within its body to the right side, because it had become habituated to turning to the right when the food was to be obtained on that side. Now that the food is to be obtained on the left, it still goes to the right, and does so over and over again. Each attempt of this sort is of course recorded as a failure. How long the animal will continue to run into the wrong com- partment is a question of how long before it will begin to attend to the means to be employed in obtaining the food, or, it is a ques- tion of how long before the impulse to venture in search of food VISION IN THE MOUSE 595 will be overcome by the impulse to avoid the unpleasantness of the shock which ensues in case of failure, so that the animal be- comes unwilling to venture at all. In the case of the human being, if several repetitions are neces- sary to bring into the focus of consciousness the movements that that being is making, then we must not expect much of the mouse, and we must believe that a large number of the cases recorded as failures were not necessarily failures to discriminate between stimuli but, rather, secondarily automatic movements. When a visual stimulus succeeds in calling forth a change in mode of action as its response—thus overcoming the kinaesthetic influences, which would tend to bring about a former mode of action—there is involved a discharge of energy through the higher association tracts and a conscious accompaniment. The law of parsimony proposed at the beginning of this section, and also our general observations, forbid us to assume that such conscious accompaniment is involved, but rather make for the view that the mental processes of the animal rise into consciousness of its movements at intervals not so frequent as in the human being, and that the kinaesthetie sense is the predominant directive sense in the mouse. IV. STRUCTURE OF THE EYE OF THE MOUSE The animals used for the anatomical study of the eye were from among those that had been used in the experiments. Some gray mice were chosen and some albinos; also a few dancing mice were included among those studied. Metruop: The heads of mice that had recently died or had been killed were preserved in formol-alecohol. When ready to work with them, the eyes were removed and put through alcohols of increasing strength, for about an hour in each, up to 100 per cent. They were then transferred to xylol where they cleared over night and the next day they were put in soft paraffin for an hour. They were next placed in hard paraffin for a half hour and later imbedded in this same paraffin. Care was taken in the imbedding 596 KARL T. WAUGH that the exact orientation of the eye, dorsal-ventral and anterior- posterior, should be known and preserved as the paraffin hardened. In those eyes to be used for the study of the retina it was found necessary to remove the lens because, being made harder by the alcohol than the rest of the eye, the knife in striking it would have a tendency to tear the more delicate retina. The lenses were removed best after the eye was in the hard paraffin. A knife was passed through the block just cutting off the cornea, the lens was then easily picked out with a needle and the whole block was reimbedded. In other eyes the lenses were removed while still in xylol. The paraffin blocks were next cut by the microtome into sections ten micra in thickness. These were taken in serial order and mounted by the water method upon slides smeared with albumen. The slides were warmed, washed in xylol and plunged into abso- lute alcohol, after which they were run through alcohols of decreas- ing strength in order to use an aqueous stain. They were then stained in Delafield’s hematoxylin for three hours, washed in water, then counterstained over night in orange-G solution. After having been removed from the last solution they were run up through the alcohols again, immersed in xylol and finally mounted in Canada balsam. Resutts: o Bhs Oe Beta es 6 Ale Pe he OER ad be tal ee a PUP Sa aoe eee '‘* A Pa a 6 th . Ot ees re oe See O88 one le ae oe oe -* nat, KG A? / *~ yor CM ee ah & were sf 4 55 953) tar 4 Petters? wy 5, ce etat: ; 44, sete ane? ese On f é : ” 75 one Ph Pn Mere e > 4 aoe . . ~ v n i ; Pah ASI Ah petits VIrs a ad oe) Ca +-3- o? Sc . . oy